.wav files detected as audio/wave when maybe they should be audio/wav

[simonw]

As far as I can tell, the "correct" type to return for a .wav file (with 52 49 46 46 xx xx xx xx 57 41 56 45 66 6d 74 20 is audio/wav - but this library returns audio/wave.

I got very confused looking through the code because I came across these two lines:

puremagic/puremagic/magic_data.json

Line 103 in 763349e

["57415645", 8, ".wav", "audio/wave", "Waveform Audio File Format"],

puremagic/puremagic/magic_data.json

Line 1118 in 763349e

["52494646", 0, ".wav", "audio/wav", "Resource Interchange File Format"],

I've found it hard to research the correct resolution though, as both audio/wav and audio/wave are entirely missing from what I thought was the official RFC for these! https://www.iana.org/assignments/media-types/media-types.xhtml#audio

MDN lists audio/wav https://developer.mozilla.org/en-US/docs/Web/HTTP/MIME_types/Common_types

I'm not sure there is a correct answer to this question.

[simonw]

Tried this:

python -c 'import puremagic, pprint, sys; pprint.pprint(puremagic.magic_stream(open(sys.argv[-1], "rb")))' output.wav

And got:

[PureMagicWithConfidence(byte_match=b'RIFFH\xe0\x02\x00WAVE', offset=8, extension='.wav', mime_type='audio/wave', name='Waveform Audio File Format', confidence=0.8),
 PureMagicWithConfidence(byte_match=b'WAVEfmt ', offset=8, extension='.wav', mime_type='audio/x-wav', name='Windows audio file ', confidence=0.8),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.4xm', mime_type='', name='4X Movie video', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.cdr', mime_type='', name='CorelDraw document', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.avi', mime_type='video/avi', name='Resource Interchange File Format', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.cda', mime_type='', name='Resource Interchange File Format', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.qcp', mime_type='audio/vnd.qcelp', name='Resource Interchange File Format', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.rmi', mime_type='audio/mid', name='Resource Interchange File Format', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.wav', mime_type='audio/wav', name='Resource Interchange File Format', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.ds4', mime_type='', name='Micrografx Designer graphic', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.ani', mime_type='application/x-navi-animation', name='Windows animated cursor', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.dat', mime_type='video/mpeg', name='Video CD MPEG movie', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.cmx', mime_type='', name='Corel Presentation Exchange metadata', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'RIFF', offset=0, extension='.webp', mime_type='image/webp', name='RIFF WebP', confidence=0.4),
 PureMagicWithConfidence(byte_match=b'WAVE', offset=8, extension='.wav', mime_type='audio/x-wav', name='WAV audio', confidence=0.4)]

[simonw]

I had a similar issue on llm-gemini where puremagic was returning audio/mpeg for MP3 files but the Gemini AI wanted audio/mp3:

• Solve multi-modal models with a new concept of "attachments" simonw/llm#587 (comment)

It turned out in that case puremagic was correct and Gemini was wrong - the official mimetype for MP3 is indeed audio/mpeg.

[NebularNerd]

Tried this:

python -c 'import puremagic, pprint, sys; pprint.pprint(puremagic.magic_stream(open(sys.argv[-1], "rb")))' output.wav

And got.....

.wav files are part of the RIFF family, so their detection is currently split across the initial RIFF at byte 0, then the WAVE at byte 8 (IFF and PK (aka Zip) based formats are similar). Due to the size of the .json some duplication and occasional hiccups appear. @cdgriffith is working on a v2 update which will improve detection and collectively we should be able to remove duplication. Looking at the results, next time I do a PR I personally would look to roll the standalone WAVEfmt entry into RIFF family to reduce duplication but like you say which MIMETYPE should we go for?

The MIMETYPE's are a nightmare, according to The Library of Congress: .wav page, both are valid and they even mention another few.