making samples sound of higher quality

[mickomoo]

I'm using Kontakt 4's default samples and when I record tracks with some of them, they sound like a 16 bit recording. I can't tell if it's my speakers, or my daw or the sample rate. How do I improve the tone of these vsts?

[dannthr]

I'm using Kontakt 4's default samples and when I record tracks with some of them, they sound like a 16 bit recording. I can't tell if it's my speakers, or my daw or the sample rate. How do I improve the tone of these vsts?

They probably are 16-bit recordings.

The DAW, Speakers, and Sample Rate have nothing to do with your samples sounding like they were recorded, perhaps your speakers to a degree may have a lower dynamic range, it might make 24-bit and 16-bit samples difficult to distinguish, but generally, it probably is 16-bit and if it sounds like it, then everything is working okay.

16-bit is not good enough? I mean, 16-bit is basically required for CDs or iTunes mp3 distributions, but if you're looking for higher quality samples, well, then the bit-depth has to be recorded in from the source. Increasing the bit-depth will not do anything to the audio, only decreasing the bit-depth will have an impact in the dynamic range and quality of your sounds.

So there is no way to "improve" the sound.

You can get 24-bit samples, they are becoming more and more common now that computers have really made great leaps in the last year and a half or so with regards to RAM.

[Rozovian]

Do you really mean "like a 16-bit recording" and not "a recording from a 16-bit console"? There's quite a difference. The latter would be solved with better humanization and/or mixing.

[mickomoo]

I think the latter. My father pointed out that the sound simply sounded "16 bit" and that his Daw, pro tools, allows you to record at multiple bit depths. He was talking specifically about this recording:

I think he was specifically refering to the saxophone. I'm aware that it needs to be mixed but he pointed out to me that even without mixing the quality of kontakt's samples should sound better. Which some of them do I think. However the trumpets, saxes, and strings are the only samples that are really hard to make sound decent for some reason or other.

[Fishy]

I dunno what your Father is trying to tell you but 16-bit is professional quality of bit-depth. Every regular CD in every CD store is 16-bit. They might record albums at 24 but it's always released at 16.

The quality of the original sound that was sampled and the way you use it is the only thing that should affect how 'good' it sounds. The last thing I would complain about for the sample you just posted is it's bit-depth. The problem is it sounds pretty robotic. The dynamics and timing are a bit everywhere also.

[dannthr]

Yeah. 16-bit is the Red Book standard by which all audio CDs are burnt at... Bit depth is about smooth dynamic range. Increasing your bit-depth increases your dynamic range potential by about 6 dB. Poor bit-depth results in sound that seems like it was played over the radio or has a crunchy quality. 1-Bit Depth sound basically means you either hear sound or you don't.

Your samples are actually fine, they don't sound crunchy, but they do sound like they were performed by circus robots, and you have to appreciate the fact that as someone interested in creating a smooth performance from samples, you are the performer--you are the band, the conductor, recordist and mixer--so this falls on you.

You can't just plug your samples into your music and expect them to sound like a band, that's not how it works, that's not how people work in real life either.

You need to coax a performance out of your samples by assembling phrases that have a beginning, middle, and end. By LISTENING to other performers, to recordings, and mimicking those performances as best as you can.

Finally, your music is weird--awkward even--and that results in a weird and awkward sound that masks any built in effort on the part of the sample libraries themselves to create a natural sounding performance.

[eilios]

Finally, your music is weird--awkward even--and that results in a weird and awkward sound that masks any built in effort on the part of the sample libraries themselves to create a natural sounding performance.

This. I wasn't really having any major problems with the sounds(it all sounded somewhat okay), but they seem really badly matched together and odd.

[mickomoo]

I'm aware that I need to mix and humanize, but for some reason my dad was saying that in pro tools, premixing the depth of the library sounded better, idk I have to ask him again

[dannthr]

I'm aware that I need to mix and humanize, but for some reason my dad was saying that in pro tools, premixing the depth of the library sounded better, idk I have to ask him again

It does.

Working in a higher bit-rate in your projects gives you a greater dynamic range within which to work and that can be extremely useful for creating deep and dynamic sounding mixes--even if the ultimate destination bit-depth will be 16-bit.

But what is standing between you and a professional is your skill at instrument programming, not your technical audio quality.

[mickomoo]

It does.

Working in a higher bit-rate in your projects gives you a greater dynamic range within which to work and that can be extremely useful for creating deep and dynamic sounding mixes--even if the ultimate destination bit-depth will be 16-bit.

But what is standing between you and a professional is your skill at instrument programming, not your technical audio quality.

I know that I need to mix, and also edit the notes of each instrument, but I guess what I didn't get was that you could change the bit rate of the project. My Daw doesn't do that. It only changes bit rate during mixing I believe.

