[x264-devel] What is the problem of x264? Is it a bug?

肖卫华 xiaoweihuacom at 126.com
Sun Apr 6 14:10:56 CEST 2014


,jdjdjdjdudkjd

发自我的 iPad

在 2014-3-24,10:38,kevin bai <maosheng.bai at gmail.com> 写道:

> Hi all,
> 
> When use x264 for 2 pass encoding, I find there is a big VQ gap at I frame. I captured some logs as below:
> The avg QP of index 6299  I frame jumps to 54.79, which make the total frame is blur. 
> The content of the source steam is courseware, the most frames are still only with some characters change.
> 
> 
> 1st pass
> in:6290 out:6292 type:b dur:2 cpbdur:2 q:24.17 aq:26.93 tex:84 mv:10 misc:154 imb:0 pmb:3 smb:877 d:s ref:0 ;
> in:6292 out:6293 type:b dur:2 cpbdur:2 q:24.17 aq:16.00 tex:0 mv:0 misc:128 imb:0 pmb:0 smb:880 d:- ref:0 ;
> in:6296 out:6294 type:P dur:2 cpbdur:2 q:25.19 aq:21.69 tex:139165 mv:3188 misc:559 imb:212 pmb:222 smb:446 d:s ref:0 ;
> in:6295 out:6295 type:B dur:2 cpbdur:2 q:26.24 aq:25.08 tex:104779 mv:4030 misc:495 imb:92 pmb:324 smb:464 d:s ref:0 ;
> in:6294 out:6296 type:b dur:2 cpbdur:2 q:26.60 aq:27.30 tex:95845 mv:3841 misc:514 imb:84 pmb:306 smb:490 d:s ref:0 ;
> in:6298 out:6297 type:P dur:2 cpbdur:2 q:28.68 aq:19.39 tex:144086 mv:3561 misc:449 imb:301 pmb:177 smb:402 d:s ref:0 ;
> in:6297 out:6298 type:b dur:2 cpbdur:2 q:28.76 aq:29.67 tex:90747 mv:3950 misc:463 imb:102 pmb:330 smb:448 d:s ref:0 ;
> in:6299 out:6299 type:I dur:2 cpbdur:2 q:31.59 aq:18.22 tex:308469 mv:24386 misc:441 imb:880 pmb:0 smb:0 d:s ref:;
> in:6303 out:6300 type:P dur:2 cpbdur:2 q:31.97 aq:19.20 tex:12110 mv:496 misc:754 imb:26 pmb:172 smb:682 d:s ref:0 ;
> in:6301 out:6301 type:B dur:2 cpbdur:2 q:32.40 aq:14.00 tex:8 mv:17 misc:151 imb:0 pmb:3 smb:877 d:s ref:0 ;
> in:6300 out:6302 type:b dur:2 cpbdur:2 q:32.40 aq:17.00 tex:11 mv:3 misc:138 imb:0 pmb:1 smb:879 d:s ref:0 ;
> in:6302 out:6303 type:b dur:2 cpbdur:2 q:32.40 aq:17.00 tex:11 mv:3 misc:138 imb:0 pmb:1 smb:879 d:- ref:0 ;
> in:6307 out:6304 type:P dur:2 cpbdur:2 q:31.26 aq:16.81 tex:13638 mv:453 misc:725 imb:22 pmb:182 smb:676 d:s ref:0 ;
> 
> 2nd pass
> in:6292 out:6293 type:b dur:2 cpbdur:2 q:6.03 aq:3.00 tex:113 mv:267 misc:468 imb:0 pmb:532 smb:348 d:- ref:2048 0 0 0 ;
> in:6296 out:6294 type:P dur:2 cpbdur:2 q:15.83 aq:14.33 tex:165460 mv:18106 misc:498 imb:122 pmb:312 smb:446 d:s ref:752 209 210 53 23 1 ;
> in:6295 out:6295 type:B dur:2 cpbdur:2 q:15.68 aq:14.62 tex:123695 mv:11211 misc:502 imb:38 pmb:280 smb:432 d:s ref:626 63 19 1 0 ;
> in:6294 out:6296 type:b dur:2 cpbdur:2 q:43.20 aq:12.03 tex:37729 mv:4327 misc:352 imb:17 pmb:165 smb:669 d:s ref:288 17 1 4 ;
> in:6298 out:6297 type:P dur:2 cpbdur:2 q:39.84 aq:31.40 tex:20417 mv:8994 misc:573 imb:125 pmb:367 smb:388 d:s ref:464 46 477 344 96 29 12 ;
> in:6297 out:6298 type:b dur:2 cpbdur:2 q:39.49 aq:44.37 tex:7298 mv:3565 misc:625 imb:12 pmb:335 smb:521 d:s ref:370 65 164 8 ;
> in:6299 out:6299 type:I dur:2 cpbdur:2 q:54.79 aq:45.91 tex:22558 mv:7107 misc:439 imb:880 pmb:0 smb:0 d:s ref:;
> in:6303 out:6300 type:P dur:2 cpbdur:2 q:35.56 aq:24.53 tex:159533 mv:19446 misc:221 imb:477 pmb:393 smb:10 d:s ref:1572 ;
> in:6301 out:6301 type:B dur:2 cpbdur:2 q:56.92 aq:3.00 tex:0 mv:0 misc:136 imb:0 pmb:0 smb:880 d:s ref:0 ;
> in:6300 out:6302 type:b dur:2 cpbdur:2 q:55.51 aq:3.00 tex:0 mv:0 misc:136 imb:0 pmb:0 smb:880 d:s ref:0 ;
> in:6302 out:6303 type:b dur:2 cpbdur:2 q:55.33 aq:3.00 tex:0 mv:0 misc:136 imb:0 pmb:0 smb:880 d:- ref:0 0 ;
> in:6307 out:6304 type:P dur:2 cpbdur:2 q:22.31 aq:12.04 tex:224603 mv:10127 misc:478 imb:98 pmb:712 smb:70 d:s ref:2202 580 58 8 ;
> 
> 
> What's the problem? Is it a bug of x264?
> 
> Thanks,
> Kevin
> 1 attachments
> Frame_6299.jpg(58K)
> download preview
> <Frame_6299.jpg>
> _______________________________________________
> x264-devel mailing list
> x264-devel at videolan.org
> https://mailman.videolan.org/listinfo/x264-devel
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x264-devel/attachments/20140406/7fba1cff/attachment-0001.html>


More information about the x264-devel mailing list