/musicprod/ - Music Production

i'm very unconventional in my ways and in a constant change. i haven't worked as described below for some time, and it is more of an exercise that may give rise to useful material.

one of my first attempts involved simple sampling techniques.

easy and fun, perfect for starters.

this is how i did it.

1. get audacity and youtube.

2. find interesting audio on youtube.

3. record whatever you wish on audacity, preferably of instruments or music.

4. use the samples to arrange, experiment. see what you can do.

first or all attempts might be useless but you will learn and it will be entertaining.

hopefully this won't be followed by a series of ridicule on my part.

one example of an audible result is here:

https://www.youtube.com/watch?v=RMorkMS-ro8

as i said, very unorthodox and sort of primitive.

>>131

Very interesting stuff. Can you remember one of the videos you used?

>>140

barely

think the melody was from this movie

https://www.youtube.com/watch?v=GvrKFZlYW_k

but i doubt my words, think it was some therimin.

found another inferior example which i have linked here. please lower the volume.

http://a.pomf.se/tndhbj.mp3

everything on this one apart from the drums was from the above clip. drums were just bongo drums lesson layers.

I think we should never have more than two stickied threads, and I think the banner and welcome stickies are more important. We can cycle this one.

>>168

Yeah tulpar stickied it.

Hi just found this thread because I'm a retard and didn't check the catalog before making a post.

Could anyone help me recreate the detuned synth found at 3:36 in this song https://www.youtube.com/watch?v=CVrtxeDwqAI

I'm still very new to sound design so any help is much appreciated

>>187

Oh I typically work in FL Studio, but the current project I'm working on is on Logic, so creating it in massive would be preferred

>>187

I don't think I'll be able to recreate the synth, but it's easier than it sounds. Imagine a long, tuned tone in this synth. Add some detune possibly, and use portamento (the slide-notes in the piano roll) violently and without regard to harmony.

You can probably achieve a more original and interesting sound if you try it out on different sounding synths.

>>187

Certainly, allow me to see what synth you have in mind, I'll try and decipher what it is.

>>187

Band passes. I'm willing to bet its band passes.

>>189

>>190

>>191

Awesome, thanks for the help. I knew recreating it wouldn't necessarily be possible, but I only wanted something similar anyway. I'll be sure to post a finished product

what do you ladies think of bitcrushing and tape emulation? i've been trying to get a BoC sound. shit's tough.

>>260

girls? here?

speaking of boc, have you ever noticed this vocal effect they use

http://a.pomf.se/vifzcs.mp3

>>263

isn't that just ring mod + pitch shifting? isn't that exactly how they make the dalek voice?

>>187

Savant uses FL Studio

The vast majority of his sounds are samples which are distorted and effected and distorted and effected and stretched and looped and chopped again and again, the rest is super minimalist design. He did about a half an hour demo video of how he makes music, and it was all in FL, with a fuckton of samples, and he did everything right there giving zero fucks the entire time, and what he made ended up becoming one of his later releases.

Strangely I think the spot that you picked out is the exception to this and is some of his basic minimal synth design. there's 2 sounds i'm noticing, both of which you could call detuned, One of them is a basic hoover, and the other is a plonky metal thing that is clearly both FM and RM. If you're using FL you can make both of those sounds in Sytrus, or Poizone, if not I think FM8 would work for the plonky metal thing, and by plonky metal thing I mean "sounds kind of like a transformer" and the hoover is that thing that sounds like it goes Woo woo woo wo-wo-woooo wooo

And it's litterally just a bunch of saw waves out of phase and slightly detuned from eachother, all doing a pitch sweep downwards to the root note over about 1/16-1/8 before it finally holds and is basically a reese. then you just add some reverb to it and a bit of EQing.

the metal dense reverb plonky transformer noise on the other hand, I mean, I'm pretty sure you could also pull it off harmonically in harmor, using prism shaping and a ton of other shit, but I can never seem to use prism for any good effect. I'd say check out seamlessR on youtube for some tutorials on how to use different synths.

>>267

slight correction for the hoover explanation, the pitch goes up sharply, then sweeps back down. I forgot it has an attack.

>>267

>>268

FFFFFFUUUUUUUU—–

I never use some of those synths so I got the name wrong, I meant to say Toxic Biohazard, not Poizone, TB is FM, Poizone is subtractive(?)

>>260

Bit crushing can sound cool if used in a creative way. Most attempts I've heard have just sounded annoying though.

I definitely over use it to some people, but I really like the sound it makes

>>277

meant for

>>260

I use bitcrushing for 3 things

1: cheap fast formant-y sounds, that I can then add more effects to

2: texturing waves to re-sample. For instance I like bitcrushing sine-waves in different ways to get a ton of artifacts in them, and then drop a single cycle of it into sytrus for oscilators, and then make all my automation on steps, so that it sounds like super old FM that is janky as fuck. (I typically set the bitrate way down, and do +Xdb before crushing and -Xdb after, for instance +55db pre, -55db post, this actually creates a different sound than degrading it by the numbers, depending on plugin used and how your DAW handles audio. FL is 32bit float so it handles it like a boss.

3: ambiance. I like making bitcrushed pad/chorus sounds. I don't do it often, but it works to super sweet effect IMO.

>>294

>2

This. For example, if a wave is too pure for any effect to properly affect the sound, bitcrushing/excitation might make it work and sound interesting.

Anyone know how to make vocal-like pads for massive or sylenth? Everything I make is usually too robotic, harsh, or cold.

>>303

you can do that with filters. the key is effecting the right frequencies. just look up formant frequencies on google and wikipedia, find out what the frequencies are, and just start fucking with it. it's one of those things you have to learn, you gotta know what frequencies, and how much, and which effects before and after it and all that shit, so you can just know how it's gonna sound on the other end.

vocal-like sounds are usually pretty hard.

the cheapest way is to just bitcrush to 8-24 bits and add a lowpass filter BEFORE the bitcrusher and mess with the lowpass, but that sounds also obviously bitcrushed.

the second cheap way is a vocoder and a vocal modulator, but I garauntee this isn't what you want, unless you're using a damn good vocoder and know wtf you're doing, which is admittedly HARDER than my initial suggestion, but requires less wrote memorization.

As for doing it WITHING massive or sylenth, in massive there's vocal filters and vocal-oriented WTs in massive, and FM, and using a triangle (with some extra frequencies) wave to FM another wave, and modulating the FM ammount/intensity, also creates a vocaly-formanty sound, but you're better off using an FM dedicated synth for that, it is not massive's strongest suite.

Sylenth on the other hand, I've got no goddamn clue.

once again I'm gonna refer to my earlier post >>267 and say to look up SeamlessR on youtube, he makes a MOUNTAIN of tutorials, and while he does use FL, he explains the technical aspects of things quite frequently especially with his earlier tutorials, and those processes carry over to other DAWs and other synths, for instance if it works in sytrus it most likely works in FM8, etc.

>>303

How about using processed chokr samples?

>>306

Choir* lel

>>303

specifically in massive or sylenth I have no idea, BUT I did record myself making weird noises and put it in harmor and edited the fuck out of it and turned it into a thing. it was 2 or more years ago though and I don't remember the specifics beyond having been in harmor and tweaking the fuck out of it additively. and I wasn't trying to be a choir, but rather a growl, which I sort of used as a morphing pad. you can check that out here if you want to give it a listen,

https://djquade.bandcamp.com/track/halpha5

I'm not sure if i still have the project file but what I do know I did was import the wave to harmor, and I messed with at least: formant and both scales on the image tab. other fun things for pad including unison (Especially in Hz mode) and harmonizer (if you use it right).

even making a "robotic harsh or cold" vocal-like sound can be softened, warmed, brightened, and humanized, if you do it right. Though humanizing it might require use of prism, and custom prism shapes (which you can set, and I don't fully understand how it works but I know a few of seamlessR's random tutorials explain it pretty well but fuck if I remember which ones I'm slowly realizing his videos are a mess of not tagged or organized, and never referred to pile of info-dump with no context)

so for your sake you could just record a short clip of you going "ahhhhhh" into a mic, adjust the pitch, throw it in harmor, and then fuck with it for half an hour and make the perfect sound. I'm not even good at it and I can make some interesting shit.

>>722

You're looking for synths. You want to learn how to sculpt synths into what you're imagining, and sound design (for example turning samples into something completely different with editing and effects).

Vision is a much more refined sense than hearing, so this skill takes a lot of time and effort to achieve. You're not gonna find any shortcuts. With audio, usually, one change affects the entire sounds, while in 3D sculpting, it only changes one area, and you can see all parts at once because it's spatial, while audio can only be heard one slice at a time through time.

Find some drone tutorials and see if that's more up your alley.

>>722

Also, here's an example of how I do sound design.

https://soundcloud.com/audiopolis/the-removal-of-music-background/s-IeoQX

It consists of heavily processed samples. There's no such thing as making music out of nothing (unless you consider a synth "nothing"), and I can guarantee you that the music you linked to was NOT made from nothing.

>>725

>There's no such thing as making music out of nothing (unless you consider a synth "nothing")

There is no such thing as creating a 3d model from nothing, unless a modeling software is nothing. unless a cube or a sphere is nothing. unless polygons are nothing

there is no such thing as a painting from nothing unless brush canvas and paint are nothing.

Synths are not sampled sounds, they are tools, like photoshop or zbrush, but for sounds, where primitive waveforms like Saw and square, are your colors, and filter types and re-sampling and effects are your brush types, and your envelopes are your brush stroke intensity.

Yes, there IS such a thing as making music out of nothing. just ask Bobby McFerrin.

I honestly hope you're a troll because you just straight up made me RAGE with that assumption, and if you're not a troll, there is no future for humanity.

>>733

A cube in 3D and paint in painting are comparable to sines in music. You can make music out of edited sine waves (a basic waveshape of a synth).

But if you want to make a guitar song, you need a guitar to record, or you need realistic guitar VSTis.

>>735

I guess FL Studio could be better for you. I feel like it's a more visual experience to work in. Besides that, you just need the correct plugins and skills. There's an endless combination of effects you can use.

>>737

That's underwhelming. Really, no software just for doing this one thing?

>>738

What exactly is this "one thing" to you? Model sounds? Sure, like I said, download a synth. The plugins I'm talking about are the software that does that one thing.

A synth is the one thing you use to model a wave into a sound you imagine. Among many other means of doing this.

>>739

>>738

I mean, it sounds like you're actually mainly interested in making presets for synths and plugins. Sounds a little boring to me.

>>735

the DAW is your "Digital Audio Workstation" due to your pic, I shall use a work-out analogy.

The DAW is your Gym, your song is your workout routine, and your various VSTi/Synths are your workout machines.

I suggest you check out my post in this other thread

>>706

you can think of different synthesis types as different tools that get different results.

>>736

>a cube in 3d and paint in painting are comparable to sines in music.

>you can edit sine waves

Editing a sine wave by its shape, is called waveshaping, it is done with effects, and stacking other wave shapes. This is called

>Subtractive Synthesis

There is also additive synthesis which is stacking multiple different sine waves into a spectrum of sound. the best Additive synths can in fact Take samples of audio such as single notes, and re-construct them additively, allowing you to play it as a more complex instrument. Meaning you can use it as a "realistic guitar VSTi", with synths. Check out harmor. Go ahead and download the fully-featured demo, and test out how resampling works. watch some tutorials

>>738

yes, yes there are, you're just being incredibly vague.

Serum lets you draw waves, meaning the actual shape of the wave. you can create any wave shape you want, and you can even morph between multiple wave shapes.

Harmor lets you use something like paint or gimp to draw spectrums, where vertical is frequency, horizontal is time, and brightness is loudness. as well as a second layer where brightness is pitch bending per harmonic.

every subtractive synth lets you stack different wave shapes and filter types to create the wave shapes you want, and on that note, if you understand which harmonics result from different shapes (such as square vs saw harmonics, and which shapes make them present) you can construct all manner of rich textures.

you're acting like there's no software to make all sounds. the problem is there's a ton of software and you have to be way way WAY more specific.

FM synths like sytrus and FM8 let you shape waves and spectrums by modulating other waves in speed and amplitude to re-shape them over time.

you must be clear about what you want it to do specifically.

Collectively this sounds like you simply don't understand enough about how sound even works in order ask your questions.

>>738

to continue off of my post >>741

There is no ONE synth that does EVERYTHING.

But EVERYTHING can be done with MANY DIFFERENT synths.

