Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: MP3 Normalizers (Read 18400 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

MP3 Normalizers

Reply #25 –
Quote from: papadoc on
If a difference of 0.75dB is as close to it as I can get, then I'm happy with that.

A difference 0.75dB should be as far as you can get.

I ABXed about a 5 second sample of some music at normal volume and the volume reduced by 0.75dB.  My score was 9/10 but I'm not sure what would happen if I were to ABX an entire song, which is probably what I should be doing considering the idea is to match the perceived levels of tracks, not snippets.

All I can say is try it for yourself.  If you aren't satisfied with a resolution of -1.5, adjust your source files with replaygain or wavegain before encoding them to mp3.

Now I have some questions:
Isn't the reference level of replaygain 83dB while mp3gain is 89dB?
If so, why the difference?
Or is it that the default for mp3gain is 89dB but it still puts the difference in the ape tag relative to 83dB?

MP3 Normalizers

Reply #26 –
Quote from: greynol on
I ABXed about a 5 second sample of some music at normal volume and the volume reduced by 0.75dB.  My score was 9/10 but I'm not sure what would happen if I were to ABX an entire song, which is probably what I should be doing considering the idea is to match the perceived levels of tracks, not snippets.


The real life situation is even harder to ABX: the different levels are associated with different tracks

Quote
Isn't the reference level of replaygain 83dB while mp3gain is 89dB?
If so, why the difference?


http://www.mars.org/mailman/public/mad-dev...ary/000993.html

MP3 Normalizers

Reply #27 –
Quote from: greynol on
I seriously doubt that you'll be able to hear the difference considering that it should be no more than 0.75dB.

Not quite, for example:

Target: 89 dB   

Track A: 91.2 dB -> possible corrections -1.5 dB (89.7 db) and -3 dB (88.2 dB) -> -1.5 dB is closer ->  89.7 dB

Track B: 89.8 dB -> possible correction -1.5 dB (88.3 db) -> -1.5 dB is closer than uncorrected ->  88.3 dB

Difference: 89.7 - 88.3 dB = 1.4 dB

MP3 Normalizers

Reply #28 –
Quote from: Alex B on
Quote from: greynol on
I seriously doubt that you'll be able to hear the difference considering that it should be no more than 0.75dB.

Not quite, for example:

Target: 89 dB   

Track A: 91.2 dB -> possible corrections -1.5 dB (89.7 db) and -3 dB (88.2 dB) -> -1.5 dB is closer ->  89.7 dB

Track B: 89.8 dB -> possible correction -1.5 dB (88.3 db) -> -1.5 dB is closer than uncorrected ->  88.3 dB

Difference: 89.7 - 88.3 dB = 1.4 dB

0.75dB is still as far as you can deviate from the reference that papadoc wanted to exactly match.

Thanks for the link Patsoe; and you are absolutely right: it should be more difficult to notice a difference between the level of different tracks even if the difference is 1.49dB.

MP3 Normalizers

Reply #29 –
I can understand people who want to normalize. What good is replaygain when regular players like Winamp won't use it anyway? Average Joe does not want to be bothered with installing a plugin for Winamp to read it...

MP3 Normalizers

Reply #30 –
If you use MP3gain on MP3 files, then even though WinAmp's default MP3 plugin don't do replaygaining, the audio itself has been reversibly replaygained by modifying the "global gain". That's the benefit of MP3Gain versus ReplayGain a la foobar2000.

But WinAmp's Ogg Vorbis input plugin do support replaygaining. So, all is not lost.

MP3 Normalizers

Reply #31 –
Any downsides to using mp3gain over foobar gain for mp3 files?

MP3 Normalizers

Reply #32 –
I can't believe you haven't summarized the thread yet

Main drawback: Limited resolution (i.e. 1.5 dB steps).

Another drawback: You must choose between Album or Track mp3gaining before processing the files. Cannot have them both.

Potential drawback: If a stupid program botches the file's tags, the mp3gain becomes irreversible.

MP3 Normalizers

Reply #33 –
I use both systems.

First I run MP3Gain in album mode.

Album gain adjustment in MP3Gain makes different albums to have a more even overall volume level and usually prevents the files from clipping on decoding stage. Sometimes I adjust the reference level a bit if the album clips with the default value. 89 dB can be a too high reference level for highly dynamic music like classical symphonies.

I have found that normally the album gain mode is better also with a portable or car mp3 player because it keeps the quieter tracks as the artist has meant them to be. I don't want to make a quiet ballad to be as loud as a loud rock track.

Because the APE tags are problematic with some programs I have disabled the "undo" tag writing in MP3Gain. I have never had a reason to revert the files because album gain preserves the track volume level differences inside each album. (I have set MP3Gain to log the changes, but the huge log file seems to be completely redundant for me.)

Afterwards I scan the files again with foobar. Foobar calculates new replay gain values that are based on the adjusted files. This makes possible to use foobar's precise playback adjustments and occasionally switch to track gain mode on PC playback.

