mp3 loss question

[mayan]

I realize that mp3 compression causes some "loss" of quality. Do folks have any suggestions as to what effects can be used on mp3s to help ameliorate the loss. For instance, I have Ozone and a number of the Voxengo effects. Are there things that can be done with -say- an harmonic exciter or eq that will help to off-set some of the loss without making the sounds plasticine, tinny or overly steroidal? Thanks!

[JackDark]

The loss comes from compression of the .MP3 itself, the only thing that'll make the "loss better" is if you choose a higher compression ratio (KBPS rating), like 320 instead of 128. The price you pay is larger file size.

[JackDark]

Of course, if you're trying to make a crappy MP3 sound better, do some low level noise reduction, clip it, EQ it, and add some light compression. There's not a whole lot you can do once the damage's been done.

[james0tucson]

The loss comes from compression of the .MP3 itself, the only thing that'll make the "loss better" is if you choose a higher compression ratio (KBPS rating), like 320 instead of 128. The price you pay is larger file size.

There may be specific eq's or other manipulation of the frequency domain that "improve" MP3 sound for given audio. Further, the masking curve introduced by MP3 encoding might actually improve the sound. And the spatializing done by Joint-Stereo encoding can increse the perceived presence of bass. As someone else once put it, "an MP3 of Steely Dan just sounds more like Steely Dan."

[JackDark]

I know, I mentioned EQing in my second post.

If you get lucky and an MP3 compression ratio squashes the good part of your mix into the forefront, that's great, best case scenarios are fun, aren't they? But usually MP3s do one thing very well, they murder harmonic dynamics, compress the crap out of high end, and generally make a mess of the whole frequency spectrum, when used at common 128KPBS at least. That's why I don't even post demos lower than 160, but I prefer 320 when possible.

It's all subjective as hell though so whatever.

[mayan]

Damn...Thanks for the responses...I have a piece I've been working on for months. WOuld love to post it but am somewhat intimidated by the loss of quality.

[tony tony chopper]

it's the mp3 encoder that matters, they're not all equal, far from it. They'll be the ones that will have to judge what to lose, and it's rocket science. I suggest LAME.

[mayan]

IO know painfully little about this...what do you mean by the encoder having to judge...different encoders damage different parts of the spectrum? Actually, I'm trying to get my own web site up and running...hopefully I'll post higher quality mp3s or oggs there. What's generally considered to be an acceptable mp3 compression?

[tony tony chopper]

different encoders damage different parts of the spectrum?

definitely. Use a good encoder, the best ones being the fraunhofer (or something) one (mp3 comes from them), and LAME.

[mayan]

thanks, tony!

[Peel]

There's usually not much that can be done; the MP3 encoder tries its best to preserve the audio quality of the original; generally, messing with the MP3 just boosts ugly bits that were previously hidden.
However, I once had good results using a harmonic exciter on a low-bitrate MP3.
LAME is considered by many the be the best. I actually don't think Fraunhofer is considered to be so top-notch any more, except perhaps at low bitrates.
Xing is the worst, but is commonly used because it's also the fastest.

[rockstar_not]

I use LAME at 128 kbps for posting. However, I've found that whatever encoder WMP is using to make 128 kbps .wma files, it sounds a little better than my LAME encoded 128 kbps. Particularly on cymbals and frickative consonant sounds.

Does anybody know if there are multiple possible .wma codecs?

-Scott