What's new

Is there a worthwhile benefit to 24-bit samples over 16?

Mike Greene

Administrator
I've asked myself this same question about whether 24 bits is worthwhile for a sample library. For the reasons Charlie stated really well, my opinion is that it isn't. We (Realitone) record at a 24-bit sample rate, but we then normalize all samples, so at that point, my opinion is there's no need for the extra dynamic range that 24-bit offers.

For example, a quiet guitar sample when normalized turns into a very loud guitar sample, so in the mapping editor (or in the scripting), that sample gets volume-reduced by 30 or 40 db so that it will sound right. So the effective dynamic range of that guitar zone is 30 or 40 db plus whatever the dynamic range of 16 bits would be. That's plenty.

So if I were making libraries for myself, where I record 24 bit and then normalize samples individually, I'd make the final samples 16 bit. Partly to save hard drive space (minor issue) and partly for the instrument's RAM footprint (bigger issue). Plus I assume 16-bit is easier on the processor, although I'm not sure about that part.

But ... I'm not making libraries for myself anymore, so not all my decisions can be based on cold hard facts. I'm trying to sell these things, and many potential customers have a lot of preconceived notions. In many people's minds, bigger is always better, whether it's in total gigabytes, or whether it's number of round robins, or whether it's bit depth. So for that handful of sales I might lose to the guys who think our quality isn't up to snuff if we use 16-bits, I stay at 24-bit.

That's all just my opinion, mind you, and by no means have I cracked the mystery of how to run a successful sample library company, so take it for what it's worth. In fact ... having said all that, I'm considering switching (quietly) to 16 bit for an upcoming library where the RAM footprint and processor load will be a major issue.
 

Fleer

Feeding the Trolls
I think you’re right about that, Mike. Otherwise, 48-bit should be next. Better to focus on mic options and the like. Those are the options that matter, apart from good source material and recording quality.
 

GtrString

Active Member
If you do decide for 16bit, please write it in the specs. I read that for everything, and will not buy a library with 16bit samples (maybe except for something cheapo, retro sounding stuff).

I even prefer 24bit, 48khz. Nobody is even close to the perfect sample library, so skimping on specs is the kiss of death, imho.
 

Wolfie2112

Senior Member
I've asked myself this same question about whether 24 bits is worthwhile for a sample library. For the reasons Charlie stated really well, my opinion is that it isn't. We (Realitone) record at a 24-bit sample rate, but we then normalize all samples, so at that point, my opinion is there's no need for the extra dynamic range that 24-bit offers.
+1 to this. I never use 24bit samples if available in a VI, always seems like overkill. 24/48 for recording and stem deliver? Absolutely, but not for the instruments themselves.
 

charlieclouser

Senior Member
Yes, I agree with Mike Greene's view. As long as the original signal acquisition was done at 24-bit, and the samples have been normalized during the editing process (while still at 24-bit), and then reduced to 16-bit, all is well - and your cpu and storage system will thank you!

I certainly don't bother with attempting to down-convert Kontakt libraries to 16-bit, or take any other tedious data-saving measures like that, but I have no qualms with using 16-bit sample libraries.

But I generally keep my own samples at 24-bit all the time, since more than once I've zeroed in on some tiny little squeak or squonk at the end of a sample and decided that's the bit I want to normalize and turn into a featured sound, so it makes sense to keep everything at maximum resolution just in case.
 
OP
DSmolken

DSmolken

Senior Member
In many people's minds, bigger is always better, whether it's in total gigabytes, or whether it's number of round robins, or whether it's bit depth. So for that handful of sales I might lose to the guys who think our quality isn't up to snuff if we use 16-bits, I stay at 24-bit.
Ha, this is what it really comes down to! Staying at 24 bits lets me skip the effort of normalizing and then compensating for the normalization gain of each sample in the sampler. Sell a few more, and move on to making the next instrument faster. I think I'll do that with any instruments that aren't really resource-heavy or intentionally lo-fi.
 

Dex

Member
If you do decide for 16bit, please write it in the specs. I read that for everything, and will not buy a library with 16bit samples (maybe except for something cheapo, retro sounding stuff).

I even prefer 24bit, 48khz. Nobody is even close to the perfect sample library, so skimping on specs is the kiss of death, imho.
On the other hand, for most things I prefer 16 bit libraries (assuming they’re done right, as above) for the ram, cpu, and disk space savings. I deleted my 24 bit sonokinetic libraries and just use the 16 bit versions now.
 

