I just want to know how to decipher the MP3Gain logs so I can reverse any changes I get. For instance, this is what is printed in the change log:


"8/10/2002 2:56:11 PM E:MP3's Marvin GayeMarvin Gaye - Sexual Healing.mp3 -2
8/10/2002 2:56:46 PM E:MP3's Marvin GayeMarvin Gaye - Sexual Healing (Extended Version).mp3 -2
8/10/2002 2:57:09 PM E:MP3's Marvin GayeMarvin Gaye - Take This Heart Of Mine.mp3 -2
8/10/2002 2:57:32 PM E:MP3's Marvin GayeMarvin Gaye - Try It Baby.mp3 -2
8/10/2002 2:58:09 PM E:MP3's Marvin GayeMarvin Gaye - What's Goin' On.mp3 -2
8/10/2002 2:58:34 PM E:MP3's Marvin GayeMarvin Gaye - When a Man Loves A Woman.mp3 -2
8/10/2002 2:59:04 PM E:MP3's Marvin GayeMarvin Gaye - You Sexy Thing.mp3 -2"

I would imagine the -2 is what is changed. However, I know that I changed it by -3db. Does it really mean that it was changed by 1.5db X 2?

Exactly. Just multiply the step-incremental numbers with 1.5

Thanks, that's what I thought, just wanted to make sure.

Yes, in older versions it showed the actual integer gain change applied to the mp3 file.

As of version 0.8.2 (I think that's which version it was), I "fixed" the gain change logs so that they show the same scale as the rest of the program. I think I forgot to mention this before.
It's not hard to tell which "version" of the logs you have: If you see only whole numbers (1, 2, 3, etc.), then it's the old version. If you see multiples of 1.5 (1.5, 3, 4.5, etc.), then it's the new version.

-Glen

Thanks, where do I get the new version?

http://www.geocities.com/mp3gain/

Another question. What do most people use as the target normal volume? I think the default of 89dB is way too low. Hell most rips will come in around 97 or 98 dB. Right now I'm using 96 dB and it seems to do well for most things. Just wondering what the general consensus is?

A setting that high should cause clipping with most music.

Why is the default 'too low'? You can turn up the volume...

If you use something besides 89dB, you'll be incompatible with other formats ReplayGain implementations.

--
GCP

I was just thinking that it's too low, cos apparently everthing I rip is clipping according to mp3gain. Especially newer music, usually up around 98 or 99. If I use mp3gain on a few tunes, those are tremendously lower volumed than anything else. I could see 89 if you did every single one of your mp3's. Just my $.02

Right now I have version 0.7.6 with backend v1.07 installed. What should I install from the site? Preferably, I'd like to uninstall and then install from scratch. Or should I just install the .EXE and helpfiles, or the GUI and CLI ? I'm a little confuesd. It;s late and I think I've had a few too many tonight.

"You can fight without ever winning, but never win without a fight."

That's exactly the point of using mp3Gain: make ALL your mp3's sound equally load!

Sikke

I understand that, but I don't think that should involve lowering the volume on most of them, instead it should raise the low ones.

But that would introduce clipping! If you mp3Gain your total collection to say 95 dB, you can be sure that a lot of the mp3's will clip! Also, 89 dB is the standard...

just wondering why that would mean that almost every thing i rip with eac and then encode using --aps would be lowered? Does that mean that it's clipping already? And if that is so, shouldn't the presets be using scaling? Just a thought

Most modern Pop albulms are so highly compressed that they will clip very often when compressed to mp3. Check your current mp3 collection with mp3gain (just analyze) and see how many will clip...

haven't analyzed the whole lot, but according to mp3gain, even some of my old rush discs (not the remasters) clip

just so I have this right, my understanding is that mp3 encoders can not handle the full dynamic of music. IE, if something is peaking at o or -.01 dB, it will automatically clip as an mp3?

What is the magic # for original gain then? And isn't this a problem with the encoder?

*edit* btw, most newer pop albums clip anyway...

Damn it, they did an awful job mastering Vapor Trails...

It's not taht mp3's cannot copy with the full dynamics of a CD. I think ff123 has some excellent info on his website why mp3's can clip. (http://www.ff123.net/norm.html)

MP3 encoders have no problem handling the full dynamics of music. The problem is that the digital processing that is done to it (and this isn't restricted to mp3, but to any lossy codec or any kind of processing applied to the file) occasionally expands it, and that your playback hardware and software can't handle this.

MP3Gain fixes this by lowering the volume until the dynamic range 'fits' again.

--
GCP

ahhh, so mp3gain makes the file fit the player? What about a player that compensates for that? Is that what the MAD decoder does?

nother question. I have been using v0.7.6. I just downloaded the newest version (v.0.9.5) . Can I install over the old installation or do I have to uninstall and then install?

You can install over the old version.

thanks...


will that keep all the setting I have set?

I don't know... To make sure, write your settings downs...

Prob will, just wondering if someone's done it

Yes, it will keep all your current settings.

-Glen

So, question(s)....

1) If i 'Album Gain' all my CDs.... then sometime down the road want to revert them back to their orginal setting.... do i have to file through the logs and slowly change them one at a time back, or is the an 'automated' way to restore all my album back at once.?

2) And... when MP3Gain says 'file will clip', and you go through with it, does that permentally clip your file, or if you revert the change, will your file 'unclip'.... ?

Thanks...

From what Snelg's been saying, it sounds to me like clipping isn't a problem with the MP3 itself, but rather goes back to limitations with the hardware/software decoding it. Hence, lowering the volume back should work just fine for "unclipping" it.

Just my 2 cents...

1) Yes, it's painfully manual. Sorry, this is one of the biggest "To Do" things I've had on my list for months now. It's still not done

2) Yes, if the mp3 is originally not clipping, and you apply a gain that makes it clip, then if you apply the negative gain it will "unclip". I can give you a technical explanation if you want...

-Glen

Hi,

2) No thanks, i think i understand why.... As long as i'm not making any permenent changes i'm happy

1) As for undoing the gain manually via the log files.... my question is, At this stage, do you know that if in the future when you implement an 'automated universal gain undo', will it be likey to use the current logs as they are generated now, or do you think you will have to change the way the logs are created for it to work. -- just wondering wheather i should wait to gain my entire collection... to avoid the painfully slow process of undoing it...

Thanks

The "undo" will probably be through tag info dumped directly into the mp3 file itself (yes, yes, I know the vehement opposition that some folks have to tags).

Duh. I'm a freakin' moron. I just realized that a simple undo function based on the Change logs is not that difficult. I can scan through the change log for a given filename, find all changes made, sum those changes, and apply the negative of those changes. So if you've done +1.5 dB and -6 dB, mp3gain would just have to apply +4.5 dB to cancel it out.
This would only work if you make sure that you always log the changes, of course.

So go ahead and change 'em now, but make sure to keep the logs. Even if I don't actually include this as a feature (I'm a bit paranoid that people will not understand/mis-use it), I can write a quick app to do the same thing.

-Glen

Ah, yeah, that's what i thought you'd do....



Of course...

However -



This seems like a 'cleaner' method. Less chance of an undo error occuring. (due to bad log keeping, or a directory changing, or something similar....)



I don't, what would an argument against tags be ?