[mythtv] [RFC][PATCH] Avoiding scaling

Ed Wildgoose lists at wildgooses.com
Mon Dec 13 09:47:35 UTC 2004


>>Another possible way to tackle this might be to allow resolution 
>>changing as well?
>>    
>>
>
>  Ah. I can see in xvmc_videoout.cpp that VideoOutputXvMC::InputChanged
>calls display_res->switchToVid before it calls MoveResize (where I've
>added my code). I hadn't looked at what switchToVid does but, scanning
>through it now, it seems to only pick a mode if it's an exact match to
>the source. I guess that could be changed so it picks a "best" mode, but
>I'm not sure what "best" would be here. Certainly it would make sense to
>allow a fuzzy match (within 5%) if my snap patch is in use. On an
>unrelated note, perhaps it should also take account of the refresh rate
>for the mode?
>
>  None of this would be useful to me at the moment (I only have one
>video mode for my TV-out) but if it would help get this committed, I'm
>happy to do the work if someone else can test it.
>  
>

Indeed, changing resolution would be nice.  Agree that checking refresh 
rate is important though!

I would have thought that matching on refresh rate first, and then 
looking for the mode which is closest in total area to the desired res 
(ie width x height) would be the one to pick, perhaps with a rule to 
match the zoomed mode, rather than the one with borders, in case of a tie?

I think my video card can do a 704x576x50Hz output mode for example?

If you were willing to do some work on this, then I am very interested 
in testing.  I probably don't have enough time right now to help with 
programming though - too much audio coding still needs doing for my 
setup.  As for committing to cvs, just post here with Patch in the 
subject and ask explicitly for it to be committed.

Thanks and good luck

Ed W


More information about the mythtv-dev mailing list