bouncing to disc and audio quality loss

gizeh12 · 12th February 2007

Im having this discussion with my co-producer on bouncing to disc.

I tend to hear a decrement in audio quality of a bounced audio file (mostly of a total mix) as apposed to the mix in Cubase sx3.

So is it true that bouncing to disc causes loss of quality???

elambo · 12th February 2007

ProTools Bounce-To-Disk will degrade your audio a bit. Not as bad as prior versions of ProTools. I don't know much about Cubase but I wouldn't be surprised if it suffered in the same way.

The solution in ProTools is to record the entire mix back into a stereo track of ProTools and let that be your final. When I do that I don't hear a difference between a mix running live and the printed final.

Matteran · 12th February 2007

why would this be? It's just binary? wouldn't it be an exact digital copy?

theblotted · 12th February 2007

Quote:

Originally Posted by Matteran

why would this be? It's just binary? wouldn't it be an exact digital copy?

i think mostly how automation doesn't respond well during BTD.

Matteran · 12th February 2007

oh okay. I guess that makes sense.

joris de man · 12th February 2007

Whether there is loss of quality completely depends on how you bounce to disk.

What you are basically listening to during playback is a 32 bit mix, as Cubase's internal mixbuss is 32 bit float. If you bounce to disk to a 24 bit or 16 bit file, there will be loss of quality. Bounce to 32 bit and your mix should sound the same, unless there is clipping going on at your master fader.

Putting on a dither plugin or better still, a limiter as well so that you don't have any overshoots and you're dithering down to the correct bit depth should improve the sound of your mixdowns quite a bit.
Also, having a session at say 48 and boucing down to 44 can reduce quality. Cubase and Nuendo's resampling algo's aren't the best, so you're better of using an external resampler (wavelab 6 apparently has a really good one, or r8brain pro) if you have to go that route.

All the best,
Joe

gizeh12 · 12th February 2007

Or maybe record to an oldschool DAT recorder, or put the da ad in the recording path in stead of bouncing.

I think ill buy a 2nd hand rme card for my old pc to record the end mix.

gizeh12 · 14th February 2007

anyone else got some usefull input here?

Jrussell99 · 14th February 2007

When going down to 16-bit from 24-Bit, truncating occurs. It takes 8 of the least significant bits, usually reverb tails. Dithering is your best options. Insert it into your master fader and dither away. Dithering increases white noise to the signal, thus boosting the least significant bits, in this case the reverb tails, to significant bits. White noise then becomes the least significant. Correct me if I am wrong.

maz · 14th February 2007

Don't mean to steal this thread, but a related question:

Say you have sonar 5 and recorder in 96/24bits and now need to burn to 44.1/16bit cd quality.

First option just sonar 5 to convert from 96/24 to 44.1/16.

Second option: use ditter pluggin in sonar to do the above.

Third option: use hardware such as appogee rosetta to ditter down from 96/24 to 44.1/16.

Of the above, which would you think would give you the best quality?

I'm assuming 3>2>1, but wanted to get other opinions.

Thanks.

Jrussell99 · 14th February 2007

I hear, that going down from 96 to 44.1 isnt a great idea. Quantization error etc. If going down to 44.1, record at 88.2. If going down to 48, record at 96 etc. I am not totally sure if this is proven. It is just a theory I believe. Number 2 or 3 is your best option though. Nothing a little trial and error won't take care of.

KingDaddyO · 14th February 2007

You're probably already doing this, but always keep an original copy (copies?) of the final source file, before any conversion. That way you can redo the process over if need be, until you are satisfied with the results.

It's also advisable to keep the bit resolution as high as is practical, and static, throughout the entire editing process. Conversion would be the last step before hand off to mastering or manufacturing.

maz · 15th February 2007

Ah rats, I already recorded a couple of songs / vocals at 96. Why would you go to 48k anyways since CD's are at 44.1.

I'll look into recording at 88.2.

Thanks for the suggestion.

Maz

El Fatso · 15th February 2007

Quote:

Originally Posted by maz

Ah rats, I already recorded a couple of songs / vocals at 96. Why would you go to 48k anyways since CD's are at 44.1.

dvd

Jrussell99 · 15th February 2007

You got it el fatso.

