[mythtv-users] artifacts in pvr-150 recordings, only recently

[jarpublic at gmail.com]

OK, I think I am narrowing down on things. I don't think it is because
I have too high of a bitrate. The artifacts are showing up on
recordings that have 5000 kbps avg. Also I haven't changed my
recording profiles at all and they were working fine at those settings
before this week. Turning all the fans on full speed didn't improve
anything. I was able to confirm visually that recordings were actually
fine through March 1st and then all of the recordings since the 2nd
have artifacts. Then I searched through the backend log more carefully
on those dates and I found these errors started showing up in the
backend log on the first recording on the 2nd.

2009-03-02 05:21:00.759 [mpeg2video @ 0xb730f7a4]invalid cbp at 42 29
2009-03-02 05:21:00.762 [mpeg2video @ 0xb730f7a4]Warning MVs not available
2009-03-02 05:21:00.787 [mpeg2video @ 0xb730f7a4]00 motion_type at 26 20
2009-03-02 05:21:00.790 [mpeg2video @ 0xb730f7a4]00 motion_type at 24 21
2009-03-02 05:21:00.816 [mpeg2video @ 0xb730f7a4]slice mismatch
2009-03-02 05:21:00.876 [mpeg2video @ 0xb730f7a4]00 motion_type at 8 23
2009-03-02 05:21:00.877 [mpeg2video @ 0xb730f7a4]ac-tex damaged at 2 24
2009-03-02 05:21:00.879 [mpeg2video @ 0xb730f7a4]00 motion_type at 13 25
2009-03-02 05:21:00.881 [mpeg2video @ 0xb730f7a4]00 motion_type at 7 26
2009-03-02 05:21:00.882 [mpeg2video @ 0xb730f7a4]invalid cbp at 6 27
2009-03-02 05:21:00.884 [mpeg2video @ 0xb730f7a4]00 motion_type at 12 28
2009-03-02 05:21:00.885 [mpeg2video @ 0xb730f7a4]00 motion_type at 2 29
.
.
.
2009-03-02 05:54:10.365 [mpeg2video @ 0xb730f7a4]invalid mb type in B
Frame at 39 29
2009-03-02 05:54:10.394 [mpeg2video @ 0xb730f7a4]Warning MVs not available
2009-03-02 05:54:10.402 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 8
2009-03-02 05:54:10.409 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 15
2009-03-02 05:54:10.414 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 9
2009-03-02 05:54:10.417 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 16
2009-03-02 05:54:10.425 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 17
2009-03-02 05:54:10.426 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 18
2009-03-02 05:54:10.427 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 19
2009-03-02 05:54:10.428 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 20
2009-03-02 05:54:10.429 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 21
2009-03-02 05:54:10.431 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 22
2009-03-02 05:54:10.432 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 23
2009-03-02 05:54:10.433 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 24
2009-03-02 05:54:10.435 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 25
2009-03-02 05:54:10.436 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 26
2009-03-02 05:54:10.437 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 27
2009-03-02 05:54:10.439 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 28
2009-03-02 05:54:10.440 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 29
2009-03-02 05:54:10.442 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 10
2009-03-02 05:54:10.445 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 11
2009-03-02 05:54:10.447 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 12
2009-03-02 05:54:10.448 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 13
2009-03-02 05:54:10.449 [mpeg2video @ 0xb730f7a4]00 motion_type at 0 14

I think these show up while commflag is running, because the recording
actually started like an hour before. From what I found in the
archives these seem to just confirm that there is corruption in the
mpeg2 file. The artifacts only show up for a couple of frames and
occur randomly from a few seconds apart to as much as 10 or 15 minutes
apart.  People in the other threads suggested that this is caused by
pci latency or running out of system resources. It is a brand new
system with a ASUS M3N78-EM motherboard, AMD Athlon 64 X2 7750, 1 GB
of RAM, I am recording to a 640 GB SATA drive, and two old 160 GB IDE
drives in a RAID0 array, all in the default storage group. I feel like
I am getting closer to the problem but at the same time I am stumped.
I can't think of anything I did that would cause the recordings on
Sunday afternoon to be fine, and everything since the early hours of
Monday morning to have problems.