[mythtv] AVSync2 Refinements

[Tim Pletcher]

On Mon, Dec 16, 2019 at 12:22 AM David Engel <david at istwok.net> wrote:

> On Sat, Dec 14, 2019 at 03:00:24PM +0000, Mark Kendall wrote:
> > On Fri, 13 Dec 2019 at 18:21, Tim Pletcher <pletchtd at gmail.com> wrote:
> > >
> > >
> > >> I was able to do a little more testing today and figured out how to
> nicely parse the log playback,timestamp data into a comma-delimited file to
> enable me to look more closely at some real data.
> > >>
> > >> First, I found that a 0.4 gain and 0.9 filter as I initially proposed
> does work well.
> > >
> > >
> > > Apologies Mark.
> > >
> > > I just noticed you already committed the 0.04.  I don't mean to appear
> so indecisive on this :)   I think it is probably splitting hairs at this
> point as the control seems to work pretty good with either settings.
> >
> > Tim
> >
> > No problem. I'm minded to go back to 0.4 as it generally seems a
> > little slow to respond at the moment.
>
> Count me firmly in the 0.4 camp.  I didn't see much difference with
> progressive content yesterday but did today with interlaced content.
> 0.04 takes noticably longer to settle down.
>
> David
>
>
For me, the 0.4 gain with a 0.6 filter coefficient causes jumps in video
playback on my low end apollo lake series frontends using VAAPI.  Without
sufficient low pass filtering, the adjustments are too much response to the
fairly regular larger spikes in sync measurement when using deinterlacing (
https://imgur.com/YpmvIG7).

Adjustments of more than roughly 5 ms with any regularity do not provide
smooth video playback and this is particularly highlighted when watching
programs with lots of motion such as sports. I provided an example plot
earlier in this thread with a gain of 0.4 and a filter coefficient of 0.9
and those settings provide satisfactory video playback on these apollo lake
machines.

Tim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mythtv.org/pipermail/mythtv-dev/attachments/20191216/f145c6d7/attachment.htm>