There is no one-stop shop magical program that makes everything. the closest you will get to that is harmor and serum because they can re-sample and edit the samples in very complex ways that only they can do, but in that respect you're not creating those sounds, you're sampling, which is only relevant for real-world non-synth instruments, or for vocals.barring those they can generate any sound without resampling, IF YOU KNOW WHAT YOU ARE DOING. IT IS A COMPLICATED PROCESS.

they can do these things, you just don't understand what they do, or what you even want them TO do. that's the problem, a fundamental lack of understanding.

Remember that a 3D model looks simple without a texture. You can draw it yourself, but you're only human and probably won't get very realistic results. A natural sounding sound has a much greater resolution than an acceptably natural looking texture, so you're gonna needsome tools. You can try to draw the waves of an entire composition, but you're gonna need to find the fountain of youth first.

>>744

if you're trying to make a song by drawing the entire songs waveform yeah,

you can think of each instrument as a texture, and the melody as a shape, and the whole song is a collection of shapes resulting in a finished model, if we're gonna stick with that analogy

If you're recording audio and need the rest of the composition playing in headphones to play/sing to, cut a lot of the bass and highs with a temporary EQ on master to minimize bleeding.

Maybe this isn't really sound design, but I feel that my question doesn't deserve its own thread, so whatever.

My music uses heavy drum programming with several different patterns (picture Squarepusher or Ricardo Villalobos). The problem is that this variety causes a hell of a problem within my current workflow. I use FL Studio and so I need to create a pattern for every single iteration of my elements leaving with dozens of different, very similar patterns that are almost impossible to organize or even remember sometimes. Is there a better way to manage drum programming with FL? I thought maybe I should arrange the samples directly in the timeline, haven't tried it yet. Should I switch my DAW?

>>773

Switching daws won't fix your problem. Simply put all the drums in a single pattern and stop it with the loops. You'll get a sense of the drums being a real performance, and you can change the drums (velocity change, crescendos, small variations, breaks, imperfections, etc.) exactly where in the song you need to.

I don't even loop patterns at all anymore. I put everything from one verse, chorus, bridge or part in the same pattern. That way it feels like a much more authentic live performance, and going in and changing something is actually faster to me because I don't have to look for the correct pattern and stuff, and I don't need a new pattern for every little variation I have. You could just put each variation next to each other in the piano roll and chop it up in the playlist. Also, label and color code your patterns.

>>773

Also, don't be scared of making unimportant threads as long as there's not already a good thread for it. We need more threads in the catalog so new anons stay.

>>773

FL Isn't meant to be thinking about patterns that way. you make a basic pattern or 2 (yes I'm including thinking of squarepusher). Consider this your backing drums and for builds first. Then make your first break and place it as all your patterns in the playlist first, then say "I want this spot to be a different break pattern" and click the top-left of if and select "Make unique", and it will create a duplicate pattern and replace that instance with the new one, you can then double-click it to open that patter and make the changes to it, without fucking up the rest of your track. you can do this to each individual loop in your playlist, and if you want to re-use something, you can just erase one from the playlist and copy your earlier break.

I tend to keep each "layer" or FX channel in 1 row on the playlist. it's designed to be arranged that way since you can name rows. For instance, a breakbeat row.

Even with something like squarepusher or AFX, I can only really see you needing 20-30 patterns at a max, unless you're doing 1-beat patterns chopped up all over the place at bloating it to 80-120. which I should point out FL supports, what, 999 patterns?

Also one thing you can do is apply channel filters per instrument.

that's the drop-down on the bottom left of your pattern/step sequencer. it lets you hide different groups of instruments.

the other thing you can do is give categorized names, and color-coding to your individual patterns, which can be done in the top of the pattern/step sequencer.

This means when you view a dropdown of all your patterns, they are color-coded and can sort by name, and I *THINK* tree-view but I think I could be wrong there, I'm not sure if they have tree-view for patterns yet, they do for presets though. and when cycling through patterns you can hide irrelevant instruments so that you can find only the patterns with a given list of channels.

Maybe questions like this should be in.. I dunno, a FL Q&A thread? next time anyone has a question for FL just start like an FL General or something.

>>793

>FL thread

Made.

>>776

>>793

Thank you, guys. I'll try to test both approaches this weekend and see which works best for me.

Serum;

for a dirt bass just add flanger, no rate, modulate depth.

>>1134

I've been just fine with fruity EQ 2 honestly.

>>735

I would recommend you Fl studio,the piano roll is the most "aproach" of a visual modeling,and this DAW is very newbie friendly,diferent of Ablenton.

Hey guys,some questions here.

https://clyp.it/rqyqzgxd

What is this "wheezing" at 50s?I am using reverb,and when this automation clips come makes this noise,I puted a convolver on it and looks worst,lol.How can I ease this?

And this single snare at 13s-14s?Have more than one on the full audio.I am using a single pattern,and they are not colliding,so i have no idea what could be.

I am using Fl Studio btw,thanks for reading.

>>1140

Instead of automating the reverb's wetness, automate the reverb plugin's mix level (the round little knob in the plugin list).

>>1140

that crackling sound is called a buffer underrun as far as I am aware.

If you automate the durration/decay of a reverb (or any delay other than a dub delay which is unique) what you're doing is re-setting the buffer length of a delay, which erases what's in the buffer. this means for a split second, even individual cycles, there is no reverb, and then it kicks back in at full force. If you automate smoothly this happens many MANY times over the course of your automation. If you switched from a smooth automation to steps, you would get one pop at the start of each step instead.

If you're curious about the exception being dub-delay, instead of clearing and re-sizing the buffer, it stretches the buffer, stretching the contents of the buffer to match. but i've never heard of a dub-delay based reverb so it's not particularly relevant.

as >>1141 says, if you want to have less, or more reverb, you want to adjust the wet/dry mix of your reverb, don't edit the actual reverb itself.

>>1141

>>1147

thanks for the replies,i got it solved. :3

I've been considering buying Valhalla Ubermod or U-he Uhbik. Which one do you guys think would be the better buy? I already have a few Valhalla reverb plugins, but I'm impressed with the effects U-he include with Diva. I'm mainly looking for delay and phaser plugins right now, but Uhbik looks like a good deal considering all the effects included that I may be interested in in the future.

Anyone have any experience with either of these plugins?

>>260

I've played around with bitcrushing and didn't really like it. I don't have much experience with tape emulation, but it seems like it works nicely if you want a reel-to-reel sound without all the hassle of maintaining a tape machine. If you want a lo-fi sound, you could always pick up an old 4-track cassette recorder for under $100. You could get some interesting results by screwing around with one that allows pitch control.

Hi, I hope this is the right post to ask this question, but where does one start ? I mean I tried doing some beats on fl studio, but they're just a mess. Can someone tell me where did you guys start and how do you guys improve your craft ?

I really hope I can get some good feedback from you guys, and if this isnt the right place to ask, please direct me where I could ask such questions.

Thank you very much

>>1291

No, this is the perfect place. Try the link in the sticky, although it's far from finished, and watch some of the videos.

First of all, you need to get familiar with how FL Studio works, and how you use its different features.

Do you record or sample?

>>1292

That's the problem man, I tried both, but it didnt give me anything good, I'll check out the sticky, I'll come back if I have more questions, if you want to help me out more feel free to link me up to youtube videos/forums/etc...

thank you :)

>>1293

Try getting a sample library and some synths, then try making melodies in the piano roll. Make sure you learn how to use the mixer and playlist properly.

>>1293

Also, try asking about specific problems you're having. Where to start? That's a tough one. I personally started at age 10 (20 now) by trying to replicate the soundtrack of Lord of the Rings with Fruity Soundfont Player. I was terrible, but I eventually got the hang of harmony and rhythm and stuff, and later got into making realistic music (both orchestral and band) with sample libraries. I didn't read or watch tutorials at all, so it took me years to figure out how the mixer and playlist worked, which I did by accident.

Do you already play any instruments?

>>1295

I have basic knowledge on the piano.

I've been trying to do a beat for my friend lately with a sample from Caravan Palace - Brotherswing, I'd like to know how to structure a beat for a hook and verses, sometimes in beats theres reduced drums or other melodies when the verses are spoken but sometimes its the opposite, I'd like to know more on how to structure beats with lyrics in mind, could you help me with that ?

>>1296

I don't really make beats, but it sounds like mixing is the fix. You just have to make sure that the instruments don't get in the way of the vocals sonically. The first stuff you make isn't going to be good. What genre are we talking here?

>>1297

trap, rap, from hard banger beats to more smooth r&b beats,

that is what im aiming at

I guess it really just comes naturally with time and effort

I just realised that no amount of links can really help as much as practicing, thank you for your answers,

I know 8chan gets a bad wrap but you guys aint so bad, good luck out there !

ill keep you posted up and you guys can give me more tips, thanks alot <3

>>1298

...mo.... I know a mo. I'm going to briefly have a giggle as if you are that mo, and now seriously answer some things.

so, as a point of comparison, if you ever listened to dubstep, you'll know that a lot of the emphasis is on "the drop", where it's building tension, and then it releases it and the rest of the song kicks in.

That process, of building up tension and then releasing, is the reason you get "beats" that change over the course of a song between different parts. For instance if you're doing rap, you might have the tension build with a break in the drums where it becomes instrument and A Capella with no drums, and then the drums snap back in on a downbeat, with that beat being accentuated, creating a brief moment of tension and release mid-verse.

another thing is where instruments might be louder in the chorus of a song. this is because the vocals of the chorus are typically also louder and more pronounced which means it's easier to hear them. Another reason they might sound different beyond being just louder, is compression for instance it might be going through one compressor, but then get pushed into it harder during the chorus.

The way you mix also more depends on genre, tempo, and instrumentation. there is no "one way" to do any of those. if you're doing R&B or rap, you've typically got something that's very down tempo, which means you can afford to have a heavier beat, with lyrics which are spaced to be primarily between beats, but still starting on beat. I'm not much for rap, but i've listened to a bit here and there, and the best example of this that comes to mind right now is eminem's mosh, because you can clearly hear the very basic rhythm, and there are many clear points where he gives a pause either before or after a beat, and provides different emphasis on different lines based on this, which is most likely somewhat more specific to what is being said.

I know everyone else here seems to think this is mostly about mixing, but Honestly I'm reading into this more as trying to work out how to structure a song, and build tension, which is the basis of even making a good hook in the first place.

What i'd suggest doing is picking a couple of songs that are good examples of what you're looking for in terms of quality, and mood, and take notes on how the song is structured, when each change happens, and anything relevant to it that you notice for instance lyrical content, change in mood, etc, as well as the basic song structure, is it just verse/chorus/verse or does it do anything else? does this have an intro, a solo, does it get a key change. does it follow a structure that is not verse/chorus/verse and if so, what is it.

Look at it as "why would I do things this way in a song? what would doing this step here accomplish?"

It's kinda sad. I know all this stuff, and i've been told i'm good at giving advice, and that it usually works out... but I never seem to take my own advice.

Also, don't just re-create someone else's song. that's the one risk with deeply analyzing an artist's work. you could end up sounding too much like them. which is why I suggest looking at multiple songs, from multiple artists.

>>1301

thats not it btw i cant put drop fl studio filess apparently...

do you guys have an alternative ?

>>1302

well you can render to MP3 and use soundcloud

and if you want to share an FLP you can use pretty much whatever to upload. for reference I use cracked FL 11.something so I can't load anything from FL12 and beyond to compare

here's my soundcloud

https://soundcloud.com/milest3hr4t

you can message me there, and if you want to add me on steam, message me on soundcloud, I don't want my steam page linked on a chan board because of bots scrounging the nets to scam and hack steam accounts that are active. i've had people try to log in as me and tripped steamguard about 3 dozen times this summer, and gotten like 90 fake steam invites trying to phish. shit be insane, which is why I only link to it in private groups and private messages.

but seriously upload a few of your things to soundcloud or find some other things people on here have used to upload, and use that. (i forget what all people use) and at the very least I'll check it out, though hiphop and trap aren't really my thing, I at least understand it somewhat.

>>1303

I'm starting in rap,trap and banger types cause that's the type of music I grew up in, but I listen to everything, about a year ago I started widening my music, from pink floyd to lost prophet, run the jewels, etc...

I'm just trying to start with what I'm familiar with and hopefully I'll widen my range, I hope I can do more summer house type beats by the end of the year, you know songs that dont need lyrics to be enjoyed, but I know summerhouse is quite complex, so tackling it by starting the more simpler stuff.

I don't have much free time right now but I'll upload some beats to soundcloud and you can send me some feedback and stuff, if you want of course.

Thanks for the help, and I hope we can be good friends, those are quite rare to find these days.

>>1306

I personally suggest you just make crap for a few weeks without even trying to make it good, so that you get familiar with the software.

>>1307

thats exactly what im doing lmao

>>1308

Well, how's it going?

