Mastering MP3's,.. wtf??

[Nordenstam]

MP3's do not have a set internal bit depth. Feeding a 24 bit signal may possibly give better results than a 16 bit signal, if the encoder accepts the longer word length. Ditto with the output of the decoder.

As for mastering from mp3's - one thing to keep in mind is that the output of the decoder can clip if the destination is an integer PCM word. The best practice is to output to floating point. Or lower the gain in the decoder throgh the use of something like the free MP3gain software.

[Edward_Vinatea]

Originally Posted by Lupo

MP3's do not have a set internal bit depth. Feeding a 24 bit signal may possibly give better results than a 16 bit signal, if the encoder accepts the longer word length. Ditto with the output of the decoder.

Yes, this rings truer to me too. I don't know where I read (or heard) that is "recommended" to convert first from 24 to 16 bit (44.1?) . I never spent anytime trying to prove or disprove that. So, thank you, I will take face value what you are saying. --- thumbsup

As for mastering from mp3's - one thing to keep in mind is that the output of the decoder can clip if the destination is an integer PCM word.

Agree. Also, a mp3 resulting file for sure will be clipping even if you've set your max peak at -.10 dBFS.

Best regards,

[E-Irizarry]

Originally Posted by jayfrigo

I'm not sure why sound forge or wavelab would be any better than other DAWs at working with MP3s. We're not talking about MP3 encoding, just playback from MP3 for mastering. Decoding is not nearly the unpredictable crapshoot that encoding can be.

Awwwww I think WMP gives the least bias when going through your ASIo sound engine. I hope that I´m following the thread now. My badd.

[Lagerfeldt]

Originally Posted by Edward_Vinatea

I don't know where I read (or heard) that is "recommended" to convert first from 24 to 16 bit (44.1?)

On the contrary.

You get better mp3s by not converting to 16 bit before the mp3 processing.

[Edward_Vinatea]

Originally Posted by Lagerfeldt

On the contrary.

You get better mp3s by not converting to 16 bit before the mp3 processing.

Yes Lagerfeldt, Lupo said the same. Thanks.

[GYang]

Some younger listeners prefer sound of MP3 at 320 BR over CD quality, as they are just more used to.
I had several requests to mix directly to highest possible quality MP3, as they don't see point to have 16/44 files.
The funniest thing is that efforts done during tracking and mixing OTB are appreciated as 'better sound' compared to proposed same ITB project by same clients eventhough everything ends up as MP3.
When I tried to master such tracks myself, I've found that tracks that were played as MP3 and mastered via analogue outboards sounded better than after additional MP3 to wav files decoding applied.
Seems it's time for high-end AD-DA MP3 file converter.

[Marcus67]

Hey Ondray,

So how did it turn out? What did you end up doing?

THX

[ONDRAY]

Originally Posted by Marcus67

Hey Ondray,

So how did it turn out? What did you end up doing?

THX

since some were even 192khz mp3's , it was impossible to maintain album consistency. I converted them to 24bit/48Khz, mastered as I would. Obviously attention had to be kept to spruce up the lost highs with out getting it harsh. I found that when tweaking the lows it was almost useless on the 192 files. Anyway not my proudest or must enjoyable session. Working with shit files sucks! BUT it's a good challenge and it gets you trying new things.