Jason Tackaberry wrote: > There definitely used to be an aspect bug with expand plugin (it was my > fault, I added it :)) but I submitted a fix for it that was merged a > long time ago. > > But I'm not sure if this is same problem you ran into, but I'll try to > reproduce it. > > I guess this explains why video played by popcorn was no longer being > expanded. Incidentally I use kaa.popcorn to play pretty much everything > I watch on my HTPC, so I notice when things don't work properly pretty > quick. :)
The expand plugin does not work nice together with the on-the-fly aspect changes you encountered. > For purposes of zooming, I think you can accomplish the same thing as > VO_CROP by passing negative values for the first tuple (x,y) if you want > crop on the top/left side, and increasing the (w,h) values if you want > to crop on the bottom/right side. If you want to crop _within_ the > window, this won't work, but if I understand what you want to do with > zooming, I think you don't need to use VO_CROP here. I don't understand what you mean with "within the window" but the negative values should work. I tried it and it did not show the results I expected. I guess I should try again. >> That should work. The only case where the code you commented out makes >> a difference is when you have a new width, height or aspect and again >> a new aspect _one_ frame later. > > So you mean you're getting two aspect changes in two adjacent frames? > And the second aspect change is not representative of what's in the stream? Yes. As example I get aspect=1.33 to the function. I do some calculations and return something. Next frame (!) I get aspect=1.35 (close but not equal) and I calculate again. This results in aspect=1.33 the next frame and so on. This happened for some window size and movie size combinations. > Anyway, I do admit this is confusing. I still can't get my head wrapped > around the non-square pixel video / non-square pixel display case > without taking a few minutes to remind myself how it's supposed to work, > and as soon as you start dealing with a combination of frame aspect and > _pixel_ aspect, it's a real mindfuck. Consequently it's hard to get > right. (Also I'm getting old and increasingly senile, and as a result, > stupider.) I think we should enumerate all the possible configurations > to help at the very least devise a test plan. 1. (non) square video pixel 2. (non) square monitor pixel and this merged together with 1. keep it as it should be with black bars if needed 2. force 4/3 or 16/9 with black bars 3. ignore aspect, fill window 4. cut off stuff so it fits without black bars and this also needs to be combined with 1. zoom in (cut of x% of the movie, e.g. overscan in recording) 2. expand filter should always be full window >> I have no idea _what_ the parameter I return are about. You know, >> missing and doc and such things. :) > > Well, _xine_frame_output_cb is the python version of frame_output_cb, > which is called frequently inside xine. Kaa just passes the call along > into python space. So the lack of documentation in this area is less my > fault and more xine's. :) But see > http://xinehq.de/index.php/hackersguide for a bit of info. I will. > The one you get is the _video_'s pixel aspect ratio. This, for example, > helps you calculate the appropriate display width/height for the given > video. The one you return is the _display_'s pixel aspect ratio. For > almost all cases, this value is going to be 1.0. In your case > (1280x1024 4:3) it's > going to be 0.937, or (1280.0/1024)/(4/3). I need to return something else when I want to change the aspect as described above. Or should I do it somewhere else? Dischi -- How can something be 'new and improved'? If it's new what was it improving on?
pgpS6L0uspSa5.pgp
Description: PGP signature
------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/
_______________________________________________ Freevo-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freevo-devel