Jrussell99 · 15th February 2007

i like you

El Fatso · 15th February 2007

the idea of recording at 88.2 for 44.1 or 96 for 48 is that the sample rate is cut exactly in half and would be a better representation of the original sample rate. supposedly when the sample rate conversion involves two numbers that are not multiples of each other it doesnt divide so cleanly.
to get to 44.1 from 88.2 you divide the samples in half. pretty simple
to get to 44.1 from 96 you divide the samples by 2.17657

im not sure if its been proven that this is true or not or if it even matters.
there was a thread a few months back on this and someone illustrated this by using images as an example. it was the letter "A" drawn on like 10x10 grid paper. then transfering it to 5x5 and a different transfer to 7x7. the 5x5 looked much better than the 7x7 since its a multiple of 10x10.
although people argued that since audio is sound and not image, that example doesnt convey it correctly.
try searching.

maz · 15th February 2007

OK, 176.4 it is then

maz · 24th February 2007

I'm still doing some searches to find this answer, but it seams that today's algorithms (especialy if using products such as apogee) can very easily handle converting 96 to 44.1 without loss/distortion.

IOW, loss of 96 to 88.2 is higher than dithering 96 to 44.1

I'll post if I find more info.

Tx

theblotted · 24th February 2007

Quote:

Originally Posted by maz

I'm still doing some searches to find this answer, but it seams that today's algorithms (especialy if using products such as apogee) can very easily handle converting 96 to 44.1 without loss/distortion.

IOW, loss of 96 to 88.2 is higher than dithering 96 to 44.1

I'll post if I find more info.

Tx

not sure if you've already seen this or not:

http://src.infinitewave.ca/

El Fatso · 24th February 2007

Quote:

Originally Posted by theblotted

not sure if you've already seen this or not:

http://src.infinitewave.ca/

so according to that is izotope the best thing to use?

maz · 24th February 2007

Funny you should mention, but I just saw that site yesterday.

The thing is that it doesn't talk about hardware based dittering/src. I'm curious to see what it would be for rosetta 200 since that's what I'll be using to sample rate convert and dither.

Daedalus77 · 27th February 2007

OK, I'm confused. Can someone tell me how to use / read that site?

12th February 2007	#1
gizeh12 Lives for gear Joined: Nov 2005 Posts: 908 Thread Starter	bouncing to disc and audio quality loss Im having this discussion with my co-producer on bouncing to disc. I tend to hear a decrement in audio quality of a bounced audio file (mostly of a total mix) as apposed to the mix in Cubase sx3. So is it true that bouncing to disc causes loss of quality???

14th February 2007	#12
KingDaddyO Lives for gear Joined: Mar 2005 Location: East Coast, U.S.A. Posts: 1,799	You're probably already doing this, but always keep an original copy (copies?) of the final source file, before any conversion. That way you can redo the process over if need be, until you are satisfied with the results. It's also advisable to keep the bit resolution as high as is practical, and static, throughout the entire editing process. Conversion would be the last step before hand off to mastering or manufacturing. __________________ ~8^) The enemy isn't liberalism or conservatism, the enemy is bullshit -- Lars-Erik Nelson Now, when there's no longer surface noise and you actually have the ability to have the most extraordinary dynamic range, people aren't using it -- T-Bone Burnett The problems that exist in the world today cannot be solved by the level of thinking that created them -- Albert Einstein I'm not black, but there's a whole lot of times I wish I could say I'm not white -- Frank Zappa

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread starter	Forum	Replies	Last Post
Patchbay Splitting => Quality loss?	vls	So much gear, so little time!	13	15th June 2009 06:50 AM
FL studio transfer to protools quality loss	optikstrange	Music computers	3	18th February 2007 03:14 AM
USB and the loss of 24 bit audio file?	deadhorses	Mastering forum	1	28th September 2006 07:32 AM
running a mult w/o imedpance matching/quality loss	alphajerk	Geekslutz forum	5	14th August 2005 01:26 AM
Pro Tools: Upsampling From 48k to 96k Quality Loss?	Perspectrum	Music computers	1	25th February 2004 11:22 AM

12th February 2007	#2
elambo Lives for gear Joined: Aug 2004 Location: Up here Posts: 6,181	ProTools Bounce-To-Disk will degrade your audio a bit. Not as bad as prior versions of ProTools. I don't know much about Cubase but I wouldn't be surprised if it suffered in the same way. The solution in ProTools is to record the entire mix back into a stereo track of ProTools and let that be your final. When I do that I don't hear a difference between a mix running live and the printed final.

