UTF-8 Characters Not Displayed Correctly with MKAs

[thomascheney]

When MKAs have special vowels (ú, etc.) or characters from languages that do not use the Roman alphabet (Japanese, etc.) in the tags, they are not displayed correctly. Tags that I generate on my own with the UTF-8 encoding or with other programs such as MP3Tag do not display correctly in Poweramp but do display correctly in all other programs I've tried.

Poweramp version: build-945-bundle-play Full Version 64 bit.

Phone: Samsung Galaxy S21

[andrewilley]

MKA is not a very common audio format (the container being more used for video) so perhaps you could send a sample file containing a few simple accented letters that don't work, either by uploading to Dropbox or Google Drive, or email to gpmaxmpz@gmail.com. Add a comment in here and DM @maxmp to say you've sent it (and where from) so he can look out for it.

[Edit] I just found an MKA file in my testing folder, and added a few special characters to its Title tag using MP3Tag ( " André ⓵⓶  ➊➋ " ) and while they work fine in foobar2000 and VLC on my PC, I also found that they do not show correctly in Poweramp, either in the Library or in Info/Tags. My other go-to tag editor program, TagScanner doesn't even read MKA files at all, so I can't cross-check it there.

Andre

[andrewilley]

I just looked at the tag data with a hex editor, and the Title tag containing special characters seems to be using the condensed form of UTF-8 where regular ASCII characters are coded using a single byte, then $C3A9 is used for é and $E2xxxx for ⓵, ⓶, etc. PA is treating the whole string (including $C3 and $E2) as simple 8-bit Extended ASCII though, as there is nothing at the start of the string to indicate otherwise.

TagName = TITLE
- Tag 0x05a3 (5 bytes):
  373294: 54 49 54 4c 45                                  [TITLE]
TagString = Title Here Andr.. ...... ......
- Tag 0x0487 (31 bytes):
  37329c: 54 69 74 6c 65 20 48 65 72 65 20 41 6e 64 72 c3 [Title Here Andr.]
  3732ac: a9 20 e2 93 b5 e2 93 b6 20 e2 9e 8a e2 9e 8b    [. ...... ......]

I'm not an expert on how mixed-modes should be coded so will have to leave this to Max to have a look at.

Andre

[maxmp]

@thomascheney could you please share such track with gpmaxmpz@gmail.com for tests? Thanks!

[andrewilley]

@maxmp I've sent you the sample file that I created (see my post above).

Andre

[maxmp]

@andrewilley thanks!

@thomascheney should be fixed in the next beta build - mka strings should be always processed as utf-8 now as standard says.

[thomascheney]

Hi, sorry, do you need me to send a sample still? I would certainly be willing to do that if you need it.

 

Edit: I attached one.

白色雑音.mka