ReplayGain tag information inside mp3 files

[Elias Villalobos]

This isn't really a bug report, but rather a question related to how Poweramp works.

I have an mp3 file to which I have correctly inserted the TXXX tag: REPLAYGAIN_TRACK_GAIN with a value set to -10 dB

I played around with the ReplayGain settings.  But the problem is that Poweramp doesn't seem to recognize the "REPLAYGAIN_TRACK_GAIN" tag when it is inside an mp3 file.

According to Poweramp, what is the correct ID3 tag to store ReplayGain information inside mp3 files?
 

[andrewilley]

First, it's worth checking that you actually have ReplayGain enabled in Poweramp, in Settings=>Audio=>Replay Gain.

Poweramp reads standard MP3 ID3 Replay Gain comments via the 'TXXX' tag, as used by other software. I use foobar2000 to batch create my RG data, but MP3Tag and TagScanner both show the contents too in their Extended / Advanced tags sections. Lower case entries used to be preferred for some tags, but PA is now case-insensitive as both variants exist in the wild.

For example, as underlying ID3 tags:

TXXX: (replaygain_album_gain) -3.20 dB
TXXX: (replaygain_album_peak) 0.999270
TXXX: (replaygain_track_gain) -5.02 dB
TXXX: (replaygain_track_peak) 0.918534

You can long-press on a song that is currently playing and view its Info/Tags to check that Poweramp has read the tags correctly:

For a bit more discussion on this, see this thread:

Andre

[Elias Villalobos]

@andrewilley Hello. Thanks for answering. Unfortunately I can't use foobar to add tags to files because I don't have a desktop device, and the Android version of foobar doesn't seem to offer RG tag insertion.

However, I am developing my own application capable of inserting ReplayGain tags in Audio files. I added all the tags that you suggested, but still did not achieve the expected result.

It is worth mentioning that I have a file with an opus extension, to which I inserted the R128_TRACK_GAIN tag with a value set to -1500, and in this case Poweramp did successfully recognize said tag, thus obtaining the expected result for ReplayGain.

Can you send me the mp3 file that you showed me on the screenshot?

My email is: eliasjose3112@gmail.com

 

[andrewilley]

The tag data from the discussed file is:

UserDefinedText = (replaygain_album_gain) -3.20 dB
- Tag 'TXXX' (32 bytes):
    0025: 00 72 65 70 6c 61 79 67 61 69 6e 5f 61 6c 62 75 [.replaygain_albu]
    0035: 6d 5f 67 61 69 6e 00 2d 33 2e 32 30 20 64 42 00 [m_gain.-3.20 dB.]
UserDefinedText = (replaygain_album_peak) 0.999270
- Tag 'TXXX' (32 bytes):
    004f: 00 72 65 70 6c 61 79 67 61 69 6e 5f 61 6c 62 75 [.replaygain_albu]
    005f: 6d 5f 70 65 61 6b 00 30 2e 39 39 39 32 37 30 00 [m_peak.0.999270.]
UserDefinedText = (replaygain_track_gain) -5.02 dB
- Tag 'TXXX' (32 bytes):
    0079: 00 72 65 70 6c 61 79 67 61 69 6e 5f 74 72 61 63 [.replaygain_trac]
    0089: 6b 5f 67 61 69 6e 00 2d 35 2e 30 32 20 64 42 00 [k_gain.-5.02 dB.]
UserDefinedText = (replaygain_track_peak) 0.918534
- Tag 'TXXX' (32 bytes):
    00a3: 00 72 65 70 6c 61 79 67 61 69 6e 5f 74 72 61 63 [.replaygain_trac]
    00b3: 6b 5f 70 65 61 6b 00 30 2e 39 31 38 35 33 34 00 [k_peak.0.918534.]

MP3 ID3 tags in general are pretty well documented though.

Andre

[Fitzian]

Last time I checked android version of fb2k did have replay gain scanning and insertion into tags. Not the full settings as desktop would offer, but it's certainly there.

[andrewilley]

48 minutes ago, Fitzian said:

Last time I checked android version of fb2k did have replay gain scanning and insertion into tags. 

