24 bit vs 16 bits

[turtle_michael]

I was programming ezdrummer, and noticed something that seemed odd.

I originally programmed it set on 16 bits, then I opened a new project and used 24 bits, to hear if I could hear a noticeable difference.

I mixed them both down to waves, and did hear a remotely slightly fatter sound from the 24 bit, aklbeit almost neglibile. I had to keep playing them both 4 or 5 times each to hear if the difference i was hearing was real. Seemed to be, yet slightly.

But, what I noticed about the wav files was strange. The wave from the one done at 16 bits had these specs, 48.7 mbs, and 2822 kbps bit rate. The wave from the one done at 24 bits had these specs, 24.3 mbs, and 1411 kbps bit rate


Why it seems strange to me is that arent 24 bit files bigger and take up more room on the disk than 16? So why would the wave file for the 24 bit mixdown be twice as small and twice as slow and the 16 bit one?

or do the wav specs not reflect anything about 16 bits and 24 bits?

Seems like the 24 bit should be a bigger file taking up more mbs than the 16 one, but maybe there is an explanation for it

 

[jonny deep]

24 bit audio should indeed give you larger files and higher bit rate; 1.5 times as large.

You seem to have double the bit rate, which could be explained by double the sample rate (44.1kHz vs 22.05kHz, for example). Do any other attributes change when you change the bit rate? If you switch the project to 24 bit, does this affect the drum samples - if they are only 16 bit to start with, amending the project settings won't give you any improvement in how the samples sound.

 

[turtle_michael]

Well, Im not sure if ezdrummer samples in the daw have any bit, cause they arent really recordings they are virtual recordings of drums.

not until you mix them to wave would they have characteristics. But the 16 bit project and the 24 bit project must have a difference because the specs are different when mixed down to wave.

Only the difference im seeing seems to be backwards

 

[Ethan Winer]

Why it seems strange to me is that arent 24 bit files bigger and take up more room on the disk than 16?

Maybe one was mono and the other stereo? That would explain the 2-to-1 size and bit-rate ratio.

--Ethan

 

[turtle_michael]

Maybe one was mono and the other stereo? That would explain the 2-to-1 size and bit-rate ratio.

--Ethan

Dont know about that one. I know that to use 24 bits on a laptop I had to use an external sound card, the soundblaster for laptops.
because the regulr sound card wouldnt create a 24 bit project.

But i would think that the better sound card, would if anything produced better results. it has to be something like that though, it doesnt make a lot of sense.

 

[jonny deep]

Well, Im not sure if ezdrummer samples in the daw have any bit, cause they arent really recordings they are virtual recordings of drums.

not until you mix them to wave would they have characteristics. But the 16 bit project and the 24 bit project must have a difference because the specs are different when mixed down to wave.

Only the difference im seeing seems to be backwards

The samples are recordings of drums with all the characteristics of any recorded audio (bit depth, sample rate). They have been recorded at a specific depth and rate; changing the bit depth in use by your project won't increase the quality of the samples, although it may mean that the quality isn't reduced as much at mix down. It should be simple to find specs for the samples on the ez drummer website (one would imagine).

 

[jonny deep]

Look here: Toontrack - EZdrummer

The samples are 16 bit, 44.1kHz.

 

[Ethan Winer]

The samples are 16 bit, 44.1kHz.

So much for the perception that the 24-bit mix sounds a tiny bit better. Seriously, this is a valuable lesson that I hope is not overlooked.

--Ethan

 

[miroslav]

I would think you would need the first hear Toontrack recorded samples at 24 bits....then compare them to the same Toontrack samples at 16 bits....and then come to some conclusions about which is better.

Since Toontrack samples only come in 16 bit.....there's no way to make those comparisons or conclusions which is better (24 or 16 bit samples).

I'm sure they were recorded at a higher bit/sample rate than 16/44.1....but they probably don't offer those sample libraries due to file/distro size.

 

[Farview]

I'm sure they were recorded at a higher bit/sample rate than 16/44.1....but they probably don't offer those sample libraries due to file/distro size.

And because it really doesn't matter in the context if a mix.

I recorded all of my samples at 44.1/24 bit, chopped them up and made both 16 and 24 bit sets. I could not tell the difference. The only reason my samples are sold in 24 bit is because I didn't want to add an extra step to the process which could lead to accident getting the samples out of order and not being able to match the ones that go together (direct snare and overhead of the same hit, for example)

 

[Bobbsy]

So much for the perception that the 24-bit mix sounds a tiny bit better. Seriously, this is a valuable lesson that I hope is not overlooked.

--Ethan

Isn't this entirely the wrong argument anyway? The bit depth (24 vs. 16 bit recording) doesn't change the sound quality by one iota. The on effect changing to 24 bit has is to increase the possible dynamic range--but even there, the extra range is at the quiet end of things, not an increased level at which clipping starts to occur. I sincerely doubt that anyone is recording so near the noise floor that this comes into it. The best reason I know for recording 24 bit is that "everyone is doing it" so it makes your files more exchangeable.

No, the real argument isn't bit depth...it's sample rate. I'm a Nyquist man myself but I've spent many a happy hour debating with the 96kHz folks!

 

[DM60]

Isn't this entirely the wrong argument anyway? The bit depth (24 vs. 16 bit recording) doesn't change the sound quality by one iota. The on effect changing to 24 bit has is to increase the possible dynamic range--but even there, the extra range is at the quiet end of things, not an increased level at which clipping starts to occur. I sincerely doubt that anyone is recording so near the noise floor that this comes into it. The best reason I know for recording 24 bit is that "everyone is doing it" so it makes your files more exchangeable.

