I'd appreciate your reply to my questions below.
The speed of channel changing can be improved in other, better ways. And
by 'better', I mean, 'applicable to every user, not just some'.
I agree that keeping PVR capabilities whilst achieving speeds equivalent to
those which are already accepted in current TVs is the best goal. Is that
achievable though?!? Maybe I grab the wrong end of the end, but people seem
to suggest that the buffering/ringcache is responsible for most of the delay
and now that eliminating it can't be done?
You are replying to *MY* post where I stated that I had already patched
an earlier version of myth to change channels on DVB in usually under
500ms and sometimes much less. This was done by fiddling with timeouts,
buffer sizes and removing redundant pause statements
I really feel that this moaning about changing the implementation is
heading in the wrong direction and just alienating the few people who
can help fix it rather than helping
If someone has instrumented parts of tv_play and has a recent patch to
show your code then please post it. I think this helps other less
technical people play with some instrumentation and track down their own
timings. ie it's fairly easy to cut and paste some code once someone
has stuck the basic outline code in roughly the right place
The delays ARE solvable, they just need some investigation and time.
It's NOT trivial to do this though because Myth is a multithreaded
program and it's quite easy to accidently setup some race condition
which breaks things. But it IS possible, and *YOU* can help make it
happen by reporting back on channel change timings based on some
profiling of the code (not based on using your wrist watch and flipping
channels mind)
Please help and get involved. I would love to see this all get faster
Ed W
_______________________________________________
mythtv-dev mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
