[mythtv] Feedback on svn

Mark Spieth mark at dclabs.com.au
Fri Oct 21 06:00:56 UTC 2005

hopefully Ill get to this this weekend. havent had too much time recently
for myth hacking.
also the audio sync adjust stuff I want to do.
what Id really like to do is use the bitrate value and use that to determine
how many frames are in the local ringbuffer and use that as a better metric
for how far we have to go before end is reached. still thinking about this.
maybe isaac you have some ideas about this.

On Oct 19, 2005, at 3:30 PM, Mark Spieth wrote:

1. The time stretch feature doesn't set the value to 1.00 again when I
watch LiveTV delayed and LiveTV reaches the end. It was setting to
1.00 when the end of the buffer was reached a few weeks ago, now the
picture/sound are stuttering instead of switching back the time
stretch value to 100%.

yes I know but there are a few issues with this. always have been.
the current threshold is set to 2 secs worth of frames.
Ive found this works ok on a fe on my be but not very well on my xbox fe
which needs more like 4 seconds of buffer due to network latencies.
which scenario do you have?
If you want to test, see function NuppelVideoPlayer::IsNearEnd and change
the factor for margin (video_frame_rate*2) to something else like 4 or 6
I probably need to get some sort of metric on how many frames are in teh
ringbuffer on the local machine and work off that instead. Ive been thinking
about this for a while. or else just a fudge for non local fe's.
 If you do, let me know and Ill generate a patch for a separate nonlocal fe

I know this is not exactly along the same lines as this topic, but would it
be possible to increase this local buffer on remote frontends?  Sometimes
when the wireless signal is marginal it would be nice to have about 10
seconds of buffer.  Ie get as much as you can when the signal is good and
use it when there is high latency.


mythtv-dev mailing list
mythtv-dev at mythtv.org

More information about the mythtv-dev mailing list