>>1309

each time i practice, it gets a little better,

this weekend I'll try to focus on a big project...

I'll post it up if im happy with the results.

>>1310

You should post it regardless. Might get useful feedback.

>>1311

Yeah I guess feedback is always good

So, it wasn't in this thread, but it was some other thread that I can't be assed to find, but It was on sound design, and I was attempting to make Metalic sounds, and at one point using a phaser or a comb filter came up.

So I figured out how to make a metallic sound with a phaser, but it wasn't pitched, and I wanted to be able to pitch the thing, and be able to pitch bend. so I got to work trying to figure out "How in the hell do I keytrack a phaser with FL 11 and fruity phaser?!"

This is the result.

More detailed description is in the actual thing

https://soundcloud.com/milest3hr4t/keyboard-tracking-a-phaser-for-pitch-is-hard

I am not happy with this, but it's a thing. looking for questions, comments, and criticism.

>>1317

ayy lmao

hey /musicprod/, i was wondering if you guys new of any chiptunes plug-ins for Acid Studio

>>1226

Disregard this post. Diva keeps crashing FL Studio on me, so I don't want to spend money on any of U-he's stuff until they're more stable. On the other hand, it could just be FL Studio's fault.

I'd still be interested in recommendations for a good phaser effect, though.

>>1321

Look into VSTs like Famisynth, Chipsounds, and FM Drive.

>>1322

are there any vst's that will let me filter the instruments into 8bit sounds?

>>1324

There are bitcrusher VSTs out there if that's what you're looking for.

>>1325

>bitcrusher

just looked up a video on it

that's what i was looking for & more. thank you so much

now you guys helped me alot,

heres my progress for now

couldnt upload the webm, i know you guys dont like links sorry about that: https://soundcloud.com/m0ras/mellow-shit-1

>>1329

Nicely done.

Maybe it's the style you're going for, but I'd like to hear more consistent and on-beat rhythm (paying attention to beats and bars and quantizing).

Also, maybe it's on purpose (could also be my shitty, temporary headphones), but the whole thing seems pretty inaudible. Put some gain on the drums maybe, and EQ out useless frequencies (mainly low end) that might occupy a lot of headroom.

>>1331

Check out Gross Beat if you use FL.

>>1330

I'll try to clean it up a bit, but I'll have to inform myself on how to do that.

its true I heard my beat like about 30 times while editing it and I had a headache at the end.

could you inform on how to trim the ambient sound ?

>>1333

You're using FL Studio, right? Put a Fruity Parametic EQ 2 on all your instruments, scroll up/down on the shape of the far left (purple) one and pull down on the square beneath it to make it steaper, then drag the purple circle to the right until you can hear an audible/unwanted loss of depth, then pull it slightly left again. I'll post pictures when I get to a computer, terrible description, sorry.

>>1334

tried fiddling with it, its not an earbleed as before.

https://soundcloud.com/m0ras/mellow-shit-cleaned-up

sorry about the link again do you guys have any mp4 or just sound recording softwares in mind ? That'll help me out

thx

>>1336

I figured it out, tried to open back again after the first try the file is corrupted... its doomed to stay like that i guess

but thanks now I know a little bit more, working on a new song

>>1337

http://forum.image-line.com/viewtopic.php?t=8771

>>1338

thanks, got the file back <3

Also, on bass instruments (particularly bass guitars) you sometimes want to EQ out the high end and only the sub-sonic low end (<20 Hz). I also like to limit the fuck out of my bass guitars (Fruity Limiter works fine, lower the ceiling and up the gain so that the attack and most of the sustain are above the ceiling).

If you want an instrument to sound far away, you want to EQ out a good portion of the low end and low mids, and add reverb (with the low cut maybe 1/5 up).

>>1340

huh, okay thanks for the tips,

I'll post results when I'm finished

>>1331

if you use gross beat, set up custom patterns and don't rely on randomization.

Also check out

http://destroyfx.smartelectronix.com/

particularlly buffer underrun is quite nice, I used it in https://soundcloud.com/milest3hr4t/error-delta-v

it's what draws out the piano chords and makes the higher buzzing, though I have it before reverb. If you ever get a video that sticks or skips, or a game crashes with repeating audio, or if your DAW runs out of ram and hangs or crashes with looping audio, that's exactly what this vst is imitating. it can be turned on and off, with customized sample size and sample rate, and can be selectively beat synced or not. can be LFO'd internally, can be controlled manually.

I would also suggest checking out the sequencers here

http://www.xoxos.net/vst/vst.html

I used "Circuit" in combination with dblue glitch, to create the "drums" in this https://soundcloud.com/milest3hr4t/virtual-factory-man

and by drums I mean this track consists of an 808, a distorted kick, and a drone that kicks in 50% of the way through. yeah. all those sounds are an 808 with a very weird pattern/sequence with a glitch effect.

and finally if you're experimenting with glitchy sounds as added sounds, try just generating weird loud harsh noises and turning them way down, and distorting them harshly with things like waveshaper (if you don't use FL I don't know an alternative). this is for like weird feedback, broken tube amps, etc. basically if you know what the type of glitch/error/hardware damage sounds like, you can re-create it manually, or just kinda fudge your way through it. If you're talking about adding glitches to a sound you already have, either you're cutting and filtering it manually and adding sounds, or you're actually glitching it. If you databend a sound however, be aware you'll most likely only get staticy crackles

I did that last thing to create sort of psude vinyl popping on a track, I made a single pop and sequenced them semi-randomly in tight clusters and not so tight clusters, and with variable delay, just to add some ambiance to a track. specifically this https://soundcloud.com/milest3hr4t/influenced

but you know, that's just my shitty 2 cents. i'm not 100% sure what kinds of glitchy sounds you're wanting, and that's always hard to give examples for. glitches especially, because there's glitch music, there's IDM style glitch, and then there's authentic audio glitches, or you could add the aesthetic of image glitches as we perceive them and do similar things to the music (which is, again, more like IDM)

there's lots of stuff out there.

Unrelated to my previous reply, I have done another sound design experiment.

I was attempting to re-create the types of sounds that the geth made in mass effect. I... thought it would be cool to have a voice like legion for vocals.

These are my results, https://soundcloud.com/milest3hr4t/geth_test

the first bit is random-ish noises, the rest is made from an acapella of laugh at life by mayhem. semi-detailed description is in the... description.

I will say though, that I cannot for the life of me remember how I set up the vocodex. I did, something weird, but I was half asleep when I did it, and can't remember what it was, but I do have it saved, and can nolonger make heads or tails of it.

Am looking for comments on the sound.

I'm thinking of adding auto tune to it, but can't decide if I should auto-tune the modulator, or if I should auto-tune the output. I'm also not 100% sure how I should EQ this to be clear, I can't clearly understand all the vocals, and I'm not sure what to do to make them more clear.

>>1342

I just wanna let you know that i love you. This type of tips are great for somebody, who is a complete newbie into making music.

>>1345

Most of what I said was just knowing a few common glitch-type sounds & styles, with how they work, and then through coincidence knowing of some plugins that do those things. technically you don't need any of those plugins, and simply need to understand what they're actually doing, because really with the possible exception of the buffer underrun, there's absolutely nothing in that list of links, that you can't do manually, if you only knew how.

For instance, I mentioned the vst "circuit". what that does is lets you set up a series of patterns per note, where you're controlling the probability that a note will play at any given time in that pattern, and setting the probability that each pattern will play. you could literally write music that way using dice. It would take way longer to do, but you could do it. because of the probability factor, if I re-load that .flp file from of "virtual factory man" and simply rendered to MP3 a dozen times, each file would be different, with different rhythms, because I didn't actually decide on a set rhythm and didn't record a clean mix of the output (my mistake honestly)

The single biggest tip, is to understand what it is that you want to make, and understand how that sound works, and then piece it together. this applies to literally 100% of the things that you will ever do, in music and otherwise, so I don't really consider that advice, because it's common sense I thought. I guess common sense is a super-power.

Then again, I suppose glitch is sort of weird to get into. like, you'd have to listen to a lot of glitch music and IDM, to get familiar with the things they have in common, and really think about what's happening for the differences, and like, get into datamoshing and databending, and learn what's actually happening. That's admittedly one that's harder to figure out, but most of it is just randomization of some core things, like rhythm and pitch.

Like, take a thing you want to do, and take it apart, make it into little pieces and look at them all, it's like taking apart a clock and putting it back together.

I'm not getting inspired, when I first started I was so ambitious, but right now im discouraged by all the thing I have to learn.

How do you get inspiration back /musicprod/ ?

>>1373

Listen to your favorite artists and strive towards their level. Just focus on the song structure and such for now with crude sounds, then imagine how the crude sounds should sound in the end, then play around with synths or samples until you achieve what you want, then focus on mixing, and finally mastering. In the process you'll get a little bit better at each.

>>1373

Just remember, Vanilla ice is a fucking idiot and he topped the billboard charts.

you don't technically have to learn everything, you just have to learn how to use a tool to do what you already know.

But like sometimes I say to myself "I'm not inspired but I haven't done anything in AGES. time to experiment. I'm gonna spend like an hour trying to make synth drums" and somehow in the process of doing that and experimenting I'll end up making a whole track and go "well damn" and then I'll also have all this un-used stuff that didn't make it into that song and I'l be like "this stuff is pretty good too, I'll mess with that" and then BAM another song.

and then i'll be like "eh... what now" and then i'll play vidya or watch a movie or derp around on youtube. I'mma have to share with you some of the horrifically embarassing shit that I USED to make back when I got started. the shit that I made back when I was a furry, before I knew what soundcloud was and posted to fur-affinity. the stuff I made back when I only had FL 4.0. back when after like 4 years I went "Oh, there's an FL9... thanks for the crack bro" and then having new toys to play with gave me more inspiration.

This is that guy that you added on steam btw, I'll send you some stuff for inspiration on there.

in the meanwhile, for me meanwhile that is, i'm finishing this post, which is for everyone.

The best source of inspiration for me has been experimenting in new ways I've never done before, or experimenting with a new synth, or a new effect, or a new type of rhythm, or a new way of structuring a song, or a new genre of music. Doing new and different things inspires me.

Another thing that inspires me is saying "hey, this was a cool thing that happened, either to me, or even to a fictional character, or a thing in a videogame, I want to do a thing about that" and i've done a few things like that with music inspired by things that happen... music that tells a story.

and lastly, remixes and covers. Sometimes I've got nothing else and i'm like "I want to derp around and play someone else's song, but with hardstyle kicks and cheezey plucks, and a vocoder! and dubstep wobble! and stupid shit!" and i've done quite a few thigns that way.

TL:DR

Experimentation is a great source of inspiration and motivation

Experimentation is also a great way to learn

and learning is a great source of things to experiment with. Like, Learn one thing, and focus on it, and do as much as you can with that one thing, until you feel burnt out on it, then grab a new thing to learn that's totally different and just keep doing that until you run out of new things and have to dig for more abstract new things, like trying to make an IDM reggae cover of two-headed boy, and ending up with something that resembles none of those things but is still kinda cool.

Does anyone know what synth was used for this dreamy pad sound that starts at 1:17? I know Uncle Varg has used a Casio CZ synth in the past, but I don't know if that synth was the origin for this sound. A good number of sounds in the track sound like ROMpler instruments or something.

I've read good things about VirtualCZ and will look into it if it's capable of those kinds of pad sounds.

>>1385

Funny, my cousin "worked" with Varg & co for a while. But I have no idea.

>>1386

Really? That's interesting. I do know he's kind of reticent when it comes to answering questions about what equipment he's used because he's wary of people becoming too consumeristic and attached to certain brands. I'd wondered in the past about those bellish sounds on Filosofem, but through a bit of searching found people on sites like Vintage Synth Explorer mentioning the CZ synths.

In this video at 1:54 you can hear the "Rundgang" bell-like patch I was looking for. I'm not sure that these synths could come up with a pad as lush as the one in "Der Tod Wuotans." It sounds like he was using a completely different synth in jail judging from that string sound and the timpani rather than the "budget DX7" timbres that Casio's phase distortion synthesizer line was capable of.

>>1387

Yeah, I've never really talked about music production with my cousin. He never took it too seriously, he's twice my age and hates people. He did something synth-y called Hardangervidda or something, which has a similar feel. If the opportunity presents itself in the near future, I'll ask for you (he probably knows), but I doubt it will.

>>1391

Is this the album he worked on? I'll have to give it a listen.

>>1392

Yeah, he's Ildjarn.

>>1394

Also, I could never really stand his music or the genre in general.

