• Welcome to the General Discussion forum for UAD users!

    Please note that this forum is user-run, although we're thrilled to have so much contribution from Drew, Will, and other UA folks!

    Feel free to discuss both UAD and non-UAD related subjects!

    1) Please do not post technical issues here. Please use our UAD Support Forums instead.

    2) Please do not post complaints here. Use the Unrest Forum instead. They have no place in the the General Discussion forum.

    Threads posted in the wrong forum will be moved, so if you don't see your thread here anymore, please look in the correct forum.

    Lastly, please be respectful.

16 bit vs. 24 bit

westla

New Member
Will I notice if I start working in 24 bit as opposed to 16 bit? Will I hear more \"clarity\" in my music?
 

geekeye

Member
the difference between 16 and 24 bit audio is similar to the difference between 16 and 24 bit colour, so yes, you will notice a huge difference.

unlike the difference between sample rates, bit depth is exponential. for example, 88.2 khz has double the detail of 44.1 khz, but 24 bit has 256 times the detail of 16 bit.
 

Dr_Jones

Member
For me the answer is a resounding yes! Once I made the switch to 24bit I really was blown away by the difference in sound. Now I don't have the most experienced ears, but the difference is immediate and pronounced. Now when I made the jump from 44.1 to 88.2 I couldn't quite hear much of a difference, so I switched back to 44.1
 
I agree with all of what has been said -- tracking/mixing at 24-bit vs. 16-bit was a huge step up for me.

But I have a question (one that I can and will answer myself eventually, but thought I'd pose it anyway): is there any value in changing the bit-depth of 16-bit files in older projects to 24-bit and then remixing? Or transferring 16-bit ADAT tapes in at 24-bit?

My guess is \"no,\" but am asking on the hunch that multiple people here have tried this already.
 

baikonour

Member
if you are going to remix your old 16 bit project using UA plugs, you should notice the difference as all extra processing will be done at 24 bit resolution. if you don't touch the files and just convert them to 24 bit you will not improve the quality of the original files.
 
Hmmm... but if the 16-bit files are inside a 32-bit Nuendo project, all processing is done at 32-bit anyway, so would there really be a processing gain by converting the 16-bit files to 24-bit?

I'm really doubtful about this.
 
lazlominimart said:
Hmmm... but if the 16-bit files are inside a 32-bit Nuendo project, all processing is done at 32-bit anyway, so would there really be a processing gain by converting the 16-bit files to 24-bit?

I'm really doubtful about this.
There is a nominal performance increase in having the audio data in 32 bit floating point files, at least on my machine.

It is really nominal though.

That way, there is no conversion when going into the mix engine from the file. It just goes straight in.

Since I have drive space by the boat-loads, I record to 32 bit float anyway and keep it all at 32 bit float.

This would not make sense though for some folks, especially those doing conversions and imports/exports to other apps/formats. It could add considerably to the time overhead involved in a project like that.
 

Dr_Jones

Member
I always kind of wondered about this. I alway record everything through my layla at 24 bit, but then I have my preference in cubase set to use 32 bit float files. So does cubase just \"tack on\" the extra 8 bits for more efficient processing by the 32 bit processor?
 
Dr_Jones said:
I always kind of wondered about this. I alway record everything through my layla at 24 bit, but then I have my preference in cubase set to use 32 bit float files. So does cubase just "tack on" the extra 8 bits for more efficient processing by the 32 bit processor?
I don't think it is quite like you are thinking. A 24 bit integer going in ends up being a normalized 32 bit floating point number consisting of a 24 bit mantissa with an 8 bit exponent. The same thing happens to a 16 bit integer going in.

You don't end up with more or better than what you started with, just an exact representation of the same number in a different form.

Normalized keeps the decimal point so that the number is in x.xxx exx format.
 

Von

New Member
The difference also depends on your choice of AD/DA convertors and wordclock generator.

I've heard 16-bit projects (properly clocked and recorded) that sound better than 24-bit projects (poorly clocked and recorded).

But if you use the same recording process, you WILL notice the difference between your 16-bit and 24-bit recordings....especially with efx processing.

Von
 

klimt

New Member
geekeye said:
the difference between 16 and 24 bit audio is similar to the difference between 16 and 24 bit colour, so yes, you will notice a huge difference.

unlike the difference between sample rates, bit depth is exponential. for example, 88.2 khz has double the detail of 44.1 khz, but 24 bit has 256 times the detail of 16 bit.
Not this horny chesnut again, but there is no extra "detail" in increasing your sample rate. You do only increase resolution by bit depth.

S'just physics :wink:
 

Ericcc

Active Member
There is a similar thread about this on Sawstudio forum :
http://www.sawstudiouser.net/forums/sho ... php?t=3434

I found it very interesting since this also explains why Sawstudio sounds better to me then Nuendo/Logic/Samplitude... It is all in the integer use...

More interesting is another comment that most DAWS don't do the mono thing 100 %.... Haven't tested it myself yet but it explains why I notice more clarity in audiotracks recorded by Sawstudio vs Nuendo for example. On a single track the difference doesn't pop up too much but when multiple mono tracks are summed you definately hear it !
 

jcat

Active Member
John Lance said:
lazlominimart said:
Hmmm... but if the 16-bit files are inside a 32-bit Nuendo project, all processing is done at 32-bit anyway, so would there really be a processing gain by converting the 16-bit files to 24-bit?

I'm really doubtful about this.
There is a nominal performance increase in having the audio data in 32 bit floating point files, at least on my machine.

It is really nominal though.

That way, there is no conversion when going into the mix engine from the file. It just goes straight in.

