From: James Courtier-D. <Ja...@su...> - 2001-10-29 00:57:17
|
What about letterbox pictures encoded in 4:3 frames with top and bottom black borders. A 16:9 TV will just put a 4:3 frame in the centre. If the image is a letterbox image, you will then get borders on top, bottom and left and right, because video out driver will think it is a 4:3 frame. If we could introduce a new aspect ratio which could expand letterbox 4:3 images to full screen 16:9, I think it could be useful. Cheers James > -----Original Message----- > From: xin...@li... > [mailto:xin...@li...]On Behalf Of Harm van der > Heijden > Sent: 28 October 2001 23:13 > To: xin...@li... > Subject: [xine-devel] Re: [xine-user] zoom > > > On Sun, Oct 28, 2001 at 10:26:05PM +0100, Torgeir Veimo wrote: > > I've bought the latest box set of movies by Stanley Kubrik, > which are in > > 1:1.33 format, but more like 1:1.77 inside this format. Thus it > would be > > nice to be able to zoom in to avoid most of the letterbox effect. > > Most video out drivers should be able to cope with removing black borders. > For YUY2 data it involves something like out_ptr = src + top_lines*width*2 > and out_height = height - top_lines - bottom_lines, if you get my drift. > YV12 is similar, with factors /4 for U and V. > > Removing bits on the left and right of the picture is not really > an option, > as it would require video out drivers to have a sense of stride (distance > between horizontal lines). Some, if not most, require that stride > == width. > > My guess, and I'm not hampered by any real knowledge of how Xv and XShm > drivers work, is that it would be fairly easy to add a hack for cutting > X lines at the top and Y lines at the bottom. A more general xine solution > would be more work. > > BTW, in the new dxr3enc driver, I'm doing the exact opposite: the dxr3 > hardware only supports 16:9 and 4:3, and in order to display square > pixel video (read AVI) correctly I add black lines until I get to the > closest supported a.r. It's all done with pointer offsets (I allocate > more YUV memory than the decoder sees, and blacken out the borders > in advance). Again, this trick only works with top and bottom borders. > Stuff at the side is not possible with the current xine architecture > (at least not without costly copy operations), due to the stride thing. > > Regards, > > Harm > > > > > > _______________________________________________ > xine-devel mailing list > xin...@li... > https://lists.sourceforge.net/lists/listinfo/xine-devel |