12th February 2007	#3
Matteran Gear maniac Joined: Oct 2006 Location: Santa Rosa, CA Posts: 183	why would this be? It's just binary? wouldn't it be an exact digital copy?

12th February 2007	#6
joris de man Lives for gear Joined: Jun 2002 Location: Amsterdam - now Steyning, UK Posts: 571	Whether there is loss of quality completely depends on how you bounce to disk. What you are basically listening to during playback is a 32 bit mix, as Cubase's internal mixbuss is 32 bit float. If you bounce to disk to a 24 bit or 16 bit file, there will be loss of quality. Bounce to 32 bit and your mix should sound the same, unless there is clipping going on at your master fader. Putting on a dither plugin or better still, a limiter as well so that you don't have any overshoots and you're dithering down to the correct bit depth should improve the sound of your mixdowns quite a bit. Also, having a session at say 48 and boucing down to 44 can reduce quality. Cubase and Nuendo's resampling algo's aren't the best, so you're better of using an external resampler (wavelab 6 apparently has a really good one, or r8brain pro) if you have to go that route. All the best, Joe

14th February 2007	#9
Jrussell99 Gear nut Joined: Feb 2007 Location: Vancouver, BC Posts: 97	When going down to 16-bit from 24-Bit, truncating occurs. It takes 8 of the least significant bits, usually reverb tails. Dithering is your best options. Insert it into your master fader and dither away. Dithering increases white noise to the signal, thus boosting the least significant bits, in this case the reverb tails, to significant bits. White noise then becomes the least significant. Correct me if I am wrong.

14th February 2007	#10
maz Gear maniac Joined: Feb 2006 Posts: 226	Don't mean to steal this thread, but a related question: Say you have sonar 5 and recorder in 96/24bits and now need to burn to 44.1/16bit cd quality. First option just sonar 5 to convert from 96/24 to 44.1/16. Second option: use ditter pluggin in sonar to do the above. Third option: use hardware such as appogee rosetta to ditter down from 96/24 to 44.1/16. Of the above, which would you think would give you the best quality? I'm assuming 3>2>1, but wanted to get other opinions. Thanks.

15th February 2007	#13
maz Gear maniac Joined: Feb 2006 Posts: 226	Ah rats, I already recorded a couple of songs / vocals at 96. Why would you go to 48k anyways since CD's are at 44.1. I'll look into recording at 88.2. Thanks for the suggestion. Maz

15th February 2007	#17
El Fatso Gear maniac Joined: May 2006 Posts: 241	the idea of recording at 88.2 for 44.1 or 96 for 48 is that the sample rate is cut exactly in half and would be a better representation of the original sample rate. supposedly when the sample rate conversion involves two numbers that are not multiples of each other it doesnt divide so cleanly. to get to 44.1 from 88.2 you divide the samples in half. pretty simple to get to 44.1 from 96 you divide the samples by 2.17657 im not sure if its been proven that this is true or not or if it even matters. there was a thread a few months back on this and someone illustrated this by using images as an example. it was the letter "A" drawn on like 10x10 grid paper. then transfering it to 5x5 and a different transfer to 7x7. the 5x5 looked much better than the 7x7 since its a multiple of 10x10. although people argued that since audio is sound and not image, that example doesnt convey it correctly. try searching.

24th February 2007	#19
maz Gear maniac Joined: Feb 2006 Posts: 226	I'm still doing some searches to find this answer, but it seams that today's algorithms (especialy if using products such as apogee) can very easily handle converting 96 to 44.1 without loss/distortion. IOW, loss of 96 to 88.2 is higher than dithering 96 to 44.1 I'll post if I find more info. Tx

27th February 2007	#23
Daedalus77 Lives for gear Joined: Aug 2006 Location: Los Angeles, CA Posts: 1,645	OK, I'm confused. Can someone tell me how to use / read that site?