GtrString

Active Member
On the other hand, for most things I prefer 16 bit libraries (assuming they’re done right, as above) for the ram, cpu, and disk space savings. I deleted my 24 bit sonokinetic libraries and just use the 16 bit versions now.
Ok, but for commercial projects I just cant justify that. The competition is so fierce that every small inch counts, and I can hear a difference in depth and dimension between 16 and 24 bit samples, as well as a difference in the top end between 44khz and 48khz (with acoustic instruments). Also developers dont count in the processing you might do in the mix stage, and I need to be able to take a little hit from da/ad conversion as well (running outboard gear). So I won’t deal with samples that juust cuts it for the intended use.

Often Im on the fence using samples at all, so while I understand the reasons for size ect (allthough computers gets more powerful all the time), I feel that is more of an alternative route. If the customer base accepts it, of course it is fine, but then Im not in it. But it does matter if we talk acoustic type samples or just electronic sounds. For electronic sounds, the differences might not be as noticeable.
 

Voider

Came from the future
We've made a little blindtest here in the forum around a year ago, me and a few users included, and we all failed to tell the 16 bit from the 24 bit files apart. But before when we compared them side by side, we felt that 24 bit would sound so much bigger, wider and more three dimensional. The blind test proved us wrong.
 

Nick Batzdorf

Moderator
We've made a little blindtest here in the forum around a year ago, me and a few users included, and we all failed to tell the 16 bit from the 24 bit files apart. But before when we compared them side by side, we felt that 24 bit would sound so much bigger, wider and more three dimensional. The blind test proved us wrong.
Whether you hear a difference is highly dependent on many things.
 

Voider

Came from the future
Whether you hear a difference is highly dependent on many things.
Or it is simply not that huge of a difference. What you said, professional violin players said about the Stradivari, but in a blind test most of the top players couldn't tell them from other violins apart. Same goes for wine with experts. We feel what we believe and expect. If the $2 wine is in the $2000 bottle it might taste like a $2000 wine, and if we read that a file is 24 bit instead of 16 bit, we might hear more than there actually is. Our brain tries to satisfy our high expectations if we got them, and the same vice versa. The $2000 wine tasted cheap and bad experts claimed, because they drank it out of a $2 package without knowing it's the 2000 dollar one.
 
Last edited:

AllanH

Senior Member
Even if I'm looking at a 24bit vs. 16 bit recording, how do I really know what I'm hearing? How did it get to 16 bit from 24? Was it recorded in true 16 bit or down-converted by the audio interface from 24 during recording? How about the DACs used during playback? I'm not really sure there is one "best answer" as there are too many steps involved from source to ears.
 

JohnG

Senior Member
I'm not really sure there is one "best answer" as there are too many steps involved from source to ears.
Exactly.

Apropos the "blind tests" some alluded to, if you're playing back on Soundcloud or something like that, it's mighty hard to know what you're really listening to at all. So I don't even think the tests are very valid, unless conducted in a pristine manner.

I don't know how one could arrange a pristine test online unless people all are downloading the original files and listening back on very good equipment, with good D/A and all that. As Allan says, "too many steps" to be all that confident about what people are reporting.
 

Dex

Member
I figure you could just make a simple kontakt vi where, for instance, C4 is either the 16-or 24-bit version of a well-recorded 24-bit sample and C5 is the other version, and have people report back which they think is which.

If you want to make sure they're not cheating by looking at file sizes, re-encode the 16-bit file as a 24-bit file so both versions of the sample have the same file size.
 

Nick Batzdorf

Moderator
Or it is simply not that huge of a difference
Sometimes it's all but inaudible, sometimes it's very audible. If you record something really complex, say a piano or a ride cymbal, and let it ring out in a room... yeah, you'll hear the difference.

16 bits is great as a release format, but it does often make an audible difference when you use 24 bits as a production format and then dither down to 16 bits at the last stage.

How did it get to 16 bit from 24?
Hopefully using dither.

Now, I'm going to be honest and say that 20 years ago when I listened to different kinds of dither, I couldn't hear a difference. And while my hearing is still very good, touch wood, I'm sure I'd hear nothing whatsoever different even better now. :)

But truncating from 24 bits down to 16 without dither can sound harsh - again, sometimes.
 

GtrString

Active Member
General online blindtests are not dependable. What matters is what you hear in your own listening environment, because that's where you react to the music and make mix desicions. Of course, this can always be contested, and there are a myriad of variables. But general consensus online don't mean shi*, really.
 
Top Bottom