Invalid duration about tinytag HOT 6 CLOSED

I'm seeing what looks like 2X with an mp3 on this code path:

if frames == 0 and ID3._USE_XING_HEADER:

mp3 generated by lame, for a 48k ABR mp3. When I encode at 96k ABR, it's not off by
a factor of 2. So it appears to be bit rate, not channels, which throws off the interpretation
of the xing header. My wife's a musician, I can encode one of her tracks (no license issues)
if you need some example mp3's.

from tinytag.

In any case, in _parse_xing_header() the TOC values are bytes, not 4-byte big endian values.
But that doesn't change the duration calculation...

from tinytag.

Wowzeee, It now takes me a year to answer github issues 😢

Anyway, if you still have any of the files at hand, please post them, so this issue can be reproduced!

from tinytag.

Sure... compare soxi on https://vsta.org/andy/pickup/qz9k29290006-48.mp3 versus what tinytag says for duration. (It's my wife's track, so copyright/music licensing is no problem, enjoy!)

BTW, welcome back.

from tinytag.

Hey @vandys , thanks for the sample file. That and your comment about an off by a factor of 2 lead me to some investigation about the mp3 format. It turns out that the mp3 frames have 1152 samples per channel. My test for the xing header was using a mono-channel file, so your stereo file was off by a factor of 2 🤦

Anyway I have fixed the issue an your file is now correctly read. I am using a truncated version of your file for the test-suite now.

Thanks for your hints and your patience! You can use the version here in git until the next version is released containing this fix.

Have a great day in these weird times, maybe you'd fancy some cake? here you go: 🥮

from tinytag.

Thanks much!

from tinytag.