>>1394

>>1395

Huh. I recognized the name but never actually looked into his stuff. I listened to Hardangervidda last night and found it to be pretty good, and I'm actually listening to it again right now. It reminds me a lot of Wongraven, but with a more transcendent, outdoorsy kind of feeling to it. Not sure I'd like his more straightforward black metal stuff, though.

Though black metal is one of the few styles of metal that still appeals to me, I can get why a lot of people don't like it. Many of the bands try too hard to sound aggressive and come across as just plain noisy and obnoxious a lot of the time (Beherit's non-electronic stuff comes to mind), which really turns me off. I think the whole idea of creating unappealing music to spite those dirty plebeian normalfags is pretty stupid.

>>1396

Agreed. While I don't listen to any metal genre, it's specifically the kind you mention that I can't see the appeal of besides being sort of a fashion/political/religious statement. Vidar/Ildjarn's vocals are so intangible that no amount of editing could reveal the words. There is minimal effort in song structure and anything resembling it, and I guess that's a "point", but it's not for me. Music, to me, is supposed to sound pleasing (without defining what pleasing sounds like). Which is why I never talked about music with him. At least he's not a phony. Full time vegan, misanthropist, anti-social, the whole package, without wearing black and spooky clothes just to make sure everyone knows you're weird on first glance.

The low production value ambient style of Hardangervidda is also not my style. I hear people compliment the melodies and harmonies, but I'm not sure I hear what they hear. Sounds "guessed" through blind trial and error to me. I imagine it was finished within days.

Anyway, just my opinion. I guess I just prefer fancy convolution reverb-based pads and granular synths than something that sounds inherently computer-generated.

>>1409

Sounds like a very cheap/free brass/trumpet plugin. Could be samples as well. Google "free trumpet vst" and find a really old one. You could also try browsing factory presets for a few synths and find a similar one, probably called "trumpet" or categorized under "brass" presets.

>>1409

Pretty sure it's a Roland Sound Canvas with effects over it.

>>1410

>>1411

Thanks lads. I'll go find me some trumpets I guess and check out the Roland. Wasn't expecting any replies this fast.

>>1409

sounds like a soundfont to me, or someone sampling something. maybe the korg M1 or something? I mean this seriously sounds like a super old sample-based synth / soundfont, like something that would be in a nintendo DS game, or an SNES game. very flat and low-fi. or like, a nintendo wii menu somewhere. and nintendo just loves their samplers for some strange reason.

that's just my impression of it. someone using a sampler, to re-sample something that was a vintage sample.

>>1412

Well I now have a shit preset-filled keyboard thing and I noticed that the French horn one sounds pretty similar to it, I guess with some clever tweaks you could make it like that. Probably not though.

>>1468

Is it a plugin?

So I was experimenting with text to speech and how badly I could fuck it, and I did this.

http://vocaroo.com/i/s1EXriVO0Ljw

I had a bad time.

>>1479

How'd you do that? I'd consider this time well spent.

>>1480

step 1. stock text to speech, Voice - Robotoid, monotone. enter text. Sample in edison

step 2. drop in harmor for resynthesis

step 3. https://www.youtube.com/watch?v=MFMgdzh0YMA

step 4. mess with the formant, formant scale, and mix scale

step 5. wonky as fuck note placement that's way too fast and has no consideration for what's being said.

With some added fiddling with newtone to get the timing of each formant to the right position in time, I can picture this doing some pretty kick ass things. I just can't write lyrics. Hatsune miku can just suck a bag of dicks.

I'm tempted to do the vocals of an existing song, but I want it to also be funny, so it needs to be a song that would be popular and meme-worthy. I just... I dont' want to use all star, gangnam style, never gonna give you up. I want it to come out of left field, like something like neil cicierega would do, except he did allstar to death already.

I could probably make people cry with hotline bling. except I fucking hate that song. So I'm open to suggestions.

>>1481

Thanks, gonna make something a little different using this method. And yeah, Miku < MS Text to Speech.

>>1482

I wouldn't really say miku isnt' as good as microsoft sam or anything like that. It is after all designed to sing, it's just not made for english. It is however, not free.

That's the distinction, sam and robotoid and speak and spell... there's free versions... there isn't a free miku. and miku is a pain in the ass to set up apparently, or else I'd be using it myself.

Other things I'd like to point out

-Auto-tuning microsoft sam

https://www.youtube.com/watch?v=Z20IzaCAgtU

also, Using a VOCODER on it

I mean.... the only thing stopping me from doing things with it is that I can't think of lyrics.

as for the thing I said about being tempted to use this to do a cover... I'm thinking of doing "Semi-charmed life" in a minor key, as a robot acapella, but that's not particularly meme-worthy, I mean a robot singing about meth addiction sex and an overdose might be cool, it's not exactly meme worthy.

>>1483

Of course, my statement that MS text to speech > Miku is under the assumption that you use a vocoder, plus time stretching and other adjustments where needed. If Vocaloids were easy to use, I wouldn't say that. But I've tried using vocaloids, and they're a nightmare to use. In English anyway. There are no English vocaloids that don't sound cringe-worthy.

And, of course, it depends on the genre. I just needed a robotic and feminine voice saying "let go" over and over. Text to speech was close, but I ended up sampling an obscure youtube video.

How niche do you want you meme-worthy robotoid cover to be?

>>1484

I dunno, how obscure was everybody walk the dinosaur? Every time I try to think of something like that though I end up with something like i saw the sign by ace of bass, and that's not nearly energetic and meme enough.

though after yesterday, I was very tempted to just do space oddity, but moonbase alpha already did it better.

In other news, apparently, Not the only ones to do this.

http://lapfoxtrax.wikia.com/wiki/Kitcaliber

"the duo's female "vocalist", whose vocals are actually generated via Cepstral text-to-speech tools and more recently Google Translate's text-to-speech function, which are then time and pitch corrected)"

I gotta say, the lyrics on D.FREQ.CRUSH sound pretty damn good for pure text to speech.

>>1485

--not all tracks on d freq crush seem to use the text to speech,

raystorm and vultures definately are, but i'm not so sure about three wonders or pseudotriton

Q:Is it possible to play a sample(or set a channel)in a exactly frequency rate?If you want examples i can post.

>>1496

Post examples.

>>1497

https://clyp.it/pr3umr0z

On 4-5 sec starts kind of a cymbal.

the quality wasn't the best

>>1499

Still not exactly sure what you mean. Nothing out of the ordinary was made with that hat. Sounds like a regular old sample.

Try to describe what you want to do. You have a sample that you want to do something with. The sample, I assume from the example, is percussive. How do you want the sample to be different from the original sample?

>>1500

When I listen to it in a headphone feels like the "cymbal" is on top of everything. i think it is connected with frequency

posting this everywhere because i really need help so

help

>want to be a producer

>dont have a library of samples and synths to use for production

How do I get them? How do I build my sound library to the point where I can sit down and produce some quality music?

I cant make anything right now because I have nothing to work with.

I have a bunch of presets, only know how to use Sylenth1 just barely.

What I want to know is how and where producers get their sounds and how theyre library is set up.

Someone add what im missing or correct me, producers get their sounds from:

>sample packs

>presets

>tweaking presets to get desired sound

>making own samples

pls respond

>>1501

I think what you're hearing is just the frequencies of the sample. All samples consist of frequencies. EQing out some room in the other instruments will make it sound less cluttered and the sample becomes clearer. There's nothing fancy going on in that clip.

>>1505

Thanks.

and for everybody that still come here

>>1501

the only thing that I noticed odd about that, is the drums that kick in part way through.

it sounds like they have a slightly different sample rate, which wouldn't effect if they're "on top" of everything, that's purely an audio quality thing, and secondly, they sound just slightly high-passed.... High pass? low-cut? are those even different?

all high pass/low-cut serves to do is give more headroom for bass, which means they can be a bit louder. being louder means that yes, they'll seem like they're on top, but just doing a lowcut on drums makes them sound really wimpy and takes out all the punch.

>>1502

the only sample packs I use, is for real drums, physical instruments I don't physically possess, and random foley sounds to edit the fuck out of.

It depends on your outlook on music. Do you want to make music because you want to make a song people want to dance to that you can play in the club, or do you want to make music because you enjoy doing it and you just wanna have fun and if people buy it along the way then kick ass.

if it's the former, the answer is sample packs and preset packs, and then tweak those presets/samples.

if you're having fun, or if you're making really complicated EDM genres or getting into experimental music, then you absolutely absolutely must make your own sounds.

For instance if you're making some kind of pop/dance track, or a hiphop beat, yeah, use samples and presets. If you're making dubstep or something, that's sort of the in-between and you can do both. Oh, you wanna make drum and bass/techstep/neurofunk or triphop or futurefunk or hardstyle? time to start making your own sounds and learning how to use your software and hardware and learn all the how-to stuff.

If you want tutorials for making sounds withthe synths and stuff that comes with FL, I suggest looking up SeamlessR on youtube, he is the best at that, otherwise, I suggest looking up tutorials for the specific plugins you have, as well as look at what TYPE of synth it is "FM, Additive, Subtractive, Wavetable, Etc" and look up general tutorials for those synthesis types.

For example, Sylenth1 is a "subtractive" synth, which means it takes basic waveshapes (sine/triangle/saw/square/noise) and stacks them together (positive and negative interference,) which you typically invert the phase on some of them to minimize headroom and keep the same character of the sound (so you can fit more sounds in the mix), and then add filters and LFO's to it.

So any tutorials for subtractive synths that are JUST subtractive, will apply to Sylenth1, and any sylenth1 tutorials will apply to other subtractive synths. The only difference is where the knobs are.

As for Samples and presets. there's tons of places, especially online, that sell presets and samplepacks. a few sources off the top of my head include Black Octopus, Loopmasters, and Cymatics.

And remember, if you sample a song, get the legal rights to do it or you will eventually get busted for it if it gets big, at all.

Not gonna tell you you have to do it one way or the other, because tons of successfull people only use samplepacks and don't have a clue what they're doing for sound design, and only know about mixing and composition. some of them don't even know that. which is a sad state of affairs for the human race as far as I'm concerned.

>>1134

It's super awesome and I bought it on a sale a few months back.

Your explanation is kinda wonky tho. It uses allpass filters to shift phases of specific frequencies, so you get that zzzzzzap on transients.

Also holy shit this board exists.

>>1522

There's more to quality than bitrate. Bitrate is about compression, and lossless wave files have huge bitrates. Resampling is more important when you know what bitrate to export to (which depends on the medium and filesize limitations). There's also the "time resolution" found in the project setup menu in FL, for example, which you'll wanna keep low in order to save CPU usage until you finally render it. The rest I can't answer.

>>1522

There are 2 indicators of audio quality.

1: Bit Rate

2: Bit Depth

Bit rate is how quickly it moves between points

Bit Depth is how much variety there can be between points (amplitude)

Vinyl and Tape are both restricted to a bit-rate based on the molecular structure of the material, vs the rate that the player scans through the physical material (Tape speed or RPM respectively)

The lower your bit rate, the smaller your frequency range for instance, 44100 hertz is optimal, but most people cannot tell it apart from 22050 hertz,

Vinyl is less than 22050 hertz, but not by much

Bit Depth effects clarity for how loud sounds can be, but also in extremes can create aliasing. Vinyl wins out for having a high bit depth, but loses out for having a lower bit rate.

you can compare this by recording vinyl to 44100 hz floating point bit depth (theoretically infinite) and then compare the spectrum's range to other HQ digital recordings of the same piece. There will be a HUGE cut off the top of the frequency range, AND there will be frequency balance issues on vinyl.

The highest quality analog recording that can be obtained with household recording/playback devices is with a Metal tape (expensive as shit) using Dolby Type S noise cancellation. However, No pre-recorded audio exists on that format because it was abandoned for CD's. Ironically Metal tape + dolby type S gets higher quality than early CD Audio, and unless you're comparing them directly, is indistinguishable from lossless.

The highest quality possible in any environment would be 2-inch metal tape with I believe Dolby Type S is better than the professional dolby system that existed before it.

The reason that Vinyl is often considered better than everything else, is because of when Tape was new all we had was feric and chrome tape, with dolby type A, or 8-track, and MOST people had access to tape through really bad car stereos. Vinyl is objectively better in every way from that point.

After that however, when CD came around, it was compared to the quality of tape, not vinyl. and those people who stuck with vinyl made the assumption that this is because CD couldn't compare to vinyl.

Early Digital recordings WHERE worse than vinyl, when we did 8 and 16 bit bitdepth with 22025 hz or 11012.5 hz, because early digital audio converters (DAC) where really shit, and the ability to store audio on the computer was limited to uncompressed PCM, which is HUGE and we didn't have the storage space.

