[mythtv] Using short skips during live TV

[Dan Wilga]

Here's another variation of the problem that's been happening to me 
from time to time, but I only just took note of the exact sequence 
this morning:

1. I started watching live TV and after about 2 minutes paused it to 
leave the room.

2. When I came back after about 5 minutes, I un-paused it and watched 
for another 2 minutes.

3. At this point, a commercial came on, so I pressed Skip Forward. 
Myth actually went *backward* about 2-3 minutes, showing me stuff I 
had already watched.

4. I used Skip Fwd to go forward a bunch of times, which worked as I 
would expect (based on the weird position it was now in). However, 
when I advanced to what Myth thought was the end of the buffer (4 
seconds behind), it was actually 2-3 *minutes* behind Live TV. I was 
able to continue watching just fine, however Myth continued with this 
skewed idea of what "live" was, always behind reality.

The only way I was able to clear it up was by quitting back to the 
menu and hitting Watch TV again.


At 11:59 AM -0500 2/5/05, Dan Wilga wrote:
>I am running the CVS version from Feb 2, with the new ivtv code, on 
>a dual tuner 250/350 system. I'm using the 350's TV-out, with ivtv 
>0.2.0-rc3f. (Tired of seeing this paragraph yet? :-) )
>
>I have my skip forward/reverse times both set to 10 seconds.
>
>Here's the problem:
>1. Start watching live TV, and let it run for 20-30 seconds.
>2. Press skip right. The position doesn't change, but the OSD 
>appears, to confirm that we're only 1 second behind.
>3. Now, press skip backward. The OSD says we're 12 seconds behind, 
>not 10 or 11 as I would expect.
>4. Press skip forward. The OSD now says 4 seconds behind. No amount 
>of further pressing will get it back to 1 second behind. The video 
>stops momentarily with each press, but stays 4 seconds behind.
>
>So, I think the problem may be that the skip backward function isn't 
>taking into account any latency in the process of actually 
>repositioning the player. By the time it has repositioned, another 
>second or two of real time has elapsed. (It seems like less than a 
>second, though, so this is only a guess.)
>
>Then, when skipping forward to what should be only a 1-second delay, 
>the latency accumulates to 4 seconds instead. Allthough, that theory 
>wouldn't explain why I can never get it to less than 4 seconds 
>behind again.
>
>I realize this is probably as clear as mud, so if you want more 
>details, let me know.

-- 
Dan Wilga                                         dwilga at mtholyoke.edu
Web Administrator                             http://www.mtholyoke.edu
Mount Holyoke College                                Tel: 413-538-3027
South Hadley, MA  01075            "Who left the cake out in the rain?"