[Rozovian]

Dude, unless you have some really weird settings, the bit rate isn't among the issues with your music. Can't tell if the track you posted doesn't sound like it's suffering from any bit depth related problems, not something that's evident to me at least. It does sound like something out of a snes game tho, with fairly raw and not humanized instruments.

I don't know what your dad is on about. Either he has good ears and is working with material where you can tell the difference, or he's going by some good advice from someone who can... or either you or he (or both) are confused about the term 16bit.

Just focus on creating a cool sound and performance, realistic or not, with the stuff you have.

[SnappleMan]

Mixing in a 16bit environment is completely different from using 16bit samples. You want to be working in a 24bit environment, that'll give you much greater headroom in terms of dynamic response from plugins and whatever processing you apply to your tracks, but you can use 16bit samples within your 24bit environment just fine.

It's all a matter of where the downsampling happens. If your DAW is set to 24bit, all your plugins will be locked to 24bit, if the samples within these plugins are 16bit, then there's no real loss in quality because the processing onto them will be within a 24 bit environment (and the samples are 16 bit to begin with). If your samples were recorded in 24bit, and are being dithered down to 16bit by the samplers engine, then you will lose some of the quality (though you wont be able to hear it unless you're in very expensive and tuned mastering studio, and even then you wont be able to hear it in a mix).

The big difference is what your DAW is set to; if it's set to 16bit then you're really going to be pushing too much through that bandwidth and things will get muddy. If you're in 24 bit then it doesn't really matter what the samples are set to because the engine that you're mixing them through has the headroom to reproduce them more clearly when mixed together.

Also, you can't increase the quality of a sound. You can only degrade it and modify it. Once a sound is recorded, that recording becomes the master, and anything done to the master is going to be lesser in "quality" as you regard it. You can EQ and compress and all that to make it fit a mix properly, but you can't make it sound any cleaner or more accurate than it is. Changing the bit rate of a sound will only add more bandwidth to it so signal processing will have more room to breathe, it wont magically make the sound clearer. You can try noise removal, which will always degrade the sound a bit, but when you try to compensate for it via high end EQ your sound will only get more and more off.

[dannthr]

Dude, unless you have some really weird settings, the bit rate isn't among the issues with your music. Can't tell if the track you posted doesn't sound like it's suffering from any bit depth related problems, not something that's evident to me at least. It does sound like something out of a snes game tho, with fairly raw and not humanized instruments.

I don't know what your dad is on about. Either he has good ears and is working with material where you can tell the difference, or he's going by some good advice from someone who can... or either you or he (or both) are confused about the term 16bit.

Just focus on creating a cool sound and performance, realistic or not, with the stuff you have.

Agreed, focus on creating a natural, human performance--reference live recordings as often as possible.

Mixing is its own can of worms, and yes, bit-depth can be set in project settings for almost any DAW. If you're working with ProTools, bit-depth is set at the opening of a new session.

24-bit is ideal as it offers the most dynamic range, but at the end of your session, you will want to Dither down to 16-bit for release.

The only reason not to work in 24-bit is to conserve memory as 24-bit files are about 50% bigger than 16-bit files.

[dannthr]

Mixing in a 16bit environment is completely different from using 16bit samples. You want to be working in a 24bit environment, that'll give you much greater headroom in terms of dynamic response from plugins and whatever processing you apply to your tracks, but you can use 16bit samples within your 24bit environment just fine.

It's all a matter of where the downsampling happens. If your DAW is set to 24bit, all your plugins will be locked to 24bit, if the samples within these plugins are 16bit, then there's no real loss in quality because the processing onto them will be within a 24 bit environment (and the samples are 16 bit to begin with). If your samples were recorded in 24bit, and are being dithered down to 16bit by the samplers engine, then you will lose some of the quality (though you wont be able to hear it unless you're in very expensive and tuned mastering studio, and even then you wont be able to hear it in a mix).

The big difference is what your DAW is set to; if it's set to 16bit then you're really going to be pushing too much through that bandwidth and things will get muddy. If you're in 24 bit then it doesn't really matter what the samples are set to because the engine that you're mixing them through has the headroom to reproduce them more clearly when mixed together.

Exactly.

One of the benefits to mixing in 24-bit is actually working effectively and accurately while still leaving dynamic headroom in your mix.

[Yoozer]

Your dad's not really nice if he won't even tell you what you should do

But a 16-bit console like a SNES has limited room for samples. That means that you can easily hear them loop - and higher pitches loop faster. It's a warbling effect, and it has absolutely nothing to do with your DAW's bit depth or the sample library's bit depth but solely with programming.

