[pvrusb2] Fuzzy Video using Hauppauge WinTV PVR USB2

[Mike Isely]

On Thu, 22 Dec 2005, cardboil wrote:

> Mike:
>
> I tried a couple of settings based on some instructions from
> http://ivtvdriver.org/viewcvs/ivtv/trunk/doc/README.tda9887?rev=2685)
>
> Here is what my ivtv section in modprobe.conf looks like:
>   # ivtv modules setup
>   alias char-major-81 videodev
>   alias char-major-81-0 ivtv
>   options ivtv ivtv_std=2 tda9887=0 tuner=38
>   options tda9887 pal=b port1=0 port2=0
>   options tuner pal=b
>
> But, this didn't do me any good.  My current symptoms are crackling or no
> audio on some channels and bad picture quality on other.  Of course, there
> are some channels which have perfectly good video and audio.

I have a suspicion about another thing which might be exacerbating this.

There is a piece of hacked up code in the pvrusb2 driver which executes 
some at-the-time unknown I2C operations.  This dates from Aurelien's old 
version of this driver.  When I converted the driver to use various well 
known V4L modules in preference to hard coded logic, I was never able to 
figure out what those mystery I2C operations corresponded to.  So I 
preserved them, thinking at the time that if it wasn't broke to not "fix 
it".  I have just recently discovered that these mystery I2C operations 
were in fact attempts to operate the same part that the tda9887 driver is 
operating - clearly that is what these operations were for.  I've 
therefore already checked in a change to the driver to remove this hack, 
but I haven't released a new snapshot yet because right at the same time 
when I removed the hack I noticed that tuning still wasn't working very 
robustly and I ended up spending the rest of the day trying to understand 
the problem.

The next snapshot I release will have that straightened out, but 
unfortunately that doesn't help you very much at the moment.  However if 
you are willing to do a bit of editing, look inside pvrusb2-tuner.c at the 
code which configures the tuner's idea of the video standard and you'll 
see a hackish piece of crap (with a suitable comment attached) that is 
doing these explicit I2C commands.  Comment that out and see if it helps. 
It's conceivable that this might be getting in the way of the operation of 
tda9887.ko.  If you try this, please tell me what happens.

I glanced through the system log info you posted and didn't see anything 
that immediately suggested a problem.  (The ioctl failures are not 
unusual.)

   -Mike

-- 
                         |         Mike Isely          |     PGP fingerprint
      Spammers Die!!     |                             | 03 54 43 4D 75 E5 CC 92
                         |   isely @ pobox (dot) com   | 71 16 01 E2 B5 F5 C1 E8
                         |                             |