[Bug 776038] audioringbuffer: do not require 4 byte multiples for encoded audio

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=776038

Vincent Penquerc'h <vincent.penquerch at collabora.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vincent.penquerch at collabora
                   |                            |.co.uk

--- Comment #2 from Vincent Penquerc'h <vincent.penquerch at collabora.co.uk> ---
ocrete did:

19:26 < v> The iffy change is changing bpf to 1. But I think it's actually
OK...
19:26 < ocrete> think you can make a cleaned up mp3 only patch and submit it
through phab ?
19:26 < v> Not sure if it's rationalization thoug h:)
19:27 < ocrete> why change bpf to 1 ?
19:27 < v> Even if it's possibly buggy ?
19:27 < ocrete> I didnt see the bpf change
19:27 < ocrete> why is that ?
19:27 < v> bpf to 1 otherwise baseaudiosink says "wrong size!" and error sout.
19:28 < ocrete> because the input is not in 4-byte quanta ?
19:28 < v> For mp3, it was always N bytes, with N about 400 or so, and
divisible by 4. Except for the first packet, which was one byte 
           less.
19:28 < ocrete> the bpf bit worries me.. I'm not sure what pulsesink expects
(or what bluetooth expects)
19:28 < v> (coming from mpegaudioparse)
19:28 < ocrete> oh from mpegaudioparse
19:28 < ocrete> hmm
19:29 < ocrete> would possibly make sense to have an upstream patch and get
some more feedback
19:29  * ocrete as no idea about the secrets of the mp3 format ;)
19:29 < v> Yes. It's encoded data after all, so bytes-per-frame has no real
meaning...
19:29 < v> I guess I'll do that.
19:29 < ocrete> all of the encoded formatts seem to have bpf=4
19:30 < v> Could be cut and paste. I'd totally do that :D
19:30 < v> And for Vorbis it's clear-cut wrong.
19:30 < v> Assuming it really is what I think it is.
19:41 < ocrete> the origin of that 4 is
19:41 < ocrete>      spec->type = GST_BUFTYPE_MPEG;
19:41 < ocrete> -    spec->format = GST_MPEG;
19:41 < ocrete> -    spec->width = 16;
19:41 < ocrete> -    spec->depth = 16;
19:41 < ocrete> -    spec->channels = 2;
19:41 < ocrete> +    spec->info.bpf = 4;
19:42 < ocrete> I think it'S completely made up

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.