No, the real argument isn't bit depth...it's sample rate. I'm a Nyquist man myself but I've spent many a happy hour debating with the 96kHz folks!

I was waiting for someone to bring this up. Based on a link in another thread where the sound engineer (I mean a real sound engineer and not a guy behind a board) was talking about sound and digital that even the old analog tape machines could not achieve (at best)higher than 13 bit (The guy who explained it knew much more than i did and I would have to find the thread).

I am sure there is a reason for 24+ bit, but I don't have one.

 

[Bobbsy]

Well, the theoretical dynamic range of a 16 bit recording is about 96dB; for 24 bit it's about 144dB.

However, the reality is that, particularly in home studios, the digital noise floor is so far below background noises in the room and your recording chain that the digital bit depth doesn't really enter into it.

I was a bit dismissive (pun intended) of 24 bit working though. What the 24 bit depth does do is allow working with a bit more headroom at the recording stage--and this is probably a "good thing" except that, when we tell people to record nearer -18 than 0, we hear about how low this looks. If you understand what's happening, though, 24 bit working can provide some advantages--but not increased sound quality.

Of course both 16 bit an 24 bit are integer systems. If you REALLY want some advantages, go to 32 bit floating point! A lot of people will argue that the advantages this gives promote sloppy working--but if I can be sloppy and still get a technically pristine mix at the end, I don't mind being a sound slob!

 

[miroslav]

And because it really doesn't matter in the context if a mix.

Possibly...though I wouldn't always judge everything only within the "context of a mix". If you deal with very long tails/fades that slowly trail off in a mix, it might expose just the sample, appart from the mix. In some cases you might hear issues with lower bit/rate samples in those long tails, but I agree that's not going to happen often.

We all tend to work almost by default at 24 bits these days (sampling rates vary based on personal choice)...but it was mostly likely a file/size/DVD decision, and 16/44.1 just makes it easier, but I bet they didn't record the samples for the libraries at that bit/rate.

 

[DM60]

Possibly...though I wouldn't always judge everything only within the "context of a mix". If you deal with very long tails/fades that slowly trail off in a mix, it might expose just the sample, appart from the mix. In some cases you might hear issues with lower bit/rate samples in those long tails, but I agree that's not going to happen often.

We all tend to work almost by default at 24 bits these days (sampling rates vary based on personal choice)...but it was mostly likely a file/size/DVD decision, and 16/44.1 just makes it easier, but I bet they didn't record the samples for the libraries at that bit/rate.

It feels like some sanity and logic are entering into the discussion. If you do gain head room on higher bits, then this makes sens (I haven't tried, but am taking Bobbsy's word on it, after all a boring ol git probably researched the hell out of this point)

@Miro - just like the guy on another thread going to 24 and EZ drummer sampling in 16 (it states this in the specs). So, there is a least one area to confirm your point.

 

[Bobbsy]

It feels like some sanity and logic are entering into the discussion. If you do gain head room on higher bits, then this makes sens (I haven't tried, but am taking Bobbsy's word on it, after all a boring ol git probably researched the hell out of this point)

Just for clarity, there's only more headroom if you lower your recording level to take advantage of the reduced digital noise floor. Zero dB(FS) remains at zero no matter what you do as long as you're in an integer system. That's why I mention 32 bit floating point (not applicable to this EZ Drummer discussion but doable on many DAWs) because that DOES cheat the system to give headroom about the 0dB level.

 

[Farview]

Possibly...though I wouldn't always judge everything only within the "context of a mix". If you deal with very long tails/fades that slowly trail off in a mix, it might expose just the sample, appart from the mix. In some cases you might hear issues with lower bit/rate samples in those long tails, but I agree that's not going to happen often.

We all tend to work almost by default at 24 bits these days (sampling rates vary based on personal choice)...but it was mostly likely a file/size/DVD decision, and 16/44.1 just makes it easier, but I bet they didn't record the samples for the libraries at that bit/rate.

I can't tell the difference even on its on, even in the cymbals samples that actually have a long decay. This is mainly for the reasons that were already discussed, the noise floor of the room, the Mic, and the rest of the signal chain is higher than the noise floor of a 16 bit recording, so the digital noise floor is never reached.

 

[Ethan Winer]

Isn't this entirely the wrong argument anyway? The bit depth (24 vs. 16 bit recording) doesn't change the sound quality by one iota.

Exactly.

The best reason I know for recording 24 bit is that "everyone is doing it" so it makes your files more exchangeable.

Even that is a silly reason to waste 50 percent of your hard drive storage capability. All DAW software I know of can import one bit depth into a project set for another bit depth. My explanation for why 24 bits is standard is simple: mass delusion.

the real argument isn't bit depth...it's sample rate. I'm a Nyquist man myself but I've spent many a happy hour debating with the 96kHz folks!

Both parameters matter, but 44/16 is perfectly fine both for recording and distribution. It kills me when people argue that 44/16 is inadequate for high fidelity, then lament how much better analog tape sounds.

--Ethan

 

[Ethan Winer]

even the old analog tape machines could not achieve (at best)higher than 13 bit

Yes, and analog tape also has a worse frequency response than digital audio - even with frequent alignment - and much higher distortion. Just one generation loss with analog tape is obvious, versus modern digital which can withstand several generations before you can even hear a difference:

Converter Loop-Back Tests

--Ethan

 

[Ethan Winer]

both 16 bit an 24 bit are integer systems. If you REALLY want some advantages, go to 32 bit floating point!

Modern DAW software records audio at 16 or 24 bits, but all processing is handled at 32 bits FP to minimize distortion and noise. So there's no benefit to actually recording at a higher resolution.

--Ethan