Opened 13 years ago
Closed 13 years ago
#10488 closed Bug Report - General (Fixed)
HLS transcode broken for sources with ac3 audio
| Reported by: | Owned by: | beirdo | |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.25 |
| Component: | MythTV - Mythtranscode | Version: | Master Head |
| Severity: | medium | Keywords: | |
| Cc: | Ticket locked: | no |
Description
I did a git bisect and the guilty commit turned out to be:
0fe6e143faf4a987ec97203840d991de5578dbf2 is the first bad commit commit 0fe6e143faf4a987ec97203840d991de5578dbf2 Author: Gavin Hurlbut <ghurlbut@…> Date: Fri Mar 2 15:11:15 2012 -0800
Totally rework the AudioReencodeBuffer? part in transcode
Things are fine with non-ac3 recordings such as what you'd get from an ivtv recorder, it's just ATSC-type recordings that are faulty but nowadays that's all of my stuff.
Change History (5)
comment:1 Changed 13 years ago by
| Milestone: | unknown → 0.25 |
|---|---|
| Owner: | set to cpinkham |
| Status: | new → assigned |
| Version: | Unspecified → Master Head |
comment:2 Changed 13 years ago by
| Owner: | changed from cpinkham to beirdo |
|---|
comment:3 Changed 13 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | assigned → closed |
Fix HLS transcoding audio with non-MP3 input
The buffer splitting I had in place was slightly off, with the result of horrible sounding audio when transcoding from AC3. Feeding entire single frames of audio to the encoder is the trick, and finally, we are there.
Fixes #10488
Branch: master Changeset: 7fa50231618ecc8e45f75d424275f3d389c9c1b6
comment:4 Changed 13 years ago by
| Resolution: | fixed |
|---|---|
| Status: | closed → new |
I hate to say it, but the fix for DTS has rebroken AC3 audio in HLS. AC3 transcodes at any bitrate are comprehensible, but suffer from a constant "pulsing" of audio that makes it pretty much unlistenable. Reverting the DTS fix restores AC3 functionality, and re-breaks DTS.
:(
comment:5 Changed 13 years ago by
| Resolution: | → Fixed |
|---|---|
| Status: | new → closed |
Refixed in 8bf3888. Sorry about that. I think we should be good now.
Well, then we have some fixing to do. That commit will not be reverted, as it is necessary for the fifo mode in mythtranscode. I will work with cpinkham to get this working as needed.