Digital audio didn't really become viable until the 90's, which is when it finally got better than MOST tape. However in another universe there's probably a world full of Dolby type S Metal Tapes that are HD and goddamn amazing.

you CAN still get metal tape, and home recorders that can record and play dolby type S, but they're both goddamn expensive. (the cost of a single 60 minute metal cassette, adjusted for inflation, was close to $20!!!)

Vinyl does however win out over MP3.

>>1524

>Vinyl does however win out over MP3.

Bullshit. MP3 bitrates can, of course, exceed audibility to human ears. 32 bit information allows for a dynamic range far beyond any practical need. Because it is digital, it's not affected by environmental factors. You're buying into myths. The reason people love vinyl so much is because of our materialistic instincts telling us it's a possession. It's more tangible, and so it's more appealing to our hoarding instincts. Any "warmth" heard when playing audio from vinyl is essentially impurities, like throwing an analog warmth plugin onto all your music and calling it better, when it's simply farther from how it was intended.

>>1525

Im not >>1524

But mp3 is compressed file,maybe if you compare vinyl and digital music then you are probably correct.

>>1524

Straightening the question:How can i --distribute-- the best audio quality in general terms?I guess is digital music(WAV no compression) rendered with a good sound card.Correct?

Is there a pluginor something that "renders" like a vinyl sounds?

>>1526

The sound card doesn't play a role when rendering. Just bounce to lossless wav for mastering, whether you do it yourself or not. Then research the format of your medium by googling it, for example "cd bitrate". Always use the highest resampling quality, as that doesn't affect the file size.

>>1526

Also, if the final destination is vinyl, don't mix differently. You'll want the final wave/mp3 file to sound digitally like you want it to sound on vinyl. If you want a vinyl sound without the medium actually being vinyl, there are loads of tape and vinyl emulators out there. Google "vinyl emulation plugin."

>>1525

unless you are using 320kbps mp3's, vinyl is better than mp3 format (EQ issues not withstanding)

>Warmth

Warmth is a meaningless buzzword. In terms of Vinyl and guitar amps, "warmth" just means a full mid-range. you can literally send vinyl through a multiband EQ and lower the mid-range and wow look it loses the warmth. Oh lets take some MP3's and boos the mid-range, OH LOOK IT HAS WARMTH.

It is the most garbage term.

>>1526

>MP3 is a compressed file.

Digital File compression, not audio compression. Do not confuse these terms. ever.

>best format for file size.

Flac is lossless compression, OGG Vorbis is lossy compression, but still better than MP3, there are just fewer players.

>Rendering like vinyl

The audible difference is literally an EQ difference. Vinyl gets lower bass, higher mid, and low & lossy highs. This is because to make louder bass, the groove in the vinyl has to be wider. This is because the human ear detects different frequencies at different decibels, and we don't hear bass as well as mids and highs, which is why bass has to be so loud to be heard with clarity. If you tried putting dubstep on vinyl you'd get maybe 15 minutes. If you cut the bass out, you'd get 45-60 minutes. louder bass means wider grooves, wider grooves means fewer rotations per play.

In other words, you can simulate the sound of vinyl as I described above, by throwing a multiband EQ on, lowering the bass and increasing the mids.

>>1529

>Boos the mid-range

Boost*

>>1529

>best format for file size

And without the file size?It can carry gigas.

Anyone here know about SVG and its iteration with images?I dont know if something similar exist for sound,it could be interesting.

>>1524

>Bit rate is how quickly it moves between points

I take issue with that statement.

Bit rate is how many raw data bits are processed per second.

Say you have a mono wav file, 8kHz for the sake of simplicity, 1 second long, with each sample having 24 bit resolution.

There are 8000 data points in the file, each taking up 24 bits, so you have 1 x 24 x 8000 = 192000 bits per second.

Say you compress it losslessly, and you shrink it to a third of the original size. You get 64000 bits per second, despite both files having the exact same quality.

>>1522

>>can I say that bitrate=quality

The answer is 'depends'.

The only thing bitrate tells you is how much space some predetermined fragment of the file takes. Nothing more, nothing less.

>>1525

>MP3 bitrates can, of course, exceed audibility to human ears. 32 bit information allows for a dynamic range far beyond any practical need.

Not related. Also, mp3 does not allow for 32 bit sample resolution.

>>1531

>Anyone here know about SVG and its iteration with images?I dont know if something similar exist for sound,it could be interesting.

Yeah, it's called MIDI.

>>1533

I meant 16 bit, despite mp3 supporting 24 bit. No idea why I said 32. Anyway, it is relevant, because I was struggling to understand what the fuck you/he was on about, and so excluded dynamic range. When bit rates this high are indistinguishable from lower ones, vinyl has nothing at all going for it. It's basically a meme. It's a physical medium that's prone to alteration in ways digital information isn't, and in any practical situation, a vinyl disk capable of an equivalent of gigabits per second can still not be said to be of "higher quality" than mp3, even below 320 kbps.

>>1534

>prone to alteration in ways digital information isn't

Bit rot is a thing, thought it's not a problem that's prominent unless you have a big shitload of data you need to have stored pristine, and not nearly at the scale physical media are. You don't use exposed hdd platters.

>vinyl disk capable of an equivalent of gigabits per second

Are analog media considered continuous?

I know nothing really is - can't get two values from one molecule, but still - apples and oranges.

>>1533

>Yeah, it's called MIDI.

Exactly,but Midi is used for production and not distribution

In theory,if i have an midi playing "a song" i will have more detail quality(bitrate) than a streaming 320kbps or a CD 1400kbps?

Thats why I asked:

>does exist real diference(in sound) between pre-rendered(project) and pós-rendered(file)?

To understand better the questions i am:

>>1522

>>1526

>>1531

thanks for all the replies

>>1531

technically speaking sound on the computer IS vectorized. if you have 44100 samples per second, that's 44100 points at different amplitudes. that's vectors. I'm not 100% sure if the output is intended as steps or if it's smoothed before going to the speakers, but a speaker cone can only travel so fast, so the output IS always smoothed out to some degree. That basically is vectorized.

>>1533

say you have a mono file @ 8khz & 24-bit bit-depth. it takes the same space as an 8khz &12-bit bit-depth stereo file, but will have dramatically reduced quality. Plus 8Khz is spectacularly low, considering 44.1Khz is standard HQ and 22.05Khz is nyquist.

you litterally re-phrased exactly what I said, Each "sample" is bit depth, and samples per second is bit-rate. that's.... exactly what I said.

To simplify,

If you had 44.1khz bit-rate, and 1-bit per sample, you have 44100 bits per second, which basically means it's either max positive polarity or negative polarity with no zero crossing. which by the way, is really bad and would sound like ass. most bit-crushers stop at 2-bit bit depth, because it has max +, max - and 2 points of zero crossing.

>MP3 does not support 32-bit sample resolution

[citation needed]

in fact, no, Citation given,

https://en.wikipedia.org/wiki/MP3#Bit_rate

mp3 does support 32 bit bit-depth and 320kbps bit rate.

the problem with MP3 is mostly that it is lossy. very lossy in fact.

>>1535

the current best method of storing digital data in terms of bit-rot, oddly enough, is tape. many data servers use tape for data back ups TO THIS DAY, and is one of the more popular storage methods for archival purposes, because properly stored data-tape has a non-degrading shelf life of +70 years, assuming you don't fuck with it, or have to read/write more than a few times in that period.

>analogue bit-rates

the bit-rate of any analogue is limited by matter. In the case of vinyl it's based on only how fine the material can physically be shaped, and since vinyl is made by playing sound through a hot needle that is pressed in, and the sound and heat actually cut the vinyl which is turning at playback speed, the recording *SHOULD* be 1:1... except this gets very terrible low and high response due to the needle itself. more bass = wider grooves, and more high-end = deeper cuts. that detail aside the actual resolution is limited only by the molecular structure of vinyl. Theoretically if you measure distance traveled per second, and divide that into segments the size of a single molecule of vinyl (Polyvinyl chloride, or PVC) and you'll have your exact bit-rate. Measure the maximum vertical depth of the groove (at an angle if stereo) in that same increment and you have your bit-depth.

>>1537

>mp3 does support 32 bit bit-depth

And you literally posted link to bit RATES not DEPTHS.

That one depends on the other doesn't mean they're describing the same thing.

There's a difference, and I explained it in my post. Not my fault you can't read.

Actually, I don't think mp3 have a specified "bit depth" at all.

Now, I'm a little rusty on the exact format specification, but mp3 does not store data in the time domain, at least I don't think it does. It's separated into blocks, 512 or whatever samples each (less if there's a transient detected), each is transformed into frequency domain and stored that way - so, in an array of real and imaginary components of a fourier series. Decoding then iFFTs it into time domain, where we can start talking about discrete samples.

>you litterally re-phrased exactly what I said

Again, read again.

You claim that bit rate is how quickly it moves between points, but it's the sampling frequency that determines how long each sample is.

Each sample, in turn, can be made up of several bits, and this is what, indirectly, influences bit rate, but is not equivalent to it.

>>1536

SVG is a vector format - where to start lines, how bold they are, what the colour is, maybe the lines made up of several control points and joined so they make a circle, etc.

You can then rasterize it into any spatial resolution you desire.

Which is exactly what MIDI (or rather SMF - standard midi file) is - a description of what notes to play, how loud to play them, what timbre to use, etc.

Which you can then turn into audio at whatever fs and bit depth you want.

I suppose you could substitute it for any format that holds note data rather than samples, so .flp, .cpr, .rns, .als, whatever.

>MIDI is used for production and not distribution

Well...midis were all the rage back in the day.

I still have a shitload of floppies with SMFs on it, to be played back on a synth or a workstation.

>>1538

ctrl+F depth

"Compression efficiency of encoders is typically defined by the bit rate, because compression ratio depends on the bit depth and sampling rate of the input signal. Nevertheless, compression ratios are often published. They may use the Compact Disc (CD) parameters as references (44.1 kHz, 2 channels at 16 bits per channel or 2×16 bit)"

That's a 44100 bit rate, and 16 bit bit-depth per channel for CD Audio

https://en.wikipedia.org/wiki/Audio_bit_depth

"Bit depth is only meaningful in reference to a PCM digital signal. Non-PCM formats, such as lossy compression formats, do not have associated bit depths. For example, in MP3, quantization is performed on PCM samples that have been transformed into the frequency domain."

This means that MP3's are always variable bit-depth

Meanwhile back to the MP3 article linking to bit-rate

"Because of the Nyquist–Shannon sampling theorem, frequency reproduction is always strictly less than half of the sampling frequency, and imperfect filters requires a larger margin for error (noise level versus sharpness of filter), so 8 kHz sampling rate limits the maximum frequency to 4 kHz, while 48 kHz maximum sampling rate limits an MP3 to 24 kHz sound reproduction."

This means that an MP3 will always store enough detail to accurately recreate up to 24khz audio, Regardless of the bit-depth of that reproduction. So if you take 2 uncompressed wav's, one of them is a 24khz sine wave @ 24-bit depth, and the other, and a 25Khz sine wave @ 32-bit depth, merge them, the wav will be 32-bit bit-depth, then convert it to MP3, and it will reduce to 24-bit depth, because no sound at 24Khz or below requires more than 24 bit depth.

That's why it's lossy.

basically, the MP3 format is to sound, what Content Aware Scale is to images. still has a high bit-rate, and can have a very high bit depth, but something about it is just wrong.

>bit rate / bit depth semantics

at this point we're getting into being obnoxious pedants. both are "this is how quickly it goes through samples" vs "this is how big the samples are". that's what bit rate and bit depth are respectively. I never said that the bitrate/sample rate are 1:1 samples to sample rate. this is just semantics. for a fixed bit-depth, a higher bit rate will always play through faster, end of story.

>>1539

Midi doesn't dictate the sound of the audio. Midi is just control signals. in the audio to SVG analogy, Midi is the keyboard and mouse being automated.

In the case of virtually all audio, each sample point is a control point. and the sample rate dictates the curve.

>midi distribution back in the day

yeah, back in the days of DOOM and DOS games, before the Mp3 existed.

I'd like to point out that Diablo 2 was 3 disks on CD, but if you update the graphics from .bmp to .png, the audio from .wav to .mp3, and the video from .avi to .mp4, it would fit on 1 disk with tons of blank space.

the days when midi was for listening to, was the days before we had viable soundcards and audio file-compression.

I did the math. Remember Tape drives? the drives you could put a regular cassette tape in and store data? yeah. I did the math. if you encode the audio as MP3, you can store about 60 minutes of 128kbps audio on them, which Ironically is about the audio quality of a low-quality cassette for audio storage anyway. Not that the hardware that could read tape drives back in the day, had the processing power to play an MP3. which again, is the point.

