[mythtv-users] Inverted chroma on PVR-250

[Joseph A. Caputo]

On Saturday 28 February 2004 13:40, Bill Hoover wrote:
> (I sent this to the ivtv list earlier this week, but since I got no
> responses I figured I'd see if anyone on this list has run into a
> similar problem).
>
> Earlier this week I ran into a problem with inverted chroma on a
> video capture from my PVR-250.  The card is in a MythTV system
> running the latest Myth code from CVS (well last weekend). 
> Everything has been working fine, but then on Wednesday night the
> recordings on cardid 1 from 21:00 to 22:00 and 22:00 to 23:00 (on
> different channels) had inverted colors (luma looked right, but faces
> were blue, etc.).  Then the recording it made (also with cardid 1) at
> 23:00 (on a third channel) was fine.
>
> All captures are from an NTSC signal through the tuner.  720x480,
>
> I'm hoping someone has an idea as to what could have caused this, so
> I can change things to avoid it in the future.
>
> As I said, the system is running MythTV from CVS from a week ago, so
> that is that software that is initializing the capture parameters,
> etc. There are three cards in the system.
> Cardid 1 is actually /dev/video2, a PVR-250 rev2.
> /dev/video1 is also a PVR-250 rev2, while /dev/video0 is an original
> PVR-250.
> The system is an FC1-based system.  Kernel is
> 2.4.22-1.2166.nptl_35.rhfc1.at and ivtv is from the
> ivtv-0.1.9-24.rhfc1.at.i386.rpm from the versions packaged at ATrpms.
>
> I'm currently using the 21288 version of the firmware.
>
> The log messages from initializing this card are:
>
> Feb 24 23:37:38 tvix kernel: ivtv: Found an iTVC16 based chip
> Feb 24 23:37:38 tvix kernel: PCI: Found IRQ 3 for device 03:03.0
> Feb 24 23:37:38 tvix kernel: ivtv: Unreasonably low latency timer,
> setting to 64 (was 32) Feb 24 23:37:38 tvix kernel: msp34xx: init:
> chip=MSP3448W-A2, has NICAM support Feb 24 23:37:38 tvix kernel:
> msp3410: daemon started
> Feb 24 23:37:38 tvix kernel: saa7114.c: starting probe for adapter
> ivtv i2c driver #2 (0x10005) Feb 24 23:37:38 tvix kernel: saa7114.c:
> detecting saa7114 client on address 0x42Feb 24 23:37:38 tvix kernel:
> saa7115.c: writing init values Feb 24 23:37:38 tvix kernel: status:
> (1E) 0x8c, (1F) 0x20
> Feb 24 23:37:38 tvix kernel: tuner: chip found @ 0xc2
> Feb 24 23:37:38 tvix kernel: tuner(bttv): type forced to 17 (Philips
> NTSC_M (MK2)) [insmod] Feb 24 23:37:38 tvix kernel: ivtv: Encoder
> revision: 0x02040011 Feb 24 23:37:38 tvix kernel: tuner: type already
> set (17)
> Feb 24 23:37:38 tvix kernel: ivtv: Registered v4l2 device, minor 2
> Feb 24 23:37:38 tvix kernel: ivtv: Registered v4l2 device, minor 34
> Feb 24 23:37:38 tvix kernel: ivtv: Registered v4l2 device, minor 226
> Feb 24 23:37:38 tvix kernel: ivtv: loaded
>
> Any suggestions as to what causes this, and how to eliminate the
> possibility of it happening in the future, would be greatly
> appreciated.


Hmmm, not sure but I *might* be seeing this on one of my M-179 cards.  
I'm not really sure because I only just installed both of them, so I 
don't know if this is a recent problem.

I have a single frontend/backend with 2 M-179s.  On the first card (/
dev/video0), I get really screwed up color.  Faces appear inverted, and 
there is also some color banding in the image.  Using Tuner0 as input.  
I've played with the coax connection as much as I could.  Running FC1 
straight from Jarod's guide, with the latest ATrpms kernel & ivtv 
packages.  My initial thought was that I was getting RF interference, 
as this card is closer to my wlan card than the other one.  I was going 
to try moving it to a different slot sometime this week to see if that 
helps.

I'm hoping it's either RF interference or a Myth/ivtv problem (though I 
somehow doubt the latter), because other wise it probably means I have 
a bum card :-(

-JAC