[mythtv] Mythweb and Windows Media Player over Samba

[Torbjörn Jansson]

mythtv-dev-bounces at mythtv.org <> wrote:
> Torbjörn Jansson wrote:
>>> From my experience, and I may have it set up all wrong on my
>>> machine, using the myth:// URL still does not enable the seek bar
>>> inside of Windows Media Player.
>> 
>> It shoud work, unless the file has a broken seektable, but in that
>> case the filters will try to rebuild the seektable, but most players
>> can't handlefiles where the length changes.
>> 
>> If you play files that has finished recording seeking shoud always
>> work.
>>> With some other media players, I can seek backwards, but
>>> still cannot move forwards.
> 
> I've had the same problem with WMP on WinXP not allowing the file to
> seek with the recordings.  This is without backend support since I
> still have problems getting that to work with my laptop, although I
> have had success with my desktop.  The laptop doesn't have
> any problems
> connecting to mysql to pull up the titles with the explorer
> extensions though. 
> 

What kind of nuv file is this? mpeg2? Or mpeg4/rtjpeg?
What version of the filters do you use?

Does seeking never work? Or is it just some files that you can't seek in?

I'll probably release a new verson of the filters for the cvs version of
mythtv soon.
I have rewritten the buffering and fixed some bugs that affected seeking
preformance, seeking is very fast now and the network filter does read ahead
so no more jumpy playback incase the backend computer decides to do
something else for a short while.
It woud be intresting to know how well the new code works with wireless
networks.

> Does DSMYTH need to be updated to work with the latest cvs since the
> protocol has now been bumped up to number 8 I believe?  Any chance
> of including deinterlacing in the filter or is that solely the
> responsibility of the player? 
> 

Yes, it needs to be updated.

Deinterlacing shoud be done in another filter or by the player app.
The DScaler project ( http://sourceforge.net/projects/deinterlace ) is
working on creating a directshow version of their deinterlacing algorithms,
they're only in cvs at the moment, but I'm looking forward to testing them
when it's ready.
I haven't looked any closer at how they work, but I think they might be
usefull.