[vdr] Re: [PATCH] AC3overDVB 0.1.0

[s.huelswitt@gmx.de (Stefan Huelswitt)]

On 24 Jan 2003 "Reinhard Walter Buchner" <rw.buchner@freenet.de> wrote:

>> > When I try the above "can'ts" I get a noise similar to white
>> > noise except it's in a staccato mode. It seems to me that
>> 
>> The noise you're hear, is the raw digital AC3 data played as PCM
>> audio. This indicates that your DD receiver doesn't detected the
>> encapsulated data.
> 
> Thats clear, but as mentioned there is NO DD data in the "don't
> work" cases (except for the one recording, which as far as I know
> SHOULD contain DD data). Pure DD DVDs playback just fine, so
> it shouldn't be my amp that is causing the problem. As I said:

OK, I did at lot of test now, but I cannot reproduce any of this:
-I can replay recordings with DD sound
-I can replay recordings which certainly doesn't contain DD.
-Beside some "noise" right at the start of DD data (before the
receiver has detected DD), ther is no "stacato" sound.
-It doesn't matter if I start the recording with red or green
button.
-In DD recordings I can switch between DD & stereo sound with
green in main menu.

Did you tried the "old" patch from Jan Rieger? Did it worked?

> I'll have to investigate this further. Will try your debug thoughts.
> Maybe your mod is removing streams where it should not be, so
> all that is left is the capsule data ;o)) Obviously you couldn't blame
> the amp for that.

There may be a bug, yes. All I know, that all my recordings are
played fine (did not tried DVD yet).

>> message on console every time a DD packet is detected and
>> encapsulated.
> 
> I forgot to mention that I get messages (from mind, as I am not
> in front of VDR right now) on theVDR tty: error stream 1800
> != 1899 (example). Will log the true output next time. As far as

Probably "encapsulate: length mismatch 1800 != 1899".
I have seen this with some recordings too, although they play
fine. Most of the time the differenze is 6 bytes.

As I said at another place, the stream analyser isn't very robust
and have some race conditions. I'm working at it currently.

> Hmm,  I'll have to give your mod a thorough look at, but I think
> if one turns off DD / DTS in the setup VDR should behave exactly
> as a vanilla VDR (and use the vanilla routines).

It does, or at least it should. ;-)

I see two possible failure points:

1. the stream is "broken", well, not corrupted, but different
from the streams I have here. Different in a way that the stream
analyser and the encapsulation fails. Would it be possible to
upload some stream data (at least some MB from the beginning) for
testing?

2. the receiver fails to detect the DD encapsulation. I know from
1.0.4 times, that some receivers are simply not able to do this.
In the last patch I changed the encapsulation header. You could
try the version by replacing in multichannelaudio.c 
cAudioEncapsulator::NewFrame() below "// build LPCM header":
0xFF by 0x00 and
0x81 by 0x00
This is the old sequence.

-- 
Stefan Huelswitt
huels@iname.com  | http://home.pages.de/~nathan


-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe vdr" as subject.