Hi to all,
I've a little doubt about what best option I've to use to normalize my tracks.
I've a very large directory with thousands of mp3, of different genres, from classical to heavy metal.
I want only to normalize the volume of all songs, because now there are too differences in the volumes of tracks.
I've tried with the -r option to a little directory of mp3 as test, but the tracks modified sounds much less loudness than the other that I've in my big collection (not yet modified), in fact in its work the program has diminish all the gain.
This can be due that the program works on a statistic made after have read all the tracks?
What option do you suggest me to coherently normalize all this large collection without change too much the volume?
I have also a little problem when I try to convert the tracks: every 20-30 tracks the program stops itself showing me this WARNING and waiting an input from me:
"WARNING: some_track.mp3 may clip with mp3 gain change 2
Make change? [y/n]:"
but this make the work impossible, because with thousands of tracks is unthinkable to insert an input every 20-30 tracks, so, is there a method to force the program to work without it asks me answers?
Thank you
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm not surprised your tracks sounded quieter overall. IME most mp3's are over amplified to begin with. Setting a reasonable gain of 89 dB to each track generally means you will reduce the track's original gain.
I'm not sure how happy you will be with normalizing so many unrelated mp3's first making some effort to group them. As you stated you have many different genres. A heavily compressed metal song next to, say, a piano cantata with a much wider dynamic range might sound awkward if both were adjusted to the same gain.
I might suggest that you categorize the mp3's into "soft, medium, and loud" folders. Then run a track analysis on each category using the GUI (I don't see any way to do this at the command line). Then you can establish sensible gain levels for each category: most gain for the "loud", least gain for the "soft".
The /r switch should do what you want (see command line switches below). The /c switch should suppress warnings about clipping, although you may end up with clipped files. Unless you analyze first, you will be shooting in the dark.
<pre>
>mp3gain /?
mp3gain version 1.3.3
copyright(c) 2002 by Glen Sawyer
uses mpglib, which can be found at http://www.mpg123.de
options:
/v - show version number
/g <i> - apply gain i to mp3 without doing any analysis
/l 0 <i> - apply gain i to channel 0 (left channel) of mp3
without doing any analysis (ONLY works for STEREO mp3s,
not Joint Stereo mp3s)
/l 1 <i> - apply gain i to channel 1 (right channel) of mp3
/r - apply Track gain automatically (all files set to equal loudness)
/a - apply Album gain automatically (files are all from the same
album: a single gain change is applied to all files, so
their loudness relative to each other remains unchanged,
but the average album loudness is normalized)
/m <i> - modify suggested MP3 gain by integer i
/d <n> - modify suggested dB gain by floating-point n
/c - ignore clipping warning when applying gain
/o - output is a database-friendly tab-delimited list
/t - mp3gain writes modified mp3 to temp file, then deletes original
instead of modifying bytes in original file
/q - Quiet mode: no status messages
/p - Preserve original file timestamp
/x - Only find max. amplitude of mp3
/f - Force mp3gain to assume input file is an MPEG 2 Layer III file
(i.e. don't check for mis-named Layer I or Layer II files)
/? - show this message
If you specify /r and /a, only the second one will work
If you do not specify /c, the program will stop and ask before
applying gain change to a file that might clip
</pre>
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Silly me... I forgot you can do analysis only (no gain modification) from the command line by specifying no parameters. That isn't exactly obvious from the /?...
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I say thank to you for your reply.
I didn't see a gui for Linux, but is possible that it could be avoid from me, does it exist?
I can make the work you have suggested me, if it is the best solution for have all the tracks with as near as possible loudness, even if I could have some doubts on which group put some album (in example soundtracks albums, with different songs).
Another think is not too clear to me, I don't have idea about how may be the value of the gain when you tell ' most gain for the "loud", least gain for the "soft" '.
Thank you
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I don't know anything about MP3Gain for Linux, but I would hope that the command line features are the same as in the Windows version. Note that I was incorrect when I said you need the GUI for "analysis only" -- my second post addressed that.
After thinking about this more, you might just want to adjust everything to 89 db (that is MP3Gain's "default" standard level) and see how your collection sounds. If some songs seem too loud relative to the others, move those in a separate directory and apply a 3 db cut to them.
Hope this helps!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Ok, but this last advertise is different from the previous, when you tell me to divide all the albums into 3 big categories.
If this is the best solution, how much would be the loudness of each of three directories?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
best thing to do is, if you already have albums in their own seperate directory, is just run it through the whole collection using 'album gain' with the albums by directory function.
And even though 89dB does sound kind of quiet when you compare it to an mp3 that hasn't been changed, it is pretty much the best setting to use. The reason being, you want to avoid clipping. If you were to take all your quieter albums and raise their gain to match today's EXTREMELY LOUD ALBUMS then you would likely have to raise the volume quite a bit and you would more than likely introduce a lot of clipping. For this reason, it is wiser to lower the volume of the loud stuff to match the more moderate stuff. After all, you have a volume control on your amp, right? Crank that up to the desired level when you're listening. That's what it's for ;)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
album gain still uses the same 89dB, but it treats the album as a whole unit. What that means is, the relative loudness of the album's tracks to each other don't change, because it treats the whole album as though it were one song. It calculates the gain based on the one big "song", and then adjusts every individual song by the same amount as all the other songs.
I'm not sure, because I haven't used it via command line much, if you can do the same things that you can from the GUI. The GUI allows you to add a whole pile of songs, and if albums are in their own directory then you can still use album gain on that batch, but it lets you differentiate between different albums because they're in different directories. One would think that the command line would let you do that too, but I'm not 100% sure because I've used the GUI more than command line.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi to all,
I've a little doubt about what best option I've to use to normalize my tracks.
