>Wouldn't it make more sense to set it in decoder init?
>Anton Khirnov

Yes, you're right, in the actual code, this would seem the best place.
But, I have in mind to add support for s337m in wav (I am just getting back on 
this work I started some months ago).
Take a 48k wav with dolby_e in it which rate is 44800 for example (this is my 
test case actually).
Here is the diff (+frame_size set on init / -frame_size set on decode frame)
-0,          0,          0,     1920,    11496, 0x05a9c147
-0,       1920,       1920,     1920,    11496, 0x1d44d2b4
-0,       3840,       3840,     1920,    11496, 0x4e078953
-0,       5760,       5760,     1920,    11496, 0x1c73b1a1
-0,       7680,       7680,     1920,    11262, 0xfa179fc8
+0,          0,          0,     1792,    11496, 0x05a9c147
+0,       1792,       1792,     1920,    11496, 0x1d44d2b4
+0,       3712,       3712,     1920,    11496, 0x4e078953
+0,       5632,       5632,     1920,    11496, 0x1c73b1a1
+0,       7552,       7552,     1920,    11262, 0xfa179fc8

The problem is the duration of the first audio frame. The mismatch is because 
of the difference of sample_rate between the mux/wav and the stream/dolby_e.
Anyway, my proposal is that I will send my complete patchset including this 
patch, so you can review the global project and continue the discussion.
I think the way I have integrated s377m in wav is ok, but the lack of a parser 
for dolby_e (and thus early sample rate detection) makes things a little bit 
tricky.

Nicolas
_______________________________________________
ffmpeg-devel mailing list
[email protected]
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".

Reply via email to