simple question: Whats the best way to reencode MP3´s? Means i get music from various "sources" and they vary from 196 CBR to VBR or ABR.

For my portable im using -ABR 128 and thats what im targeting for since the size matter for me. So what will be the best way to get those "large" MP3´s in my 128 ABR format, without big qually looses? I know u cant reencode without looses but is there a way to minimize the loose and keep the new file still at good quality?

PS: Without knowing anything about reencoding i would decompress to wav with LAME and than just reencode with my setting?

For portable-use on a mp3-only-player I would recommened lame 3.92 or lame 3.90.2 with --alt-preset 128.

But you should give lame 3.94a13, not a14, (@rarewares) a try with this option. I think it sounds much better than lame 3.92 or 3.90.2 and is nearly comparable with OggencGT2 -b 128 (www.sjeng.org)
because lame 3.94a13 has less problems with preecho than any other lame compiles.

I use multi frontend for decoding and encoding.
It works great.

It's really not advisable to use alpha versions for anything other than testing, even if they "sound better" in some respect than an officially-released version (e.g., 3.90.3, etc.). Just because one thing sounds "better" doesn't mean that another is not broken (i.e., block-switching, noise-shaping, masking, etc.), and wouldn't it suck if you had to re-encode everything again due to a newly-noticed problem with encoding? For now, I would advise the use of 3.90.3 for --alt-preset 128 until the kinks in 3.94 have been worked out for an official release.

If its for a portable, I would suggest using Monkey's Audio as your frontend.

It has a "write id3 tag based on filename" feature that will add a tag to your abr 128 files. Lots of players read out the tags so this might be a good feature for you.

Isnt monkey audio a own compress format? Its not mp3?

btw no offence but maybe some1 could answer my base question? I mean its oki to discuss portable and alpha versions *grins* but my question was how i handle already compressed mp3 formats wich differ from my own target format? Just decopress to wav and reencode or do i need do something more?

PS: Btw im using EAC wich work realy well for Taging with the right user switch. But i just can hope more dudes would read those forum since i often have downloaded versions on 192kbit wich just sound like "xing" crap.. like some1 quoted i rather would use a 128kbit LAME jstereo file than a old MM6 or Xing encoded 192kbit file.

Btw is there any reason most dudes still encode in 160/192kbit CBR mode? I rarly see VBR files at all?

What has changed in 3.90.3 compared to 3.90.2?

In 3.90.3, the -Z option (which toggles the noise-shaping mode) is defaulted for --alt-preset standard, extreme and insane. However, if you are using bitrate settings like --ap 128, I don't think it matters whether you use 3.90.2 or 3.90.3.

You may also want to try dBpowerAMP, www.dbpoweramp.com. You can right-click on any MP3 file and transocde it while preserving your filenames and ID3 tags. Just be assured that going from MP3-anything to --alt-preset 128 will not sound all that pretty.

[quote=Cygnus X1,May 8 2003 - 05:05 PM] What has changed in 3.90.3 compared to 3.90.2?[/QUOTE]
In 3.90.3, the -Z option (which toggles the noise-shaping mode) is defaulted for --alt-preset standard, extreme and insane. However, if you are using bitrate settings like --ap 128, I don't think it matters whether you use 3.90.2 or 3.90.3. [/quote]
Is the -Z tag the "preferred" option for -aps? What do I gain by encoding my files with -Z rather than the default?

It is supposed to make a few problem samples sound a little better. For the vast majority of music out there, however, I doubt that you would hear much (if any) difference. For that reason, I'm keeping my --aps encodings that I did with 3.90.2 and using 3.90.3 only for new encodings.

Don't use the "-Z" option with 3.90.3, though: doing so will have the opposite effect, since it is already hard-coded into aps now (again, --aps now defaults with -Z in 3.90.3).

(Edit: typo)

Monkey's is also a great frontend for Lame, MP3enc and maybe the latest vorbis and mpc are still supported. Not sure about the last two.

Do a search for a few pre-echo torture samples like castanets, and compare the two compiles.

But on a portable in an active environment it will be hard to tell the difference.

Edit - I forgot to mention that Monkey will decode and re-encode without a permanent wav file.

AFAIK, john33 completely disabled the -Z toggle switch in 3.90.3.

I have never transcoded MP3 to MP3 but I have gone MPC to MP3, I just decompress to Wav and then re-encode to whatever I want, usually --alt-preset 150 for my portable, when doing this I use Speeks frontends MPC batch encoder to decompress and All2lame to re-encode, I suppose you could use All2Lame for both if it is MP3 to MP3. I find them easy to set up and you can set up your tags prior to encoding. Hope this is of some use.



Who knows ? maybe just lack of education

WinLAME will also let you do this, but I suspect dBpowerAmp is just as good or better. (I'm just decoding a Lame APS file (VBR) to 128 kbps ABR using WinLAME, and I note that WinLAME rc2 also offers a portable preset that purportedly averages 96 kbps). This method also preserves tags. (It also converts among WAV, MP3 and Ogg Vorbis)

In essence they'll all decode to WAV PCM format, so quality-wise, no method is much better for quality reasons, and they're all transcoding.

Foobar2000 player with the foo_clienc diskwriter plugin could also Convert to any Lame setting and could be made to send 24 bit audio to Lame (to reduce rounding errors) using --bitwidth 24 to tell Lame that's what its standard input should expect to see. It could also use DSP features (like Crossfeed) if you prefer that sound for headphone/earbud listening rather than the inside your skull stereo image. The improvement from using 24-bit will be minor compared to the degradation caused by transcoding, but you're not really interested in the utmost quality if you're not prepared to go above 128 kbps.

For those newbies who aren't aware, it bears repeating that converting from one MP3 format (or other lossy format) to another is transcoding and this always results in a loss of fidelity compared to encoding into the final format direct from the original lossless WAV.