[mythtv-users] Soft/garbled audio with SVN

[Yeechang Lee]

For those experiencing problems with audio while running SVN (such as
Curtis Stanford, at
<URL:http://www.gossamer-threads.com/lists/mythtv/users/318588#318588>),
the cause may be the new audio upmix feature (ticket #1104) recently
merged into trunk.

My frontend is connected via S/PDIF to a 5.1 receiver, with AC3 and
DTS passthrough enabled. Here's my experience with various recordings'
audio sources with #1104 and the Active Simple or Active Linear
upmix settings:

* MPEG-2, AC3 audio: No issues[1].
* MPEG-2, non-AC3 audio: No issues.
* MPEG-4, non-AC3 audio (generated via mythtranscode from MPEG-2 and,
  usually, AC3): Very soft music. Voices are almost inaudible even at
  very high volume.

Shifting to Passive seems to fix the issue with the MPEG-4
files. #1104 (the only place I can find any relevant documentation at
all on the topic) states that the Active Simple and Active Linear
settings require more CPU. My frontend is a Pentium 3.0GHz; who'd have
figured that audio decoding (as opposed to video) would be so
CPU-intensive? In any case, I've uploaded a patch (#4775) that may
help in avoiding some future confusion.

[1] As far as I can tell; I'm no audiophile and perhaps I'm missing
some nuance those with keener ears would pick up instantly.

-- 
Yeechang Lee <ylee at pobox.com> | San Francisco CA US