Play strings that lack any kind of dynamics, forget adding any positioning or room response (panning and reverb) and yes, it's going to sound fake. This does not mean putting reverb on every sample; rather, use a single reverb for an entire orchestra and position the instruments like they'd be positioned in reality.

Also, keep arrangement in mind. If you let 20 violins play a single note you get a massive single note. If you let them play a chord, 7 will play the root, 6 will play the third, and 7 will play the fifth (and most likely you should not take this as a guideline at all because it's a very naive method to split - it completely ignores any other instruments that play the same note or related notes that amplify the original chord's notes).

If you have a sample of 20 violins playing the same note, a single note may have 20 violins; but a chord will have 60. Of course that's not going to sound realistic. Listen to Mozart's Eine Kleine Nachtmusik, Allegro movement - all strings play the first part in unison, then they split up and divide roles. If you'd play that with a single kind of sample, then realism is thrown out of the window - but on a 16-bit console, you must, because there's simply not enough DSP horsepower (or memory) to play several violins separately without phasing.

[Rozovian]

Yoozers advice on splitting nots 676 and stuff is for ppl writing sheet music for real players, Id say you shouldnt care about those bits, m.

[Moseph]

Yoozers advice on splitting nots 676 and stuff is for ppl writing sheet music for real players, Id say you shouldnt care about those bits, m.

You have to worry about it less, actually, if you're writing sheet music as long as you're aware of what strings sound like played divisi.

The issue Yoozer's talking about is that sample libraries with, say, violins that can't do divisi will put the entire violin section on every note of a violin chord, artificially increasing the size of the ensemble in a way that may not sound realistic. If you're trying to do very realistic things, it's a huge problem -- I don't have a divisi-capable strings library, and I'm feeling the pinch in the things I want to do with the library.

[Yoozer]

Yoozers advice on splitting nots 676 and stuff is for ppl writing sheet music for real players, Id say you shouldnt care about those bits, m.

Uh, what?

It's exactly that phenomenon of multiplication that makes string ensembles sound unrealistic. You're not just going to make a dozen players appear out of thin air for each note.

You're not going to solve it by turning the volume down, or smacking a limiter on it, or lowering the velocity - besides the volume, the timbre changes, and no tricks afterwards are going to solve that.

[Rozovian]

The alternative is using 20 solo string patches and humanizing each of them uniquely and hoping there are enough round robins to avoid phase issues. And then their positioning and the room sound is unrealistic so you need convolution reverbs from each individual instrument position. No, you're overthinking it. Realism and quality are two very different things. A great, real performance is a bad room recorded badly and processed wrong will sound terrible no matter the realism. Sampled, decent string ensembles mixed well will sound better. This thread isn't about realism, it's abou making samples sound better, and that's a matter of sound design/processing , mixing and creating a sense of performance, not realism.

Btw, iirc synful orchestra does divisi. Haven't tried it, but it looks interesting.

[Nabeel Ansari]

The alternative is using 20 solo string patches and humanizing each of them uniquely and hoping there are enough round robins to avoid phase issues. And then their positioning and the room sound is unrealistic so you need convolution reverbs from each individual instrument position. No, you're overthinking it. Realism and quality are two very different things. A great, real performance is a bad room recorded badly and processed wrong will sound terrible no matter the realism. Sampled, decent string ensembles mixed well will sound better. This thread isn't about realism, it's abou making samples sound better, and that's a matter of sound design/processing , mixing and creating a sense of performance, not realism.

Btw, iirc synful orchestra does divisi. Haven't tried it, but it looks interesting.

Gonna have to agree with Rozovian, writing for a string ensemble and MAKING a string ensemble are very different.

[dannthr]

This thread is about understanding or more accurately misunderstanding what 16-bit audio means and has nothing to do with virtual instrument mock-up skill-though there is a real opportunity to address that (not sure what the HELL it has to do with divisi string writing since I don't even hear strings in the original audio sample).

No specific offense to the OP's dad, but we are trying to interpret the language of someone who was vague and who may or may not even understand the "advice" he imparted.

I quote the word advice because it was not actually phrased as advice but rather (at best read) a highly unhelpful and non-constructive statement that unclearly addresses any number of problems the original audio sample actually has.

If the OP's dad actually knew what he was talking about he wouldn't even have mentioned that you can record at multiple bit-depths in ProTools because Mickomoo isn't recording into his DAW from any external source and at best MIGHT be printing a RTAS source inside the box onto an audio track.

To compound this, ProTools blows for working with RTAS instruments, I don't specifically recommend it, but whatever.

[markhansavon]

Focus more on songwriting. x__x

There's a mention about using 20 solo violin tracks...It's not worth it.