MP3 Normalizers

Reply #34 –
I just tested out a few songs with MP3gain before running my whole 3000 songs through it. It showed all my songs to be very off, yet when I had them all converted in track mode and burned onto a cd I wasn't very impressed and even compared it to a before mp3gain cd with the same songs. It didn't seem to correct volume differences all that much, and I think even made some of the tracks sound worse in terms of quality. I'm not sure if the volume differences are because of the quality differences, since on the cd I burned I had songs with bitrates ranging from 128 to 360. Yet some of the low bit rate songs sounded pretty good and some of the higer bitrate songs didn't sound so great. It's kind of depressing. I wish there was a way to make all my mp3s sound better. I don't have the time or money to redo my whole collection. Hardly any of them were ripped, they were all downloaded. I spend tons of money on high quality car stereo and home stereo equipment just to have my mp3s sound crappier than the standard radio stations.

MP3 Normalizers

Reply #36 –
I'd like to weigh in here.  I use a program called 'MP3TrimPro' to normalize volume levels on MP3's.  I highly recommend it because I have show it to be a conservative modification.  For instance, if I adjust volume up 3 db, then back down 3 db, I get a byte for byte duplicate of the original.  Or if I adjust up by 4.5, then up by 1.5, that file is a duplicate of a file that was taken from the original and raised 6 db all in one step.  Thus, it is lossless.  It can be used in batch mode.

There is also a program called WAVTrimPro, which also claims to normalize volume.  However, my tests show that it is lossy.  No doubt it requantizes, and that is why.  It's like changing the resolution on a digital picture, there is resampling and rounding errors.  By the way, please turn off any normalization any of you are using within EAC.  That is a WAV normalization, which as I just mentioned is lossy.  I showed that directly for the case of EAC as well.  Once EAC has normalized the WAV, there appears to be no getting the original back.  So don't let EAC do it, use MP3Gain or MP3TrimPro instead.

I think MP3TrimPro and that MP3Gain program people are talking about work about the same way.  I think there is some kind of gain/volume control field in an MP3 that can be adjusted without affecting the rest of the MP3.

I highly, HIGHLY don't recommend any 'normalizers' that compress dynamic range.  That's just ruining the music, darn rudely.


@Scarface7810:  Then again maybe MP3Gain isn't so great, I don't know I've never used it.  But you definitely want to check out a program called EncSpot (http://www.guerillasoft.co.uk/encspot/).  It goes through your directory of MP3's and identifies the encoder used on each, and rates their quality.  It even shows details on what Lame options were used, if applicable.  If you have a lot of MP3's that show up red in the display window, you may want to focus on downloading higher quality versions of those MP3's.  In particular, watch for Xing recordings to eliminate, try to replace with Lame encoded versions.

MP3 Normalizers

Reply #37 –
Quote from: Gilgamesh577 on
I'd like to weigh in here.  I use a program called 'MP3TrimPro' to normalize volume levels on MP3's.  I highly recommend it because I have show it to be a conservative modification.  For instance, if I adjust volume up 3 db, then back down 3 db, I get a byte for byte duplicate of the original.  Or if I adjust up by 4.5, then up by 1.5, that file is a duplicate of a file that was taken from the original and raised 6 db all in one step.  Thus, it is lossless.  It can be used in batch mode.

[--snip--]

I think MP3TrimPro and that MP3Gain program people are talking about work about the same way.  I think there is some kind of gain/volume control field in an MP3 that can be adjusted without affecting the rest of the MP3.
Seeing that the volume adjusts in 1.5 dB steps, I'd bet my balls they do it identically.

For a technical description of how MP3Gain works (and why the 1.5 dB steps), see this page on HA Wiki.

MP3 Normalizers

Reply #38 –
Quote from: pepoluan on
Seeing that the volume adjusts in 1.5 dB steps, I'd bet my balls they do it identically.


Although I agree the 1.5dB steppings strongly suggest that, I would still advise against betting your balls on anything.

MP3 Normalizers

Reply #39 –
Quote from: Patsoe on
Quote from: pepoluan on
Seeing that the volume adjusts in 1.5 dB steps, I'd bet my balls they do it identically.
Although I agree the 1.5dB steppings strongly suggest that, I would still advise against betting your balls on anything.

Hey, I got lots of balls lying around my home... billiard balls, bowling balls, soccer balls, beachballs...
 

MP3 Normalizers

Reply #40 –
I just wanted to point out that since foobar2000 0.9, you can scan files for their replaygain values as normal, but you can also apply the replaygain value directly to the MP3 data (like MP3Gain).

So if you use foobar, MP3Gain is now redundant.

Aside: My method is to apply replaygain values directly to the MP3 data (either using MP3Gain or foobar), then scan the files again so they have the replaygain tag. This saves me reaching for the volume control on my MP3 player, and when playing in a replaygain-aware player I can use track-gain if I'm shuffling playlists.