I've a very large directory with thousands of mp3, of different genres, from classical to heavy metal.
I want only to normalize the volume of all songs, because now there are too differences in the volumes of tracks.
I've tried with the -r option to a little directory of mp3 as test, but the tracks modified sounds much less loudness than the other that I've in my big collection (not yet modified), in fact in its work the program has diminish all the gain.
This can be due that the program works on a statistic made after have read all the tracks?
What option do you suggest me to coherently normalize all this large collection without change too much the volume?
I have also a little problem when I try to convert the tracks: every 20-30 tracks the program stops itself showing me this WARNING and waiting an input from me:
"WARNING: some_track.mp3 may clip with mp3 gain change 2
Make change? [y/n]:"
but this make the work impossible, because with thousands of tracks is unthinkable to insert an input every 20-30 tracks, so, is there a method to force the program to work without it asks me answers?
Thank you
I'm not surprised your tracks sounded quieter overall. IME most mp3's are over amplified to begin with. Setting a reasonable gain of 89 dB to each track generally means you will reduce the track's original gain.
I'm not sure how happy you will be with normalizing so many unrelated mp3's first making some effort to group them. As you stated you have many different genres. A heavily compressed metal song next to, say, a piano cantata with a much wider dynamic range might sound awkward if both were adjusted to the same gain.
I might suggest that you categorize the mp3's into "soft, medium, and loud" folders. Then run a track analysis on each category using the GUI (I don't see any way to do this at the command line). Then you can establish sensible gain levels for each category: most gain for the "loud", least gain for the "soft".
The /r switch should do what you want (see command line switches below). The /c switch should suppress warnings about clipping, although you may end up with clipped files. Unless you analyze first, you will be shooting in the dark.
<pre>
>mp3gain /?
mp3gain version 1.3.3
copyright(c) 2002 by Glen Sawyer
uses mpglib, which can be found at http://www.mpg123.de
Usage: mp3gain [options] <infile> [<infile 2> ...]
options:
/v - show version number
/g <i> - apply gain i to mp3 without doing any analysis
/l 0 <i> - apply gain i to channel 0 (left channel) of mp3
without doing any analysis (ONLY works for STEREO mp3s,
not Joint Stereo mp3s)
/l 1 <i> - apply gain i to channel 1 (right channel) of mp3
/r - apply Track gain automatically (all files set to equal loudness)
/a - apply Album gain automatically (files are all from the same
album: a single gain change is applied to all files, so
their loudness relative to each other remains unchanged,
but the average album loudness is normalized)
/m <i> - modify suggested MP3 gain by integer i
/d <n> - modify suggested dB gain by floating-point n
/c - ignore clipping warning when applying gain
/o - output is a database-friendly tab-delimited list
/t - mp3gain writes modified mp3 to temp file, then deletes original
instead of modifying bytes in original file
/q - Quiet mode: no status messages
/p - Preserve original file timestamp
/x - Only find max. amplitude of mp3
/f - Force mp3gain to assume input file is an MPEG 2 Layer III file
(i.e. don't check for mis-named Layer I or Layer II files)
/? - show this message
If you specify /r and /a, only the second one will work
If you do not specify /c, the program will stop and ask before
applying gain change to a file that might clip
</pre>
Silly me... I forgot you can do analysis only (no gain modification) from the command line by specifying no parameters. That isn't exactly obvious from the /?...
I say thank to you for your reply.
I didn't see a gui for Linux, but is possible that it could be avoid from me, does it exist?
I can make the work you have suggested me, if it is the best solution for have all the tracks with as near as possible loudness, even if I could have some doubts on which group put some album (in example soundtracks albums, with different songs).
Another think is not too clear to me, I don't have idea about how may be the value of the gain when you tell ' most gain for the "loud", least gain for the "soft" '.
Thank you
I don't know anything about MP3Gain for Linux, but I would hope that the command line features are the same as in the Windows version. Note that I was incorrect when I said you need the GUI for "analysis only" -- my second post addressed that.
After thinking about this more, you might just want to adjust everything to 89 db (that is MP3Gain's "default" standard level) and see how your collection sounds. If some songs seem too loud relative to the others, move those in a separate directory and apply a 3 db cut to them.
Hope this helps!
Ok, but this last advertise is different from the previous, when you tell me to divide all the albums into 3 big categories.
If this is the best solution, how much would be the loudness of each of three directories?
best thing to do is, if you already have albums in their own seperate directory, is just run it through the whole collection using 'album gain' with the albums by directory function.
And even though 89dB does sound kind of quiet when you compare it to an mp3 that hasn't been changed, it is pretty much the best setting to use. The reason being, you want to avoid clipping. If you were to take all your quieter albums and raise their gain to match today's EXTREMELY LOUD ALBUMS then you would likely have to raise the volume quite a bit and you would more than likely introduce a lot of clipping. For this reason, it is wiser to lower the volume of the loud stuff to match the more moderate stuff. After all, you have a volume control on your amp, right? Crank that up to the desired level when you're listening. That's what it's for ;)
Thank you, you were very clearly.
With the "album gain" you mean the /a option of the program, right?
Does this method apply a gain of 89dB?
album gain still uses the same 89dB, but it treats the album as a whole unit. What that means is, the relative loudness of the album's tracks to each other don't change, because it treats the whole album as though it were one song. It calculates the gain based on the one big "song", and then adjusts every individual song by the same amount as all the other songs.
I'm not sure, because I haven't used it via command line much, if you can do the same things that you can from the GUI. The GUI allows you to add a whole pile of songs, and if albums are in their own directory then you can still use album gain on that batch, but it lets you differentiate between different albums because they're in different directories. One would think that the command line would let you do that too, but I'm not 100% sure because I've used the GUI more than command line.
Yes, and yes.