Also, Midi will not sound the same on every computer, because midi playback was based on whatever instrument set was on your computer. Go ahead, play a midi in windows media player, then load it up in FL Studio with Fruity LSD. Then load the same one up on a mac. and then on linux. All of these will sound DRAMATICALLY different, even though it's the same file.

>>1540

Your post leads me to believe you don't really understand what you're talking about. In my opinion, you lack the basic understanding of how digital audio works, and try to quote various bits of info in order to appear eloquent - which is why some of the points you're trying to make appear true, but they're completely unrelated to the subject matter. Let me address those points as you made them.

Your first mistake is:

>That's a 44100 bit rate

Nope, that 44100 Hz sampling frequency (which has exactly the same impact on bit rate as bit depth, as long as we're talking uncompressed, but more on that later).

>https://en.wikipedia.org/wiki/Audio_bit_depth

>"Bit depth is only meaningful in reference to a PCM digital signal. Non-PCM formats, such as lossy compression formats, do not have associated bit depths. For example, in MP3, quantization is performed on PCM samples that have been transformed into the frequency domain."

Yeah, but what are you exactly trying to argue here? That bit rate is exactly sampling frequency, which is blatantly untrue according to what you've yourself posted about mp3? That distinghishing them is pedantism? How is it then that high bitrate mp3s still have a fraction of uncompressed data bitrates, while retaining the same sampling rate?

Also,

>PCM

That's just a way of storing a signal in digitised form. It's just the underlying basis of uncompressed digital audio, but what you're trying to do is to take the implications of its inner workings and apply it to ALL digital audio - like equating bit rate to sampling rate, which - while works in the case of PCM encoded data, doesn't when frequency-domain compression formats, like mp3, come into play.

>This means that MP3's are always variable bit-depth

You can't apply bit depth reasoning to mp3s, mainly because it's not how this format stores the data. I've already explained in my previous post that it's not made up of samples, but of blocks of spectral coefficients, decoded during playback on the fly via an inverse transform into time domain.

>Nyquist theorem from Wikipedia

What you've quoted is true, but you draw all the wrong conclusions, mistaking sampling frequency for bit rate yet again.

>MP3 will always store enough detail to accurately recreate up to 24khz audio

Except when you lower the target output bitrate, so that the the frequency-domain representation is less precise, thus lowering the upper limit of the highest representable frequency.

>That's why it's lossy.

It's lossy because it discards some of the data so that perfect reproduction of the input file is not possible. It literally 'loses' some of the information in order to minimize the space needed to store the data.

>basically, the MP3 format is to sound, what Content Aware Scale is to images. still has a high bit-rate, and can have a very high bit depth, but something about it is just wrong.

Nope.

mp3 format to sound is what jpeg format is to images. Both work on the exact same basis - of converting the signal into frequency domain and chopping the bits the encoder deems uninteresting off in order to store the input signal in "good enough" form.

Content Aware Scaling and audio? That's a first. Maybe some form of resampling would be a better comparison. I'm not even sure if "resampling" is the right term here, because it's not about recreating the signal at a different sampling rate. I guess "slicing" would fit? Some magic for of slicing, too, one that can identify which segments of the audio to remove and still retain legibility - like removing all the fragments where a specific instrument plays. Read up on how CAS works and you'll see I'm right - it removes some connected pixel paths from the image in order to reduce the dimensions.

>this is how quickly it goes through samples" vs "this is how big the samples are". that's what bit rate and bit depth are respectively

>I never said that the bitrate/sample rate are 1:1 samples to sample rate. this is just semantics.

Wrong again. It's what sampling frequency and bit depth are.

Let me put it straight:

Bitrate is how many bits of the file are read per second, as long as we're talking about the 'per second'-kind of bitrates. Nothing about it implicitly specifies the internal details of the file or the format used. Bitrate has to do with how fast the file needs to be streamed in order to be read and played back in a timely manner.

Sampling frequency is how many samples are there per second. It's measured in Hertz, which is why the period of time is question is exactly one second.

In case of uncompressed data, bitrate is a function of BOTH sampling frequency and sample resolution (bit depth). There's nothing pedantic about that statement, it's just how it is. In case of compressed audio, it's not so clear.

1/2

>>1541

I'm going to get hypothetical here, in order to prove a point. I did this before, but I'll do it again.

If you have a superbly efficient encoder that can store an hour of one channel music in lossless, pristine 96kHz 64bit double-word form in a file 2000 bits long, the bit rate will be - you guessed it - 2000/(60*60)=0.(5) bps, which is how many bits need to be read to reconstruct a second of the audio.

What would be the bitrate of the uncompressed PCM form of this signal?

Well - 1 (ch) * 96000 (kHz, the sampling rate = how many samples per second are there) * 64 (the resolution of a single sample) = 6144000 bps. See how, in this case, the bit rate depends on BOTH the sampling rate AND bit depth? And yet you treat them as completely unrelated.

Now onto the MIDI discussion:

>Midi doesn't dictate the sound of the audio. Midi is just control signals. in the audio to SVG analogy, Midi is the keyboard and mouse being automated.

Using your (incorrect) analogy, SVG is the brush path being automated. Unfortunately, that's not how it works.

What SVG stores is shape data. This shape data can be read by the rasterizer and converted into raster data - pixels.

What MIDI does is stores note data. This note data can be read by a synth and converted into audio.

>In the case of virtually all audio, each sample point is a control point. and the sample rate dictates the curve.

The sample rate does no such thing. Interpolation does, but that's beyond the scope for now.

>Also, Midi will not sound the same on every computer, because midi playback was based on whatever instrument set was on your computer. Go ahead, play a midi in windows media player, then load it up in FL Studio with Fruity LSD. Then load the same one up on a mac. and then on linux. All of these will sound DRAMATICALLY different, even though it's the same file.

MIDI will be exactly the same on every computer. It's not the format's fault you use different settings and a different software synthesizer on every system. It's like setting up a different "brush", and complaining the resulting stroke is different. Well, duh.

>I did the math. Remember Tape drives? the drives you could put a regular cassette tape in and store data? yeah. I did the math. if you encode the audio as MP3, you can store about 60 minutes of 128kbps audio on them, which Ironically is about the audio quality of a low-quality cassette for audio storage anyway. Not that the hardware that could read tape drives back in the day, had the processing power to play an MP3. which again, is the point.

Please do share your findings then. I want to see the formulas and explanations.

2/2

>>1542

lets say 96khz @ 64 bit-depth, that means there's 64 bits times 96000, every second, or 6,144,000 bits, or 768 kb, per second of 96khz 64 bit PCM. multiply for an hour, that's 2764800kb, or 2.7648 Gb. 2.7 gigs for an hour of 96khz 64 bit PCM? sounds about right to me.

now, find me a soundcard that can actually handle that at speed, and a sound system that can play it back, and with those requirements find one that's factory standard on a motherboard and under $50 for speakers.

I mean, they exist, but not for the casual listener. that description is audiophile level.

>midi / svg comparisons

SVG dictates shapes, by connecting points (interpolation)

Waves, conveniently are shapes. Square, Sine, Saw, etc. or constructed from harmonics, again, formed by connecting points. (interpolation)

yes, both PCM and SVG use point interpolation. that's just what it's called when you play connect the dots with a computer. the word is fancy, what's happening isn't.

in this analogy, MIDI, is just dictating where the squares and the sines and such are placed, and how long they are, but the actual position of the points can vary depending on the player. recorded audio, and even MP3, just store the points, no interpretation.

>Midi on different systems.

the midi file will be the same, the sound output is what is different. don't be so obtuse.

>tape drives and MP3's.

Okay, apparently when I did that math, I was getting the numbers for the storage capacity on tape drives for the wrong kind of tape. I had gone to the "tape drive" wikipedia page, skipped to the 80's looked at the consumer models and saw 100mb to 400mb as the range

yes, you can fit about 1 hour of audio in 100mb with MP3. emphasis on "can" and not "will", it is possible, not guaranteed. the audio content has some dictation on file size. particular in regards to frequency range and headroom. the ozzy will compress file-wise better than say, skrillex.

However, looking now, and double and tripple checking, it appears I was in fact wrong, SPECIFICALLY AUDIO CASSETTE, the Commodore Datasette used standard audio cassettes, and each 1 hour tape (30 minutes per side) had a maximum total capacity of 1978 kB

Now, I will say this. I managed to compress the song "we used to be friends" by "the dandy warhols" to be ~about~ 100kb. granted it was at 8khz and mono at that point, which is comical, it was a 5 minute song compressed to ~100kb. it sounded like it was being played over a telephone. I did this by abusing the export options in audacity for the LAME codec.

but that's still 5 minutes of audio / 100kb, with 1978kb of storage, or about 95 minutes of audio (that sounds like it's being played over a land-line)

I never said it would sound GOOD, just that it would function.

also, digging up the file for proof... It appears that something went wrong, and at a certain point in the track.... well, you'll see.

http://puu.sh/n7fHn/ef97ebd547.mp3

I tried doing it again, but the 2nd attempt at re-doing this NOW.... well it ended up being 600Kb, which means I did something differently. Not sure what.

Lastly, Sampling Frequency, and Bit-rate are inseparable, it has to do with the Nyquist limit. it's pedantic because guess what, both dictate how rapidly you're reading through a file. full stop. the difference is what they're measured with. the difference is between saying 1 mile or 1.6Km. different measurements for the same result.

>content aware scale for sound

what content aware scale does, is scoops out information that it thinks is unnecessary and throws it out the window, and just fills in the space with whatever was around it.

what MP3 does, is it scoops out information that it thinks is unnecessary and throws it out the window, and fills in the (frequency) space with, basically aliasing, or doesn't fill it at all.

In other news, that gave me an idea.

I exported a spectragram of a vocal sample, Content aware scale'd the spectragram, and then turned that back into sound. Here's the result of that.

http://puu.sh/n6U8H/3e90c21bad.mp3 (done with harmor, then GIMP, then back into harmor)

>>1543

Edit

I can't edit.

I copyed my post to notepad part way through because i got the numbers wrong, put it in notepad, did the math better. here's the corrections.

>Now, I will say this. I managed to compress the song "we used to be friends" by "the dandy warhols" to be 268kb. granted it was at 8kbps and mono at that point, which is comical, it was a 4:30 minute song compressed toless than 300kb. it sounded like it was being played over a telephone. I did this by abusing the export options in audacity for the LAME codec.

but that's still 4:33 minutes of audio / 268kb, with 1978kb of storage, or about about 33 minutes of audio stored as data. If I could compress it to about 6kbps, or change the bit depth a bit, it would fit over an hour easy. as proof.

So there's the actual correction that I FORGOT TO RE-PASTE INTO THE REPLY BOX

>>1543

>now, find me a soundcard that can actually handle that at speed, and a sound system that can play it back, and with those requirements find one that's factory standard on a motherboard and under $50 for speakers.

Now what in the hell do speakers and soundcards have to do with how fast the file needs to be read?

Would it shock you to learn that when you watch high-quality HD movie you have to push several MEGABYTES of data per second? Do you need a theater-phile display that can handle just the reading the file from the hdd?

Just from this first bit I have serious doubts you'll listen to anything I have to say and not throw unrelated technobabble to "prove" how wrong I am.

Also, 6144000 bits = 768kB, not kb. What hard drive would have trouble reading that? Because they probably went out of use in the 80s.

>I mean, they exist, but not for the casual listener. that description is audiophile level.

First of all, I was being hypothetical in order to illustrate a point, second of all - audiophile-level description? Really?

>SVG dictates shapes, by connecting points (interpolation)

>constructed from harmonics, again, formed by connecting points. (interpolation)

Your analogy is completely wrong, because I've already explained that SVG is a format that describes what to display in abstract terms (red circle in the middle, line from top to bottom, etc), not the pixel-per-pixel result of rasterization. A digitized waveform doesn't describe what to play in abstract terms (C3, horns, 8th measure, this and that melody), but stores the recording of it sample-by-sample.

Think about it. Isn't that exactly how plain old image bitmaps store data?

>in this analogy, MIDI, is just dictating where the squares and the sines and such are placed, and how long they are, but the actual position of the points can vary depending on the player.

>the midi file will be the same, the sound output is what is different.

Just how the actual rasterized (pixel) output of the rasterization process can vary from rasterizer to rasterizer. Just because most things probably use the same library (libsvg, most likely), doesn't mean there aren't other that can give different result. For instance, there's NanoSVG, which has a rasterizer incapable of rendering all the intricacies of the format and outputting flat-filled shapes only.

>what content aware scale does, is scoops out information that it thinks is unnecessary and throws it out the window, and just fills in the space with whatever was around it.

