Here's my question.

I use lame 3.92 (preset extreme) and EAC 0.93. I enabled the normalizing option in EAC back in the day when I started using it. Lately, I got my hands on Mp3 gain and my mp3s which I encoded, happen to be mostly at 99.9 DB, although I normalized them with EAC when I made them.

So my question is, when should I normalize, how much ( I use 98%) and How come they are still at 99.9?

Thank you

Do not normalize if you use ReplayGain

The normalization peak is not comparable to the ReplayGain loudness

--
GCP

What is replaygain? The normalizing option in EAC? In mp3 gain there is radio, constant and album. However, I am not using "MP3 gain" on my own encodings, I want them to be as original as possible. So that is why I am using the normalizing option in EAC. What is the optimal solution then?

Or, mabye, the wav is normalized at 98% by eac and then, mp3 encoding goes along and so mp3gain reports 99.9 db for the mp3 although the original wave has been normalized.


Thank you

The technique that mp3gain is based on.



This makes no sense. The normalization is lossy. Mp3gain is not.



What are you trying to archieve?
Why are you using normalization?

--
GCP

First you gotta know what normalization is.

normalization lowers the volume of your wav file. That is NOT a good idea because it can never be undone and the normalization can not be changed.

However replaygain (mp3gain for mp3) is lossless and calculates a number of how many db to lower the volume upon playback. You give mp3gain a whole album and aply album mp3gain. That way when playing the decoder will lower the volume so that your album will have the peak level you set in mp3gain. This perserves the different volume of the tracks on your cd. However you can also tell the decoder (winamp plugin) to use radio gain (which mp3gain also calculated) which will give all files the same peak level and thereby about the same volume.

With normalization you can only do it one of these ways.

Prevent clipping.

I see. So I should use mp3gain with album gain option in order to prevent clipping and preserve the volume pattern within and between tracks as created in the original production. right?


I appreciated your explanations.

:flipoff:

Normalization can never prevent that...

--
GCP

YES!!!
You got it.

That is probably your best bet. Another option would be to use a decoder like MAD that can do auto-attenuation if there is clipping.

--
GCP

but mp3gain is better if you wanna preserve the dynamics of the album.

That is my main concern, I want to preserve it.

Uhm.. In what way is mp3gain lossless? Is there really a way to undo the volume change?

Yes, but only if you keep the logs.

Regards;

Roberto.

What's the best WinAmp plug-in for this? I couldn't find any matches when I searched the WinAmp plug-ins for "radio gain."

ok, i still dont get it, and iam sorry if this was explained before, but how would replaygain help with clipping? iam aware that volume is turned down mostly, but what if there is an extremly dynamic sample (like some classical music) where replaygain should actualy raise the volume to be same as others (then peaks would clip wouldnt they?) or should i just go through the replaygain math again?

tia

I'd really like a version of MP3Gain which just added tags a la VorbisGain and MPCGain (or whatever its called).

It's soooo convenient being able to quickly change replaygain modes depending on whether I want to listen to an album, a mix, or burn an album off for listening to at work.

It looks easy enough to create an MP3Gain which stores the gain info in tags, but I wouldn't have the first clue how to go about getting the functionality added to the Winamp mp3 plugin (does Peter P own this plugin?)

That was part of my original plan (and what the Replay Gain spec actually calls for), but it kept getting pushed farther down on my "To Do" list.
Part of the reason is what you brought up yourself:



Writing to tags doesn't do any good if there's no software to read those tags. Because I had a good way to directly modify the volume so that every mp3 player could see the changes, that's what I focused on.

The other part of the problem is that there was some severe disagreement about which tags the Replay Gain info should be stored in. The ID3v1 tags are pretty much out of the question (every single byte is already accounted for). Some people wanted the info in the ID3v2 tags (which would be simplest, and my personal choice), but others vehemently argued that they should go in the LAME header.
"But what if the mp3 wasn't encoded with LAME?"
"Maybe you could put them in the LYRICS3 tag!"
"In the Xing frame but without necessarily calling it the LAME header"
"Lo here!"
"Lo there!"
And so on, and so on.

The mp3gain GUI is getting close to the point where I'll officially label it a "Release" version. Maybe sometime after that I'll think about addressing the tags issue again. (unless, of course, someone else wants to do that; the code for the back end is out there and available ya know )

-Glen

With MP3Gain, all you can do is not to raise the volume to the point where clipping occurs.

If the player is ReplayGain aware, there are multiple options. e.g. applying a hard limiter.

--
GCP

Hey Snelg,

Why not offer some choices of where to write the gain info and set the default to no gain info written into the tag. How about:

option 1 - no gain info written
option 2 - gain info written into ID3V1
option 3 - V2
option 4 - Both 1 & 2.

This would give a lot of flexibility.