hello,

I heart that I could save some bits when I replay gain the wav file before encoding. However, since I use Cdex for ripping it only allows me to 'normalize' the track. Would it be save to just normalize every song to let's say 70db just to make sure that it won't clip? I use mp3gain anyway, so I could restore the volume to the way I want anyway, but it would have the advantage that I would save some space and don't have to think about of how much I file needs to be replay gained
thanks

I wouldn't do it. You only save very little, and quality would decrease that way too. Normalizing in CDex has nothing to do with Replaygain anyway.

thanks for your answer!

is this true just for Cdex normalization or for replay gain too?

gday...


i got a script that uses wavegain on a dir.
before batch encode that. the point would be obvious
but both wavegain and lame (ex) would need some improvements
"it seems".

wavegain/replaygain uses to big roundoff to scale..
but that depends on the shoosen encoder.
(and evt: decoder)
(ex: -0.71 would be rounded to 0.8 not 0.7 or .72/75)
mp3gaining the file would do about the same i guess..
but the given values is aprox. doubled.. when decoding
with lame..))

i tried it with lame 3.95.1.. and the resoult
after decoding with mad. the gain values is "messed up"
not so bad with foobar.. (don`t ask me why..)
but still lame adding extra to the peak..
if it didn`t do that.. the decoded wav would be identical to the source..
this method is better than mp3gain..
there is a slightly diff. in lame builds
but mppenc is almost perfect..

to bad this is a cmd script.. i could be funn to use
it as a loader..

@ n68

thank you very much for your response! However, could you draw a conclusion? What should I do? Just use Cdex using --aps and after that replay gain or go through all this stuff with wavegain you described

gday..

personaly i would prefer gain before a encoding..
ex. lame encodings would benefit more from that
then mp3gain..

thus. it has been a long time since i have used CDex..
as CiTay says.. the "normalizing" function in most software.. (CDex)
is NOT the same as replaygain.. it actually makes the micro dynamics
worse..

i don`t find mp3gain to be bad.. not at all.
if ure just play the tracks on pc.. mp3gain is shosen.
but i reckon everyone is going to burn them
in the end.. so gaining them before is a good thing..
to bad this method isn`t perfected..
but still better than replaygaining a existing encode.

in time.. when RG. and div encoders are perfected..
this would be without a doubth.. the best way.

(i hope the respective developers..one day.. would look into this..)
but as i say.. i do recomend it as of now.. anyway.))

ok, thanks for your answer . I have another question: When I mp3gain my mp3 files and burn then on a CD (in wav) do the mp3 files have the original volume, just as if mp3gain was never applied? (I somehow understood that from your post).

No they don't, they would have the mp3gained volume. mp3gain changes the actual data within the mp3. It's not just a tag like ReplayGain tags.

The wav files will have the adjusted volume, not the original volume. MP3Gain modifies the global gain field of every frame in the mp3 file and any decoder will generate output based on this global gain field. The only way you will get wav files at the original value is if you undo the gain changes within MP3Gain itself.

EDIT: oops, Cyaneyes got there first.

that's what I thought too =) thanks!

Edit: thanks to phwip as well! We posted at the same time