>what MP3 does, is it scoops out information that it thinks is unnecessary and throws it out the window, and fills in the (frequency) space with, basically aliasing, or doesn't fill it at all.

Yeah, two things.

1) CAS is not a compression format.

2) CAS operates in spatial domain, by literally cutting out from the image. It would be analogous to taking a knife to the waveform and cutting out samples from it. Yeah, the resulting file would be smaller, but so would be the left half of the image only.

Now, JPG does EXACTLY what MP3 does - chops the file into blocks, transforms them into frequency space, stores them in a not-quite-exact way.

>Lastly, Sampling Frequency, and Bit-rate are inseparable,

Oh for fuck's sake. Yes, in uncompressed data they're related. RELATED. In compressed, bit-rate tells you fuck all about the sampling frequency.

>it has to do with the Nyquist limit

No, that's just plain wrong. The Nyquist limit is telling you what the highest frequency can be, at a specified sample rate.

You know what? I'm not going to repeat myself again. Please refer to my multiple explanations in previous posts.

>tape drives and MP3s

And that somehow proves what, that the quality of audio is better on mp3s because you store data in a format that's unsuitable for an analog medium? Well, big surprise.

>>1545

>what do soundcards and speakers have to do with sound

check your speakers. they have a maximun frequency response rate, which means playing higher quality audio than that, will not make a difference. Most of them cap at 44.1khz, which means 96khz is wasted on them.

as for sound-cards, it has to have the processing power to parse through the data at that rate. why do you think buffer underruns happen?

>Being so pedantic as to distinguish kB/kb/Kb.

seven hundred and sixty eight kilobytes.

>your SVG description

SVG data saves individual points in a list, and what points they connect to, with what tension, and what color, and what areas are filled, from where, and in what order. This is true for all Vector Graphics, (as opposed to Raster graphics, which store pixels like BMP and PNG, or clusters of pixels like JPG or GIF)

>compressed bit-rate

if your compressed file has a maximum bitrate, you will have a maximum sampling frequency, especially if your compressed file has a fixed bitrate rather than variable.

if your file has a maximum sampling frequency of 22.05khz due to it's bitrate, and the compressed file has the same bitrate, the compressed file will still have a maximum of 22.05khz, and not have the data to go above that.

yes. it does have to do with the nyquist limit. with a set sample rate, you have a maximum frequency, therefor you can't have a higher frequency than the sample rate supports.

>tape drives and MP3's.

No. it's not supposed to prove anything about quality. the tangent about storing MP3's on a cassette drive, was just a random non sequitur that was related but not important.

>>1546

>check your speakers. they have a maximun frequency response rate

And what does it have to do with how fast the file is read?

Moving the goalposts yet again?

>Most of them cap at 44.1khz

That would be 22kHz.

>as for sound-cards, it has to have the processing power to parse through the data at that rate. why do you think buffer underruns happen?

Because the rest of the PC can't handle filling in the buffer (the contents of which are moved by the sc to the DAC) in a timely manner.

>>Being so pedantic as to distinguish kB/kb/Kb.

It's kind of important and if you don't think so, maybe you shouldn't talk about digital audio (or digital anything) until you learn the difference and use proper fucking terminology. 'k' is always lowercase, too.

>SVG data saves individual points in a list, and what points they connect to, with what tension, and what color, and what areas are filled, from where, and in what order.

No shit? Besides, it's a tree, not a list, because it's XML markup.

I think you missed the point of my little comparison. The aim was to show that vector formats don't store pixel data (at least not generally, some allow for embedding, but then the content isn't vectorial per se) in an analogous manner to pcm sample data. They abstract it away into points and shapes made of them, just like SMFs abstract precise sample information to something that is more musically meaningful than sample values.

>if your compressed file has a maximum bitrate, you will have a maximum sampling frequency

Bull shit.

Don't believe me? I can encode 8kHz to 320kbps mp3 no problem. I also can encode 44.1kHz to 320kbps mp3 no problem. And 8kHz to 32kbps, and 44.1kHz to 32kbps, and whatever combination I want.

Besides you acknowledge variable bit rates, which literally prove your point invalid. Do you think the sampling rate changes when a segment of music needed less bits to be encoded?

>yes. it does have to do with the nyquist limit. with a set sample rate, you have a maximum frequency, therefor you can't have a higher frequency than the sample rate supports.

Which has fuck-all to do with bitrates.

>>1547

>frequency response

if your speaker has a max of 22.05khz, but the audio sent to the speaker is 44.1khz, the output will only be 22.05khz.

that's not moving goalposts

>buffer underuns

there are 2 reasons

1: can't fill buffer

2: buffer over-filled (buffer overflow)

if your buffer size is smaller than what the bitrate/bit depth require, guess what happens? it won't play correctly.

>kb/kB

name one reason you would ever use kilobits instead of kilobytes, other than hardware manufacturers wanting to short-change customers (megabits and gigabits instead of megabytes and gigabytes listed on thumbnail drives)

>svg

I never said that SVG stored PIXELS, I said points, as in points on a plane. as in an X coordinate and a Y coordinate. That's why I expressly drew the distinction between Vector and Raster graphics. SVG is vector graphics, and ALL vector graphics store points, while all Rasters store pixels or pixel clusters, even JPG which stores as color-spaces, it's actually storing layers of 8x8 pixel cells that add together into 3 layers of color (y cb cr)

that said, just because it's stored in tree, not list, because of XML is irrelevant. the reading of the file is the appearance of points, lines, and color, in a sequence, in the order in which they are read.

At this point we've lost the original point.

The original question was post 1531

>Anyone here know about SVG and its iteration with images?I dont know if something similar exist for sound,it could be interesting.

SVG Is a vector graphic format, and all vector graphics go point by point and connect them with tension, and add color, etc etc, as a method of storing large images in small space

I highly doubt the poster was trying to ask

>is there a way to store audio as a data tree rather than a linear sequence

because I'm pretty sure that'd be closer to just

>lol that's how DAW's basically save project files

>compressed file has a maximum sampling frequency

yes, you can encode 8khz into 320kbps

can you encode a 44.1khz 320kbps file down into 8khz with an incredibly low bitrate?

FUCK NO

Not without MASSIVE UNRECOVERABLE LOSS OF DATA

It's called "Downsampling"

Take any song that's 44.1 @ 320kbps, and save a duplicate in 8khz 128kbps. then take that 8khz 128kbps and save it as 44.1 @ 320kbps, and compare with the original file, both playback and in spectrum view.

render audio to 32-bit uncompressed wav, then save a duplicate as an mp3, and compare the spectrums.

it makes a massive difference that you can drive a fucking truck through

>bit/sample rate

lets say you have a bit-rate of 8 bits per second and a bit depth of 8 bits. that means you're loading 1 sample per second. guess what? the highest frequency you can get is 0.5 hz, due to nyquist. therefor your sample rate, bit-rate, bit-depth, are ALL FUCKING INTER-CONNECTED.

It's like the motherfucking fire triangle. you have a ratio, Oxygen, Fuel, Heat. Will you have a fire,

in this case

Bit rate, Bit depth, sample rate, Is this combination valid?

This feels like it's just a failure to understand basic mathematical logic.

>>1531

Like a synthesizer?

>>1550

>that's not moving goalposts

But of course it is, because the original point was about bitrates and filesizes per second of audio, while you're making it about audio reproduction, and Nyquist-Shannon theorem, and how your speakers have any say about how fast the rest of the PC can read files.

>1: can't fill buffer

Uh, yeah, that's what I said?

>2: buffer overflow

That's a security violation and would warrant a crash, most likely.

>if your buffer size is smaller than what the bitrate/bit depth require, guess what happens? it won't play correctly.

First of all, as this is coming out of the underrun discussion, DAW buffers are bit-depth independent (most daws calculate samples at double-precision anyway), and sample-rate independent (2048 samples at 44100Hz give you greater latency than at 88200Hz).

Second of all, buffers are used as a safeguard in case the CPU can't provide samples fast enough - with 2048 samples of buffer, your CPU can lag 2048 samples before you get a drop-out.

So no, they have nothing to do with bitrate (as if they would, anyway), bit depth, or samplerate. They have everything to do with how fast your CPU is, though.

>name one reason you would ever use kilobits instead of kilobytes

Uh, bitrates? What do you think 'kbps' stands for?

>I never said that SVG stored PIXELS

And I never said you did, I just imply you can't comprehend the simplest of analogies.

Vector graphic formats store "instructions" for the rasterizer, because line endpoints and control points and stuff are just that.

Raster graphic formats store rasterized data (discrete pixel values).

Standard Midi Files store "instructions" for synths, because note ons/offs, velocities and stuff are just that.

PCM audio files store discrete sample data.

How is this so hard to comprehend?

All of those are stored as numbers, because it's digital, but that doesn't make the interpretation of those numbers the same.

>even JPG which stores as color-spaces, it's actually storing layers of 8x8 pixel cells that add together into 3 layers of color (y cb cr)

Technically, it's storing Huffmann-encoded, quantized coefficients of DCT base functions.

>can you encode a 44.1khz 320kbps file down into 8khz with an incredibly low bitrate?

>FUCK NO

Fuck yes, there's no technical reason why that wouldn't work. ffmpeg -i 441khz-320kbps.mp3 -af aresample=resampler=soxr -ar 8000 -c:v libmp3lame -b:v 32k audio-8khz-32kbps.mp3 should do the trick.

>Not without MASSIVE UNRECOVERABLE LOSS OF DATA

>It's called "Downsampling"

Congratulations, you discovered why you can encode 44.1kHz 320kbps file down into 8kHz one with a smaller bitrate. And proved that by specifically wanting to downsample you, uh...downsampled?

>Take any song that's 44.1 @ 320kbps, and save a duplicate in 8khz 128kbps. then take that 8khz 128kbps and save it as 44.1 @ 320kbps, and compare with the original file, both playback and in spectrum view. render audio to 32-bit uncompressed wav, then save a duplicate as an mp3, and compare the spectrums.

And that is supposed to demonstrate what, that a file with a sample rate of 44100 kHz doesn't have to have high-frequency compoments? Well, shucks, happens. It doesn't fucking matter whether there's any content in the file, you can have 96000 samples comprising a second silence, you still have a file with a sampling rate of 96kHz. Stop bringing Nyquist into this.

>lets say you have a bit-rate of 8 bits per second and a bit depth of 8 bits. that means you're loading 1 sample per second. guess what? the highest frequency you can get is 0.5 hz, due to nyquist.

You must be retarded, because it's the third time I have to tell you that WHAT IF THE FUCKING FILE IS COMPRESSED?! What if 8 bits can store more than one sample? What if there's a hundred samples in those 8 bits?

>therefor your sample rate, bit-rate, bit-depth, are ALL FUCKING INTER-CONNECTED

You bring up uncompressed data and I feel like I'm being parroted my own words back at me.

>This feels like it's just a failure to understand basic mathematical logic.

On your part, maybe.

I shan't bother responding to you any more.

>>1552

>"moving goalposts"

Speakers: there's no point to having higher quality audio than what your speakers can re-create

soundcard: there's no point to having higher quality audio than your soundcard can recreate

Nyquist and bitrate

There are, for the umpteenth time, combinations of bit-rate, and bit-depth, that render certain frequency ranges impossible, therefor frequency range is directly linked. 44.1khz audio sets a demand for a minimum bitrate and bit-depth

>buffer overflow

Except buffer overflows don't always result in a crash, that's how the effects of buffer overflows can be documented to have set results, and how exploits exist that take advantage of buffer overflows. This is general about what "buffer overflow" is,

For an example, within FL Studio, you can get what FL calls an underrun, and it has to do with something to do with what it calls oversampling. What is oversampling? when there is more data to process than there is processing ability. what is an underrun? an error with the buffer size. if your computer is trying to generate a sound that is more complex than your CPU can handle, you get a buffer underrun, despite the cause being too much, not too little, to process. in FL, this isn't an issue for rendering and recording, because it adds a buffer time when there is an underrun, to the recording. the playback will be wrong, but the recording will include the latency compensation. it's really kinda fancy.

So, if you have an audio format that has such a high bitrate and bit depth, that the computer is incapable of reading it at 1:1 speed, you get errors. this can be an underrun, because the buffer can't be filled because the read speed is to slow, but the opposite can also be true, the read speed can be faster, and output more data than the buffer can handle for playback. a buffer overflow.

So yes, the hardware used for playback does matter to some degree. this is really only relevant today when you get into the highest quality 96khz high bit-rate high-bit-depth files, regardless of compression methods, if any. However in the past, it was much more relevant.

