Questions, PRs, FMOD

[drew-512]

Hi Fabian and friends -

I posted while back to cheer on faad2 development. I'm an indie dev and have been working hard on a forthcoming Unity 3D based media player (think VLC) that uses faad2 for AAC decoding (amongst other codecs). I'm currently working on a FMOD codec plugin that wraps faad2 and it's coming along!

Questions if you have a minute:

-- Are there significant tradeoffs of FAAD_FMT_FLOAT vs FAAD_FMT_16BIT other than memory bandwidth? Of course, it's negligible on a PC/laptop these days, but how about mobile (armv7)? Is the audio quality difference from 16 bit even detectable by humans for most music? What are the pros/cons to the fixed point build?

-- My plan was to use NeAACDecInit2() with so that AudioSpecificConfig2() does the heavy lifting (unless there are advantages to NeAACDecInitDRM()) Are there advantages to using the Digital Radio Mondiale (DRM) path? Is it accurate that it has more robust error handling? Or is it moot with typical broadband data rates?

-- Is there interest for to submit a PR that adds a dir 'libfaad_fmod' (and CMakeLists.txt, etc) and builds out libs for each platform? I suspect there will be interest in it, and I thought it'd be a great way to say thanks for all the hard work that's gone into faad. I also ask because it may also bring attention.

Thanks,
Drew

[fabiangreffrath]

Dear Drew,

thank you for your kind words! As some kind of a disclaimer, I have not been involved with the development of the codec at all. I just got commit rights back then to integrate some Debian patches into the upstream code branch and now I am the last one left here.

-- Are there significant tradeoffs of FAAD_FMT_FLOAT vs FAAD_FMT_16BIT other than memory bandwidth? Of course, it's negligible on a PC/laptop these days, but how about mobile (armv7)? Is the audio quality difference from 16 bit even detectable by humans for most music? What are the pros/cons to the fixed point build?

I doubt there will be an audible difference between codecs using float or 16-bit fixed point. However, performance might be an issue on old archs without proper hardware float implementations (i.e. i386 and some very early arm processors).

-- My plan was to use NeAACDecInit2() with so that AudioSpecificConfig2() does the heavy lifting (unless there are advantages to NeAACDecInitDRM()) Are there advantages to using the Digital Radio Mondiale (DRM) path? Is it accurate that it has more robust error handling? Or is it moot with typical broadband data rates?

I can't tell for sure. I just know that we factored out the DRM enabled library into a separate build back then, because there were some compatibility issues with non-DRM use cases.

-- Is there interest for to submit a PR that adds a dir 'libfaad_fmod' (and CMakeLists.txt, etc) and builds out libs for each platform? I suspect there will be interest in it, and I thought it'd be a great way to say thanks for all the hard work that's gone into faad. I also ask because it may also bring attention.

Honestly, this code being part of this repository inevitably implies a certain degree of responsibility for maintenance etc. As I am already unable to reply to most requests posted here against the regular code, I doubt that adding another code tree will improve the situation. Also, I think FMOD is proprietary code and I wouldn't want to advertise its use (e.g. over openal-soft) here.

Thanks!

• Fabian

[drew-512]

Honestly, this code being part of this repository inevitably implies a certain degree of responsibility for maintenance etc. As I am already unable to reply to most requests posted here against the regular code, I doubt that adding another code tree will improve the situation. Also, I think FMOD is proprietary code and I wouldn't want to advertise its use (e.g. over openal-soft) here.

I figured as much, though the FMOD codec pluigin API allows devs to freely make and distribute plugins they make. You can see my discussion with them from a year back.

I'll post an update here in hopefully not long/ Anyone interested in AAC / MP4 on FMOD, lmk.

Andy

[drew-512]

After getting the fmod plugin wrapper all set up, and all the mov tag parsing all done, I was feeling pretty optimistic.

However, I'm sad to report that NeAACDecDecode() for both DRM build variants is erroring for DRM_ER_LC mdat blobs (object type 27). Search for "return 26" and you'll see where it's erroring -- I can't make any sense of what could be happening other than the DRM code there was made before the newer codec additions around 2001. I'm initing with NeAACDecInitDRM() and have checked everything is being passed in:

https://github.com/drew-512/faad2/tree/master/libfaad_fmod/src/main.cpp#L287

Unless there's a trick, I don't know how faad is usable is a world where so many consumer audio files are object type 27. Would appreciate any thoughts if anything comes to mind, and anyone is welcome to look at the code.

Happy to support anyone or explain anything to repro the issue, etc. Otherwise, I feel like I'm at a dead end.
Thanks,
Andy