From d63b469c6811a10763d0433929a1410476b5fcf0 Mon Sep 17 00:00:00 2001 From: Juergen Keil Date: Sat, 1 Sep 2001 18:11:04 +0000 Subject: Don't change ideal_width for videos with "width % 32 == 16", so avoid software scaling in yuv2rgb. Is the "ideal_width &= 0xFFFFFF0;" rounded needed at all? CVS patchset: 544 CVS date: 2001/09/01 18:11:04 --- src/video_out/video_out_xshm.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) (limited to 'src') diff --git a/src/video_out/video_out_xshm.c b/src/video_out/video_out_xshm.c index a8f1a7d80..51be649fb 100644 --- a/src/video_out/video_out_xshm.c +++ b/src/video_out/video_out_xshm.c @@ -17,7 +17,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA * - * $Id: video_out_xshm.c,v 1.28 2001/08/28 19:16:20 guenter Exp $ + * $Id: video_out_xshm.c,v 1.29 2001/09/01 18:11:04 jkeil Exp $ * * video_out_xshm.c, X11 shared memory extension interface for xine * @@ -521,7 +521,19 @@ static void xshm_calc_output_size (xshm_driver_t *this) { ideal_height *=2; } - ideal_width &= 0xFFFFFE0; + /* + * XXX: Why is ideal_width rounded down to an exact multiple of 32? + * + * AVI DivX sometimes use a width with "width % 32 == 16", so that + * "ideal_width &= 0xFFFFFE0;" modifies the width and triggers the + * yuv2rgb do_scale code... + * E.g. a 720x540 video is scaled to 704x5XX + * + * Round down to a multiple of 16 for now, + * maybe we should delete the ideal_width &= 0xFFFFFF0; ? + */ + /* ideal_width &= 0xFFFFFE0; */ + ideal_width &= 0xFFFFFF0; this->calc_dest_size (ideal_width, ideal_height, &dest_width, &dest_height); -- cgit v1.2.3