[linux-dvb] FIXME: audio doesn't work on svideo/composite - hvr-1110 S-Video and Composite

hermann pitton hermann-pitton at arcor.de
Mon Sep 22 01:20:47 CEST 2008


Hi,

Am Sonntag, den 21.09.2008, 19:12 +0000 schrieb dabby bentam: 
> Thanks for your assistant.
> 
> Yes, that makes sense, as the TV sound can be turned on.
> 
> I'll try your suggestions, changing the code to LINE1.
> 
> Has anyone tried this - Mostly it is high (0x0200000) for the radio switch, only sometimes the
>  other way round. Then you must set .gpio = 0x0200000 for the TV input
>  with the same gpio_mask. < did you try this tom?
> 
> Thanks to all, i will play around with this, hopefully get a result. Is there anyway to find the exact gpio mask?
> 
> David

please don't top post. I continue below.

> ----------------------------------------
> > Subject: Re: [linux-dvb] FIXME: audio doesn't work on svideo/composite -	hvr-1110 S-Video and Composite
> > From: hermann-pitton at arcor.de
> > To: tomlohave at gmail.com
> > CC: db260179 at hotmail.com; linux-dvb at linuxtv.org; video4linux-list at redhat.com
> > Date: Sat, 20 Sep 2008 19:15:05 +0200
> > 
> > Hello,
> > 
> > Am Samstag, den 20.09.2008, 15:29 +0200 schrieb tomlohave at gmail.com:
> >> dabby bentam a écrit :
> >>> Hi Tom,
> >>>
> >>> Sorry to bother you like this in this way. I'm emailing because i have 
> >>> a Wintv hvr-1110 card. In the kernel sources it has
> >>>
> >>> FIXME: audio doesn't work on svideo/composite
> >>>
> >>> Do you know what is required to get this to work?
> >>>
> >>> Can i help in getting this to work?
> >>>
> >>> Thanks for your time
> >>>
> >>> David Bentham
> >>>
> >> 
> >> Hi,
> >> 
> >> first, sorry for my poor english
> >> 
> >> No sorry not sure what to do, i've try many configurations and 
> >> modifications in order to have svideo/composite audio but no result
> >> 
> >> you can try different parameter (see in sources, take example on other 
> >> cards)
> >> or ask on linuxtv mailing list
> >> I forward your precedant mail to this list
> >> 
> >> Cheers
> >> 
> >> Thomas
> >> 
> > 
> > there are different models of the HVR-1110 we can see on
> > hauppauge_eeprom down in saa7134-cards.c.
> > 
> > Some don't even have analog audio in.
> > 
> > In case it is there, if the current LINE2 stereo input pair doesn't
> > work, then it is LINE1.
> > 
> > If that still does not work, you can expect an external gpio driven mux
> > chip on the board. On cards with analog audio out it can be fairly easy
> > tested by unloading the driver. Default is then to loop the sound
> > through to the analog output of the card to the soundcard, but with the
> > driver unloaded you can't test, since you need saa7134-alsa for it
> > without having analog audio out connected on the card.
> > 
> > In that case regspy.exe from DScaler (deinterlace.sf.net) might help.
> > 
> > A work around could be to plug the external analog sound directly into
> > the sound card/chip and record from there.
> > 
> > Was the radio ever tested? Since the tda8275ac1 has only one RF signal
> > input pin and the card has two antenna connectors, we usually see that
> > these inputs are switched on gpio21 including the AutoGainControl and
> > using the 5.5MHz ceramic filter for radio input.
> > 
> > See the other cards with TUNER_PHILIPS_TDA8290.
> > 
> > Mostly it is high (0x0200000) for the radio switch, only sometimes the
> > other way round. Then you must set .gpio = 0x0200000 for the TV input
> > with the same gpio_mask.
> > 

For the TV versus radio switch the gpio mask is just 0x0200000.

With this mask only gpio pin 21 will change.

By default the driver will set it to 0, only for that input gpio =
0x0200000 is set in the card's config, it will change it to 1.
(masked writes, "modinfo saa7134" gpio_tracking=1)

If changing the amux to LINE1 doesn't help for s-video and composite,
testing with saa7134-alsa there and say sox, then regspy might help.

If you use the supplied m$ app, not DScaler, and save the logs for the
card in status unused, TV, composite/s-video and radio, we will see the
gpio mask in use there and likely changing gpio pins.

They often have more pins in the mask and also more that change,
but usually one of them will be what we need to add to the mask.

It looks like manufacturers group their devices together with a gpio
mask covering so far all of them, then subsets of devices do appear,
which still do more changes on the pins, then for a specific single
device would be needed.

With a high granularity for each device, based on a doubtless and
consistent eeprom detection, the gpio settings likely could be single
device specific too, if wanted.

But since we might see unrelated gpio changes there too, dunno, first of
all make sure that amux LINE1 does not resolve the problem.

Cheers,
Hermann







More information about the linux-dvb mailing list