2018-07-28 13:34 GMT+02:00, Jeff <163j...@163.com>: > Hi Carl, > > > Sorry for late reply. > > > I make a simple test on my MacBook Pro (13-inch, 2016, 2 GHz Intel Core i5, > 8 GB 1867 MHz LPDDR3). > I read a 1088x1920 BGR24 frame from local disk, and then scale it to > 480x854-YUV420P 500 > times respectively by libswscale & libyuv. The result shows that libswscale > take about 4.19 > seconds while libyuv just take about 0.93 seconds. Which means libyuv is > about 4~5 times > faster than the libswscale on my test. Is that say something?
Not sure. My question was if libyuv is multi-threaded (if it is, it should be possible to get the scale filter to the same speed), if it outputs the same quality and what kind of asm optimization it uses. Carl Eugen _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
