RE: cx88 driver sound problems | |
| [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] | |
At 12:08 PM 6/1/05, Marko Punnar wrote:
>Hi!
>
<...>
>-
>
>I have same problem PAL-BG and Nicam.
>
>kernel 2.6.11
>=8<======8<=======
>Linux video capture interface: v1.00
>cx2388x v4l2 driver version 0.0.4 loaded
>cx2388x: snapshot date 2005-01-24
<...>
>With later snapshots I have no sound at all
>
>===========================
>Marko
>
Yes, nicam for pal BG has been disabled in recent driver versions, this
gets it back in "video4linux-20050310-100754.tar.gz"
***
Line 274 of cx88-tvaudio.c
change
> #if 0
to
> #if 1
Line 768 of cx88-tvaudio.c
change
>set_audio_standard_NICAM_L(core,0);
to
>set_audio_standard_NICAM(core);
***
I suggest also this to fix color problem:
Line 752 of cx88-core.c
change
>HLNotchFilter135PAL
to
>HLNotchFilter4xFsc
***
The sound doesn't work any better than in previous driver versions. There haven't been
any changes to the nicam-bg in the driver code. It seems that the interest of developing
the drivers for analog tv cards is dropping and most of the work is being done to
the DVB cards.
So if you have any suggestions what might be causing the initialization problem, plese tell.
I have been debugging this about 6 months and I am running out of ideas.
These have already been tested:
-disable dma for audio during "cx88_set_tvaudio" -> no effect
-disable all tv card dma during "cx88_set_tvaudio" -> no effect
-disable all tv card interrupts during "cx88_set_tvaudio" -> no effect
-do not start tvaudio thread at all -> no effect
-add delays in soft-reset and load-init parts of the audio chip initialization ->
sometimes the audio breaks every time, usually no effect (depending on the length
and position of the delay)
-do not start any dma at all -> no effect to sound, no image any more ( surprise :) )
-add debug output for nicam error registers -> values don't show any difference
between distorted sound or normal sound. They show difference if nicam signal
is weak or or no nicam at all and normal nicam signal.
-read data that has been written to the audio registers and compare it to the value
that was supposed to go there, if they differ write again -> no effect
- about 5 times I have checked the nicam values in cx88 driver and DScaler cx88 code,
no major differences found, some minor differences corrected -> no effect
I am getting worried that reason for the init failure is somewhere in the
fundamental differences in interrupt/task switch routines of windows and linux since
similar code works well in DScaler. It might also be that nicam decoder of cx88 gets
in the incorrect internal state and soft reset+register programming is not enough to
get it back to normal. But if it is so, then why repeating soft reset+regs helps? And
why DScaler works?
---
SK
-------------------------------------------------------------------------
: Money isn't everything, : Sami Kiiskilä : Nothing :
: . . . there is also : P. 03-3180690 / 050-46 23 215 : is :
: H i g h - T e c h ! ! : http://www.cs.tut.fi/~kiiskils: impossible. :
-------------------------------------------------------------------------
--
Unsubscribe mailto:video4linux-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list
|