>kbps

.... I just realized that mp3 uses kilobits per second, and not kilobytes per second. That said, that means storing MP3's on tape is a factor of 8 more dense on tape than what I had originally claimed. rather than 30 minutes on a 60 minute tape, it would be 240 minutes, on a 60 minute tape. at a maximum, at that lower bitrate. because the storage on tape is in kilobytes, not kilobits. I arrived at that number in kilobytes, from the storage listed in quantity of bytes.

>svg, pixels, pcm, and blah blah blah

SVG and PCM, Points in space, connected over time

in the case of midi, you can read a midi file on a system that is incapable of audio playback. note instructions are not sound instructions.

In this analogy, we could only accept midi as a comparison to SVG, if we consider playback to be compared to using a switchable color-pallets for an SVG, before you read the file.

Also, Vector graphics didn't always need rasterizing, after all, Vector graphics started on CRT's, before we even HAD pixels, before raster graphics where a thing which was even possible, just like how the rendering of audio as PCM was possible, before midi-data was a thing. this has now gone full circle.

>Jpg

it's a grid of 8x8 pixels, and a list of 8x8 pixel gradients that are layered together addatively. the quality of the jpg is how many of these 8x8 gradients get added to create a complex 8x8 pixel pattern. there are 3 images layered together. yellow, chrominance, and luminance. (Y, Cb, Cr) which is a small step away from CMYK (Cyan Magenta Yellow blacK)

Huffmann-encoded, quantized coefficients of DCT yadda yadda yadda, is just the way in which it arrives at that storage method.

saying that it's huffmann-encoded blah blah blah, is like describing pi and the ratio of a circle, to describe how you drive a car to burger king, because you want to describe the way in which wheels function to dictate how cars work.

I think we can let this bit die now.

>can you encode

you stopped reading and skipped

I wasn't saying "can you downsample, No, here's how you downsample"

I was saying "after you have already downsampled, can you restore the original audio FROM the downsampled version", and the answer to THAT was fuck no.

>8bps 8bit depth

if the audio is compressed, it still has to be decompressed into PCM to stream. So assume it is 8bps and 8-bit depth, post de-compression. Or otherwise, Assume that it MUST be an uncompressed file in this case.

you're ignoring the point I made and attacking a theoretical minutia. you're using postmodernist critical theory (which throws facts and logical observation out the window, in favor of winning a debate by wearing your opponent down), and it's fucking disgusting.

>>1552

a little bit more on

>can you downsample and then upsample.

what I was saying in that part is identical to

"Can you save a high quality PNG as a low quality JPG, delete your high quality PNG, then take your low quality JPG and re-create your high-quality PNG"

To which the answer is no. no you fucking can't.

Downsampling to low bitrate low bit depth, low frequency range, is identical to this.

Once you lower the bitrate, you cannot raise it back up. doing so is like scaling an image down and then back up. all you do is end up with duplicated samples.

you can't lower the bit depth and then raise it again for exactly the same reason.

and as for frequency range, that's like setting it from 32 bit color, down to 256 color pallet, or even greyscale or monochrome, and then trying to restore it to 32 bit color. all that information is just fucking gone.

>>1553

Oh fuck it, I'm bored anyway.

>Speakers: there's no point to having higher quality audio than what your speakers can re-create

>soundcard: there's no point to having higher quality audio than your soundcard can recreate

>"I'm not moving goalposts"

>proceeds to move the goalposts

Okay then.

>44.1khz audio sets a demand for a minimum bitrate and bit-depth

No.

Did you even read and try to understand my posts?

>What is oversampling? when there is more data to process than there is processing ability. what is an underrun? an error with the buffer size.

No.

What you even on about in that paragraph? Because you throw terms at random again.

>SVG and PCM, Points in space, connected over time

>PCM and space

>SVG and time

>In this analogy, we could only accept midi as a comparison to SVG, if we consider playback to be compared to using a switchable color-pallets for an SVG, before you read the file.

What the fug are you talking about?

Oh, right. You're showing you have absolutely zero understanding about anything you try to explain here. Replace "color-palettes" with "stroke styles" and you have an SVG file. Analogous to "synth patch numbers" and "MIDI".

>yellow, chrominance, and luminance. (Y, Cb, Cr)

It's luma (Y), blue-difference (Cb) and red-difference (Cr).

>I was saying "after you have already downsampled, can you restore the original audio FROM the downsampled version", and the answer to THAT was fuck no.

No, that was your conclusion. What you were trying to argue is that if you have a compressed file of poor quality, you can magically determine the sampling rate of it after looking at the spectrum.

>if the audio is compressed, it still has to be decompressed into PCM to stream. So assume it is 8bps and 8-bit depth, post de-compression.

Except the number of bits read from the file and processed stays the same.

>Or otherwise, Assume that it MUST be an uncompressed file in this case

That's a special edge case, which you're extending to all files.

>you're ignoring the point I made and attacking a theoretical minutia.

And I'll continue to do so as long as you don't understand those theoretical minutia are not so minor and have a connection to what you're wrong about.

Oh, and as long as you'll continue to spout bullshit.

>you're using postmodernist critical theory

And you're throwing as much shit as possible in order to hide your lack of knowledge.

I've already explained to you, multiple times and with examples, why you're in the wrong. The information is all there - you just don't want to learn.

>>1554

>"Can you save a high quality PNG as a low quality JPG, delete your high quality PNG, then take your low quality JPG and re-create your high-quality PNG"

To which the answer is no. no you fucking can't.

You'll have your low-quality PNG though. It's all about HOW data is stored, now WHAT data is stored.

I understand what you're trying to show here, but your lack of understanding that you indeed CAN store low-quality information in a high-quality container and there's nothing preventing you from doing it is why this is a shitty example when it comes to bitrates and samplerates.

>that's like setting it from 32 bit color, down to 256 color pallet, or even greyscale or monochrome, and then trying to restore it to 32 bit color. all that information is just fucking gone.

So? It's not like monochrome can't be represented in 32bit colorspace. Same as above - not WHAT, but HOW.

>>1555

>44.1khz and a bitrate and bit depth demand

if you are only getting 1 update per second, and only 1 bit of bit depth, you can only output a 0.5 hz triangle wave, or a an incredibly saturated 0.25 wave. you could not generate an audible tone at that bit rate and bit depth.

those numbers are based on the actual sample points, and not any method of data compression.

This is the thing you're glossing over by just saying "but you're ignoring compression", and that's because i'm looking at the final product, not the stored file.

>oversampling

Just because you don't understand it, doesn't mean it's random terms. just drop that subject at this point, we won't get anywhere with it, because it was a tangent relevent to the above point,

>SVG / midi comparison

with an SVG, you don't get huge arbitrary differences in rendering because of the system you render on. with midi, you do, and also with midi, you have very little control over this unless you're loading it in a DAW, rather than a player. If you're loading it in a DAW, it's not a midi anymore, and you're not working exclusively with midi data. for instance, control of individual samples, or synths, is not part of midi files, but is dramatically more important to the output/sound, than the midi itself.

Technically an analog drum sequencer with purely analog synth drums, would be closer to an SVG than even midi, but that's a whole other thing.

>JPGs

you're partially right,

Y is Luminance,

Cb is Chrominance (blue)

Cr is Chrominance (red)

https://www.youtube.com/watch?v=LFXN9PiOGtY

https://www.youtube.com/watch?v=n_uNPbdenRs

https://www.youtube.com/watch?v=Q2aEzeMDHMA

this is the type of shit that I watch in my spare time for entertainment. just let that sink in for a moment. in this case I have only miss-remembered a few terms.

>downsampling

saying that you can't up-sample was not my conclusion, it was a statement to construct further argument. we where specifically talking about MP3's. from wave to mp3 is automatically downsampling, with any compression method. MP3's are also Lossy, just like JPGs, which means the higher the compression rate, the more file quality is lost, again, just like JPGs. Comparing images on a spectrum, is simply the visualization of this data loss.

Seriously. Do this test.

Open any DAW of your choice, and render a project as an uncompressed wav, and as a 320kbps mp3. Also, make a separate version of the wav that downsamples to a very low audio quality.

now, according to your logic, they should all be the same output, because bitrate and bit depth don't effect frequency range. but MP3 gets a HUGE scoop out of the high end, you can see this clear as day in a spectrum view. as for audio quality loss outside of that range, you can phase cancel them. If the audio was the same, you would get perfect silence.... but you don't. And then you get the lower sample rate wave, that one is a nightmare for this comparison.

Don't JUST sit here and talk.

actually fucking do the test.

>compression of audio

Here is what happens with MP3

you have a compressed file

you open the MP3

the file gets decompressed

the decompressed data is read as PCM

the bit-rate is based on the decompressed form in memory, not the compressed form on disk.

>special edge case

This so-called edge-case is for uncompressed data, which all compressed data is built upon. that is not an edge case, it is the opposite of an edge case.

That said, stop trying to make excuses to ignore this point.

Assume 8-bits per second and 8 bits of depth, with an uncompressed file, what is the frequency range of your playback? is it 44.1 khz? fuck no. is it 22.05Khz? fuck no. Is it anything within the audible spectrum? fuck no, it's goddamn morse code. it would just be different volume pops once per second.

Therefor BASED ON THIS FACT, The conclusion is that, in uncompressed files, and therefor in lossless files, your bit rate and bit depth, have a DIRECT CORRELATION WITH FREQUENCY RANGE.

This is not an edge-case, this is the fundamentals of the uncompressed audio, upon which all compressed audio is built.

>low quality information in a high quality container

That wasn't the point I was making.

The point I was making is

YOU CAN'T STORE HIGH QUALITY INFORMATION IN A LOW QUALITY CONTAINER

>colorspace

yeah, here's the thing. In this analogy of JPGs and color spaces, and converting into monochrome... The color detail is your frequency detail and actual discernible sound quality.

>>1557

Look, this is getting so far removed from the original point that I don't even want to continue arguing. You make mistakes in your post, but I won't even bother pointing them out. We've shitted up this thread enough.

Is anyone familiar with making drones and pads? Like Katahiminan on SC. Any tips?

>>1559

Lots of things you can do to get that drone sound.

Generally the secret is having a clean base sound (guitar samples, grand pianos, some digital pads). Play some nice chords or arpeggios, sample them and load them up in your sampler, then play those. Drown them in a ton of reverb. Use them with granular synths and tweak some knobs. Load up Edison and use the blur tool.

If all else fails, download paulstretch.

>>1560

That's what I've been doing. I've been finding samples or instruments in the same frequency range as I want the pads and drones to be, then use Fruity Granulizer, which I record into Edison, then blur and loop. I've had some good results, but none have been close to Katahimikan (listen to Promise or Halcyon). I suspect he's using pad plugins, like Evolve Mutations or whatever. I'll check out Paulstretch, thanks.

https://clyp.it/gnu2r0uo

What could be this duplicated sound?I desactivated all delays and attack is ok.

>>1574

that's somewhere between a saw and a triangle wave, that's been low-passed, and has a chorus/unison applied to it.

I did the same thing trying to re-create the Playstation 1 startup sound. http://puu.sh/mAZ3q/d507589a7e.mp3

>>1575

Its not the sound itself,probably you got it wrong,because when i listened on speakers out I cound notice.On the headphone i can listen to it,it happens in exactly picks on wave drawing(1sec;4sec).

Nice start up,I through about recreating something similar too,its sounds very modern today,in my opinion.

>>1576

as far as I can tell, those pops are either

A) volume automation

B) it's 2 different recordings layered on top of eachother, and that point is where the first note starts on the second layer.

in both cases that's still definately a triange or saw that's been lowpassed and had chorus or unison applied. you can tell that much by looking at the waveform and the spectrum.

the characteristics of being a saw or a triangle is in the harmonics present when you view the spectrum.

because of the chorus/unison effect (not a phaser) is in the change in shape of the wave over time, and the phase cancelation which causes the volume to fade in and out.

you can see that at the start of the loop the waveform is mostly in-phase and lined up so the wave looks uniform, but as time progresses to around 1-2 seconds, it is clearly multiple stacked waves canceling eachother out.

additionally the shape of the wave looks like a triangle that was skewed towards being a saw, but didn't quite make it over all the way, which is something that happens when you low-pass a saw wave, or you just skew the shape of the wave (in something like sytrus).

additionally while looking at the spectrum, it has no higher frequency data, further evidence of the lowpass.

If you're talking about the sudden snapping that happens for each note, I think that's probably either volume automation or a second layer of audio as I mentioned earlier in this post. Or maybe it's still another delay at about a 1 second delay.

Either way, I tested recreating the sound based on these observations and got fairly close within about a minute of fiddling. It's a fairly basic sound in and of itself.