[ATrpms-users] 5.1 Audio Channels Mixed Up (was vdpau/bijou)

Mark Goldberg marklgoldberg at gmail.com
Sun Apr 19 17:28:54 CEST 2009

Yeechang Lee <ylee[at]pobox.com> said:

>Mark Goldberg <marklgoldberg[at]gmail.com> says:
>> Setting Max Audio Channels to Stereo fixes the problem but at least
>> in 0.21-fixes is also shuts off all audio to the rear speakers.
>All? Even from a broadcast with native 5.1 audio? That is indeed

Yes, it shuts off all output to the rear speakers even for native 5.1 audio,
both in 0.21-fixes and bijou.

>> For shows that are 5.1, I want to have the rear speakers
>> working. 0.21-fixes does have the same controls in setup. What is
>> different about the upmixer #5900?
>The page isn't quite the same thing; without the #5900 upmixer a few
>of the other controls aren't there.

I don't see any difference in the setup page.

>The short answer is that the upmixer in 0.21 without the #5900 patch
>plain doesn't work.
>> I need to do more studying, but I'm thinking there really is a bug
>> in #5900. I have seen postings that channel mapping with DTS
>> sources is wrong. This is similar. It would help alot to have a
>> test file with audio on each channel.

I have not found one yet. I could create one, but it seems that there are
issues getting it right with various tools and that would add another
variable in whatever creates the file.

>Yes, it would. Is DTS audio even used in broadcasts? I thought it was
>all Dolby Digital. And speaking of Dolby, on my system the AC3
>passthrough over optical has always worked quite nicely (before and
>after #5900) and I've, to the best of my knowledge, never had a
>mapping or other issue, other than the peculiarity with a couple of
>ripped DVDs I mentioned in the other audio-on-MythTV thread here.

Broadcast is always AC3.

I created a channel mapping as follows:

pcm.bijou {
  type route
  slave.pcm surround40
  slave.channels 4
  ttable.0.0 1
  ttable.1.0 1
  ttable.5.0 1
  ttable.2.1 1
  ttable.1.1 1
  ttable.5.1 1
  ttable.3.2 1
  ttable.4.3 1

Alsa Mapping is as follows:

0 FL
1 FR
2 RL
3 RR
4 C

One reference (Table 5.8 in http://www.atsc.org/standards/a_52b.pdf
audio coding mode 3/2) says AC3 is

0 FL
1 C
2 FR
3 RL
4 RR
5 LFE (not explicitly stated in table 5.8 but after the rest)

and the above pcm.bijou maps to that and seems to work.

So I am guessing that the analog outputs of the new upscaler
are in AC3 order rather than Alsa order. If you are using a
digital output, it probably will work fine.


More information about the atrpms-users mailing list