mp3 conversion question -Sound Forge - PC - MAGIX Community Forums Archive

Subject:mp3 conversion question Posted by: drbam Date:12/22/2004 9:26:02 AM After converting a 16/44.1 wav (mono) to mp3, the volume peaks are increased by .1 to .5 db. These are educational lectures and I edited the .wav files with the Waves L2, limiting peaks at -2 db. Anybody know why this conversion increase is happening? I'm glad I didn't push the original peaks higher or I would be into some distortion here. (using SF 7) Thanks! drbam
Subject:RE: mp3 conversion question Reply by: Sonic Date:12/22/2004 10:43:31 AM mp3 is a lossy compression format. Even at the highest bitrates it will slightly modify the audio data, which can result in small changes in individual peaks. So a little headroom is a always a good idea. To verify, try higher bitrates and see if the problem is less pronounced. 0.5 dB sounds suspiciously large, though. Since you have some good material to test this, I'd be curious to know if you can get the encoder to produce audible multi-sample clips if you remove the headroom (just peak normalize prior to encoding) or if it's better when you don't apply the L2. J.
Subject:RE: mp3 conversion question Reply by: drbam Date:12/22/2004 11:02:45 AM Thanks for the reply. I'm not applying the L2 during mp3 conversion – just during the final rendering of the edited wav file. The L2 is necessary to bring the huge dynamic range of the lecture material to a more listenable end product. I have not tried to accomplish this by using a normalizing process as the L2 is just so damned good for this kind of thing. The mp3 conversion stage is just straight forward without any added processing. drbam
Subject:RE: mp3 conversion question Reply by: Sonic Date:12/22/2004 1:37:48 PM Right, I meant do the peaks increase as much if you don't apply the L2 to the source material? mp3 essentially sneaks the more audible bits in where the less audible bits would go, so I wouldn't be suprised if your problem is worse when everything is heavily compressed (in the audio sense, not the data sense). Just a little experiment if you have a minute to try it. No big deal unless you suspect a real bug here. J.
Subject:RE: mp3 conversion question Reply by: drbam Date:12/23/2004 6:03:41 AM Thanks again. I've already completed about 20 hrs of lecture material that now include the L2 and the project deadline won't allow for additional time to re-do some this. However, during the next similar project (in about 4 months), I will experiment with the approach you suggest. drbam

Subject:mp3 conversion question

Posted by: drbam

Date:12/22/2004 9:26:02 AM

After converting a 16/44.1 wav (mono) to mp3, the volume peaks are increased by .1 to .5 db. These are educational lectures and I edited the .wav files with the Waves L2, limiting peaks at -2 db. Anybody know why this conversion increase is happening? I'm glad I didn't push the original peaks higher or I would be into some distortion here. (using SF 7)

Thanks!

drbam

Subject:RE: mp3 conversion question

Reply by: Sonic

Date:12/22/2004 10:43:31 AM

mp3 is a lossy compression format. Even at the highest bitrates it will slightly modify the audio data, which can result in small changes in individual peaks.

So a little headroom is a always a good idea. To verify, try higher bitrates and see if the problem is less pronounced.

0.5 dB sounds suspiciously large, though. Since you have some good material to test this, I'd be curious to know if you can get the encoder to produce audible multi-sample clips if you remove the headroom (just peak normalize prior to encoding) or if it's better when you don't apply the L2.

J.

Subject:RE: mp3 conversion question

Reply by: drbam

Date:12/22/2004 11:02:45 AM

Thanks for the reply. I'm not applying the L2 during mp3 conversion – just during the final rendering of the edited wav file. The L2 is necessary to bring the huge dynamic range of the lecture material to a more listenable end product. I have not tried to accomplish this by using a normalizing process as the L2 is just so damned good for this kind of thing. The mp3 conversion stage is just straight forward without any added processing.

drbam

Subject:RE: mp3 conversion question

Reply by: Sonic

Date:12/22/2004 1:37:48 PM

Right, I meant do the peaks increase as much if you don't apply the L2 to the source material?

mp3 essentially sneaks the more audible bits in where the less audible bits would go, so I wouldn't be suprised if your problem is worse when everything is heavily compressed (in the audio sense, not the data sense).

Just a little experiment if you have a minute to try it. No big deal unless you suspect a real bug here.

J.

Subject:RE: mp3 conversion question

Reply by: drbam

Date:12/23/2004 6:03:41 AM

Thanks again. I've already completed about 20 hrs of lecture material that now include the L2 and the project deadline won't allow for additional time to re-do some this. However, during the next similar project (in about 4 months), I will experiment with the approach you suggest.

drbam