[mythtv] Suggestion for improving SLOW channel changing in
LiveTV.....
Cory Papenfuss
papenfuss at juneau.me.vt.edu
Thu Sep 8 14:34:38 UTC 2005
>> Again, I think that for the most part this is largely unnecessary. Lots
>> of work for minimal gain. Once one truly accepts the PVR method of
>> watching TV, "LiveTV" becomes unnecessary and undesired.
>
>
> That's a fallacy. Don't fall into the trap of thinking that everyone uses a
> PVR the same way you do. Live TV doesn't go away for a large number of
> users.
Fair enough. I will say that when I first built my mytbox, I was
in the same boat. LiveTV was too slow, and I couldn't understand why it
wasn't a priority to "fix" it. The responses I got were basically, "Once
you use your mythbox the way it's intended, you won't watch LiveTV
anymore." I was skeptical then, but I'm a believer now.
I also didn't use LiveTV because the record/playback quality on a
PVR-250 was noticably worse than watching LiveTV on the TV. Not exactly
sure why (I've been trying to figure it out for close to 2 years), but it
is. For the convenience of time-shifting, it's a sacrifice I'm willing to
make.
I still say that a more pipelined dataflow would be a good
compromise. Basically, stream the data from the card into the *playback*
buffer (just data to dump to XV). *THEN* use the same data to either
crunch (if bttv-based card), or stream to disk (if mpeg2 card). If you
then want to pause or rewind, the ringbuffer is still there. It would
have to transition to the current mode of operation at that point
(decoding the ringbuffer)... at least channel changing would be quicker
though... at the expense of double memory transfer.
Again... lots of effort for minimal (perceived) gain. Alas, the
people who really would like this sort of thing are the ones who may not
have the skill to code it up (and vice versa... those that do don't care
to).
-Cory
--
*************************************************************************
* Cory Papenfuss *
* Electrical Engineering candidate Ph.D. graduate student *
* Virginia Polytechnic Institute and State University *
*************************************************************************
More information about the mythtv-dev
mailing list