-

Absolutely.

I mentioned in another thread that if these tracks were created by Lame (EDIT: without the --noreplaygain switch) and have Lame headers, you'll be able to retrieve RG information from the headers and alter them using the appropriate 1.5dB steps in order to restore them. Using a program like EncSpot, you can check their CRCs (after removing any APE tags, of course) to see if the files are the same as they were originally.

Certainly this is easiest when album gain has been applied. You can also take a look at the amplitude of the original file(s) in a wave editor in order to determine the proper adjustment. But if you have the originals, it may take less time simply making a new set of compressed tracks.

greynol, how do I retrieve using the lame rg tags?

If you use MP3Gain to add it, then it is perfectly safe.

MP3Gain works by modifying this "Global Gain" value on every MP3 frame, not by reencoding. Thus it is easily modifiable, and bears no effect at all to the reconstructed waveform (except for the amplification that is).

Check the wiki for more information:

http://wiki.hydrogenaudio.org/index.php?title=MP3Gain

IIRC MP3Gain by default writes all gain changes to a textfile named "changes.log" in its program directory.
If it doesn't you can activate logging via Options/Logs (by specifying a path).

With LameTag.

@Lodgikal and pepoluan,
Bourne wants to know how to restore the original global gain values assuming that the undo information is not available...
I'm sure it's safe to assume Bourne wants to know how to do this in the absence of information in the change log also.

Mmmm, smurfs!

EDIT: Picture of Gargamel eating a smurf removed.

I know that. I'm answering a specific question that he has in his OP.

Sorry to butt in... where exactly is the CRC information in EncSpot?

If the gain is undone correctly, the "Actual Music CRC" and "Music CRC" fields in the LAME header will have the same CRC... correct?

Mike

edit: typo

Right click on a file, choose "Lame Header" (second option down). In the "Lame Tag" tab, it's the last item, "Music CRC," and it will either say "corrupt" or "OK."

I believe so, but to be honest, you've had more experience with LameTag than I have. I didn't even know about it until I read your post, lol (and thanks, btw).

I should again mention that the mere presence of an APE tag will show the Music CRC as corrupt in EncSpot.

Ahhh, it sure does... COOL!

Thanks,
Mike

P.S. Is a 58 year old bald guy allowed to say "COOL"?

from your looks in the avatar, I thought you were 37... LOL

Maybe I need to change my avatar to have a grey/gray gotee?

http://en.wikipedia.org/wiki/Grey says, "Grey or Gray is the color of dull things..."

Mike

Greynol...



The headers only show track gain information, albumgain is not stored. I did albumgained the files. So, if I see -4.5db in the trackgain info, I go in MP3Gain and make that file +4.5db, and will be restored!?

Bourne, you make a good point about that.

I'd take the two tracks with the smallest change and work back from there applying the closest level or two until the CRC checks out again. Once you've found the appropriate change, apply it to all of them.

It's much easier with the Lame header information, but I've also done it by looking at the track with a wave editor and guessing the appropriate level. Having the Lame header simplifies things since it will allow you to make a more educated guess.

I just wanted to point out there's another place (in addition to tags) to look for undo information. IMHO there's a good chance the changes.log still exists even after messing with tags.

yes, i know of CHANGES.LOG, but somehow these files are albums that have lost the information only by burning them to a crap media CD-R. I no longer have infos on it or change.log.

ok, I got the recover it!!! so it led us to the conclusion that... no lame tag present, no way to revert to the perfect first bits!?

Well, you can, but it's a bit more of a guessing game. Like I said earlier, you can look at the levels using a wave editor. I've done it.

Glad you were able to fix your files.

It depends - if you start with a typical modern (i.e. loud!) CD, and you compress it to a high quality mp3, then most of the tracks will clip, but only a little bit.

So if you changed (probably reduced) the gain of the whole album with mp3gain, it's a fair bet that if you increase the gain of the whole album until most of the tracks clip a little bit, then you've got back to where you started.*

I don't know why you'd want to do this though. If you just want it "loud" again, it would be best to apply max no clip gain by album.

Cheers,
David.

* - This won't work with nicely mastered CDs though, because you can't assume that mp3s of these CDs clipped to start with.

well... i just want them to have the undo information stored... i will probably keep them where they're at, once recovering and re-storing the undo information...

but that leads me to a question.......

we know that a modern CD have hundreds of CLIPPED SAMPLES...
when we convert it to hight-quality MP3 (VBR 0), do theses files will clip equally to the CD, or will the conversion to MP3 add much more clipping?

The low-passing done during the conversion can remove much of the clipping. If the distortion prior to converting is audible, it will probably be audible after the fact, even if the flattened-out parts become rounded off.

I'm sure you already know that mp3s can cause clipping upon decoding which doesn't exist on the original source.

I should really leave this up to the experts. I see that Gabriel is reading this thread at this very moment.

you know... lemme tell ya...

I only use MP3Gain for the single sake of removing the general clipping from the tracks...

If convert a CD to hi-quality MP3 does not add MORE clipping than what it's got on the source, and if does not make the very clipped samples worse and louder, then I see no reason for why using MP3Gain since I would have listented the raw CD with these millions clips anyway....

That's why I am questioning....

Does an original PCM clipped have the same clipped samples and the same intensity clipped samples as its MP3 converted version?

The intention of MP3Gain is primarily to give your tracks/albums equal loudness.

As I already mentioned, mp3s can add additional clipping upon decoding which do not exist on the original source. MP3Gain is able to detect this and allows you to attenuate the track in order to prevent this from happening.

But as far as general clipping goes, if your mp3s are transparent then they're transparent. Bumping the level (within reason) isn't going to change anything.

what you mean by bump? raise?
yes, the MP3 are clear transparent...
you recommend me use Mp3Gain?

He means changing the volume isn't going to improve quality, unless the decoder was clipping as sometimes happens with very loud but unclipped source CDs.

Technically speaking, yes.. mp3 can add more clipping (more places, more pronounced) than what's in the original one. This is a consequence of compressors applied at the original signal so that the source sounds stronger. When an mp3 is generated out of it, it cannot reproduce the exact waveform, and goes out of bounds (clips). This clip happens in the decoding stage, when quantizing to integer values for reproduction (or rendering).
(Similarly to what can happen when resampling one audio signal)

I think that I lost the UNDO tag information of some files by editing the tags in foobar2k. Not that I even checked APE tags off of them. But somehow, re-saving a title or a album name in the general info tag makes that info that is stored in APE tag go away... at least the UNDO info of MP3Gain did go away...