Since I have drive space by the boat-loads, I record to 32 bit float anyway and keep it all at 32 bit float.

This would not make sense though for some folks, especially those doing conversions and imports/exports to other apps/formats. It could add considerably to the time overhead involved in a project like that.
I take it you mean a cpu performance gain, not a quality gain as I beleive was being asked. :?:

There shouldn't be any quality gain using 16 bit files in a 24bit or 32bit float project, as apposed to converting the files to 24 bit or 32bit float individually first.



Cheers,

jcat
 

Vaesion

Member
I believe it can take a lil time before you hear the major difference. When i a/b compared in the beginning I didnt hear much difference but when i recorded at 24 for a month then went back and listend to 16 bit version it was obvious.

One things for sure. Im my experience i will never export from nuendo anything less then 32 bit floating and then dither outside nuendo.

that is the biggest difference Ive ever heard

when u export with l2 at 16 bit from nuendo u hear a difference when playing back in wavelab........when u play back a 32 bit floating file in wavelab it sounds just like nuendos open playing it live........when u dither that 32bit file in wavelab to 16 and then a b compare them........u cannot hear the difference.
 

boody

Established Member
although I record in 24bit, I dare say it depends on how you record. If you stay on the very save side and never lt your recordings come near the 0db-red value you _might _ hear the difference. If you record 'analog style' - everything close to max, I doubt it, unless your in the high-end acoustic recording game.

Great records have been and are mixed with 16bit audiofiles. Processing is nowadays mostly 32 bit, so you won't hear a big difference with fx. 24bit is a higher resolution that will become important when recording very clean, especially with acoustic music.

Rules I follow: never feel ashamed to follow your own ears. If you don't hear the difference immidiatly, the difference isn't big anyway. No matter how much people rave about recording quality, a lesser quality doesn't mean lesser music. Comparing 24bit/16bit in the same room might show a difference, I bet you can't tell that difference hours appart in different listening environments.

Bottom line: if the song translates well emotionally it's good, no matter how it's made.

So it's all up to you. Do you hear the difference? Do you think it's important? Do you have the processing power? What do you think?

Cheers (and please don't take this reply too seriously, it's just food for thought)
Budy
 

timmcallister

Active Member
boody said:
If you don't hear the difference immidiatly, the difference isn't big anyway.
Have to disagree in a friendly way. The cumulative effect of many "very small differences" often ends up being significant in the end.

For example: Its why a good analog front end "stacks well". The difference between a $500 analog front end and a $2000 front end may not sound a lot different with 1 track, but it certainly will when you've tracked it 24 times.

Another example is AD converter quality. I recently replaced my delta 1010 with a Lynx II. When I first got the Lynx I installed it alongside the delta so I could A/B them. The differences were subtle, and of course i ask myself was the extra cost worth it. Now after tracking and stacking a bunch of tracks using the Lynx card, i can clearly say YES. The cumulative differences are noticable.

Of course ones ability to improve different parts of the chain depend on tradeoffs, or i should say money :) so that leads to your most important point (in my opinion)...


boody said:
Bottom line: if the song translates well emotionally it's good, no matter how it's made.
Absolutely! If its a killer tune recorded with lots of passion and a great performance, a 4-track cassette might be enough.

Apart from snobs (yes, I can be one sometimes) and audiophiles, what really matters to most people is:
1) Great Song
2) Great Performance
3) and lastly recording

As much fun as it is to talk gear, what really matter is the song and performance. A 16 bit recording can capture and covey the essense of song and performance as well as 24. It's simply the technical aspects of the recording will be inferior. Is it worth the extra money/hard disk space/whatever? Everyone has to make that decision for themselves.


its all good! :)
 

boody

Established Member
but we don't disagree nescesairily. In the end you hear one audiofile: the mix. If you mix in 24bit, record the result, replace all the 24bit audiofiles in your mix by 16bit versions and record the same mix: do you hear the difference? Is it a significant improvement?

I the answer yes: go 24bit. Is it no: go 16bit and save yourself the extra processing power: you'll be able to run more tracks and plugins!

By the way: what you say about the cumulative effect; understanding that is the big part of understanding the mixing process.

cheers
Budy
 

timmcallister

Active Member
16 bit
No doubt: Many million selling albums were done with 16bit recorders. Many great sounding albums were done at 16 bit.

If you have limitations that preclude you from moving to 24 bit, then be happy and record away!


24 bit
There is a great thread at prosoundweb, it's very long and quite technical, but worth reading. In it, Bob Katz, and many others, provide sound arguments for recording with peaks no greater than -6db.

Essentially what they are saying, is the math being done while processing can introduce digital distortion when summing digital signals greater then -6db.

When recording at 24 bit, you can EASILY afford this as even at -12db peaks, the noise floor is still over 100db below you. With 16 bit you cannot afford to do this.

I (and most others) have been programmed to track as hot as possible. We did this in analog for \"tone\" and noise floor considerations. We did it in 16 bit due to resolution problems. With 24bit, we need to rethink how we deal with levels. With the fidelity offered at 24 bit, there is no need to track up to -0db. In fact, as the thread points out, tracking hotter then -6db degrades the end product. Most people who have tried the -6db peak report back their recordings sound better.

Heres a link to the thread starting at page 7 (where it starts getting good), its a tough read, but worth it IMHO.
http://recforums.prosoundweb.com/index. ... 918/0/96/0

I plan on immediately using this technique and hope to make better recordings from this point on.
 

Giles117 DP

Active Member
......and in my 5 years of tracking and mixing at 24bit recording no hotter than -9dbfs I end up with a much warmer product.
 
UAD Bundle Month
Top