Foobar2000 does support this, yes. In its Tools>Utilities>Replay Gain Scanner.

Andre

[Elias Villalobos]

23 hours ago, andrewilley said:

Foobar2000 does support this, yes. In its Tools>Utilities>Replay Gain Scanner.

Andre

@andrewilley. Thank you very much for helping me with this.

I finally figured out what was the cause of my problem, and I already solved it with the help of foobar2000.

I am using Poweramp v3 build 945 (32 bits)

 

In my case, when dealing with mp3 files, Poweramp recognizes ReplayGain tags only if the keys are written in lowercase letters, which contradicts what you stated in his first comment.

My problem was because I was writing the keys of the RG tags in uppercase letters.

 

Another quirk (obviously a bug) that I noticed when PA parses RG tags in mp3 files, is that PA will completely ignore all RG tags if there are any duplicates similar to the following:

TXXX: (REPLAYGAIN_TRACK_GAIN) -5.02 dB

TXXX: (replaygain_track_gain) -5.02 dB

 

However, the following duplicate is allowed by PA:

TXXX: (replaygain_track_gain) -5.02 dB

TXXX: (REPLAYGAIN_TRACK_GAIN) -5.02 dB

 

[andrewilley]

The issue of (replay_album_gain) not being recognised if the string was quoted in UPPER case was resolved back in October last year. Track Gain always worked in both upper or lower case though. What version of Poweramp are you using? I've just checked in build 956, and it works fine for me with a test file where all four items are quoted in UPPER case.

MP3 files should not contain multiple instances of the same ID3 tags, but I agree that simply accepting the first instance and ignoring any incorrect repeats would be a more elegant way of handling non-standard content (especially as TXXX does allow for multiple instances, but for different uses).

Andre

[Elias Villalobos]

18 minutes ago, andrewilley said:

The issue of (replay_album_gain) not being recognised if the string was quoted in UPPER case was resolved back in October last year. Track Gain always worked in both upper or lower case though. What version of Poweramp are you using? I've just checked in build 956, and it works fine for me with a test file where all four items are quoted in UPPER case.

MP3 files should not contain multiple instances of the same ID3 tags, but I agree that simply accepting the first instance and ignoring any incorrect repeats would be a more elegant way of handling non-standard content (especially as TXXX does allow for multiple instances, but for different uses).

Andre

my version of Poweramp:

[Elias Villalobos]

53 minutes ago, andrewilley said:

The issue of (replay_album_gain) not being recognised if the string was quoted in UPPER case was resolved back in October last year. Track Gain always worked in both upper or lower case though. What version of Poweramp are you using? I've just checked in build 956, and it works fine for me with a test file where all four items are quoted in UPPER case.

MP3 files should not contain multiple instances of the same ID3 tags, but I agree that simply accepting the first instance and ignoring any incorrect repeats would be a more elegant way of handling non-standard content (especially as TXXX does allow for multiple instances, but for different uses).

Andre

I downloaded the latest stable version of Poweramp from this website, and the issue I described is now resolved. From now on PA also recognizes RG tags whose key is in uppercase.

 

Nevertheless. My application will need to continue to write those tags in lowercase, as I decided to do the same tests with other popular music player apps. and I found that the "jetAudio Music Player" app has the same behavior that I described in my previous comment.

[andrewilley]

The discussion the issue of ignored upper case RG album tags took place back in early October, and Max said it would be fixed in the next beta builds. I guess 945 was already finalised by that point, so the fix didn't appear until the next round of releases (948 onward).

Glad it's working for you now that you've updated to a recent build though (I did mention that the change was fairly recent in my first reply by the way). Using lower-case does seem to be the most widely occurring variant in the wild, that's what foobar200 uses when you use it to update RG tags for example, and edits made in TagScanner and MP3Tag result in lower case too. Even though the Hydrogenaudio wiki uses upper case for all ID3 tags as its house style.

Does the latest release have the same issue you described with non-standard files that contain two instances of the same tag by the way? I'm with you that although it's incorrect, one tag at least should be honoured.

Andre