Environment: -------------- Platform:IVB Libva: (staging) commit 38c94cd922473095814ed9a9f99ad98fcc9c285d Intel-driver: (staging) commit c0ef9d99df37ae45589fecb898727be495e50304 gst-vaapi: (qa) commit 01c92b04b095be9163976a53ac2c4305c84fa274 Bug Info: -------------- FREXT02_JVC_C.264 352x288 High 3.1 interlace?(Yes) Y SSIM extreme 0.904751 Y SSIM average 0.930559 U SSIM extreme 0.987066 U SSIM average 0.990115 V SSIM extreme 0.993283 V SSIM average 0.995281 Frame NO. 16 Dumped Frame NO. 17 Lost Frame 0
Libva: (staging) bfa1553c18b7b12cf1c221cafdc1784672a7b4ac Intel-driver: (staging) 04d459f88353a5be82724e648bb30ddb37286d26 Gst-vaapi: (qa) 205ae880b7696dbd76d5a2968bb858a9fa7c5958 below files are failed by ssim low, and they are passed with mplayer. ./HCAFR4_HHI.264 ./freh10.264 ./alphaconformanceG.264 ./HCAFF1_HHI.264 ./MR3_TANDBERG_B.264 ./freh8.264 ./HPCAMAPALQ_BRCM_B.264 ./Freh12_B.264 ./freh4.264 ./HCAFR2_HHI.264 ./FM1_FT_E.264 ./HCAFR1_HHI.264 ./freh9.264 ./HCAMFF1_HHI.264 ./HPCALQ_BRCM_B.264 ./FREXT01_JVC_D.264 ./FREXT02_JVC_C.264 ./freh5.264 ./HCAFR3_HHI.264 ./freh11.264 ./HPCADQ_BRCM_B.264 ./freh3.264 ./BA3_SVA_C.264 ./MR6_BT_B.h264 ./MR8_BT_B.h264 ./MR7_BT_B.h264 ./FRExt1_Panasonic.avc ./FRExt2_Panasonic.avc ./FRExt4_Panasonic.avc ./FRExt3_Panasonic.avc
FREXT02_JVC_C.264 start from 1st frame, low ssim Y_SSIM = 0.920795 U_SSIM = 0.987066 V_SSIM = 0.993604 Y_MIN_SSIM = 0.904751
Created attachment 78075 [details] [review] h264-modify-default-matrices-order default scaling_lists are set in raster scan which should be prepared in zigzag scan order. This patch is to fix it
Created attachment 78076 [details] [review] h264-fix-pps-matrices-error there is something wrong in gst_h264_parse_pps about calculate pps->scaling_lists_4x4 && pps->scaling_lists_8x8. This patch is to fix it.
Created attachment 78895 [details] [review] h264-fix-ssim fix bug when remove reference if picture frame_num is same
the new bug was created for extended profile bug.https://bugs.freedesktop.org/show_bug.cgi?id=64268
A new bug was created for "h264-modify-default-matrices-order" bug.https://bugs.freedesktop.org/show_bug.cgi?id=64272
A new bug was created for "h264-fix-pps-matrices-error" bug,https://bugs.freedesktop.org/show_bug.cgi?id=64273
A new bug was created for "h264-fix-ssim " bug,https://bugs.freedesktop.org/show_bug.cgi?id=64274
Libva: (staging) 63182cd2c03019751b8a2ded17082324be143417 Intel-driver: (staging) 5bd630d4ea732818728b0a464f45739c86e08be3 Gst-vaapi: (0.4) 533f11ac090c475fcceedcc9caf70ba766561541 below files are failed by ssim low, and they are passed with mplayer. ./alphaconformanceG.264 ./CABAST3_Sony_E.jsv ./CABASTBR3_Sony_B.jsv ./CAWP5_TOSHIBA_E.264 ./CVWP5_TOSHIBA_E.264 ./FM1_FT_E.264 ./FRExt_MMCO4_Sony_B.264 ./MR6_BT_B.h264 ./MR7_BT_B.h264 ./MR8_BT_B.h264
A new bug was created for "low ssim value with HPCAMOLQ_BRCM_B.264 & HPCVMOLQ_BRCM_B.264 ". bug.https://bugs.freedesktop.org/show_bug.cgi?id=64375
A new bug was created for "low ssim value with CVWP5_TOSHIBA_E.264 " bug.https://bugs.freedesktop.org/show_bug.cgi?id=64388
A new bug was created for "gstvaapi play alphaconformanceG.264 error" bug.https://bugs.freedesktop.org/show_bug.cgi?id=64527
A new bug was created for "lost frames when decoding FM1_FT_E.264 " bug,https://bugs.freedesktop.org/show_bug.cgi?id=64622
A new bug was created for "low ssim value with FRExt_MMCO4_Sony_B.264" bug.https://bugs.freedesktop.org/show_bug.cgi?id=64624
A new bug was created for "buglow ssim value with CABAST3_Sony_E.jsv". https://bugs.freedesktop.org/show_bug.cgi?id=64625
A new bug was created for "low ssim value with CABASTBR3_Sony_B.jsv" bug,https://bugs.freedesktop.org/show_bug.cgi?id=64627
A new bug was created for "low ssim value with MR8_BT_B.h264 " bug,https://bugs.freedesktop.org/show_bug.cgi?id=64628
A new bug was created for "lost frames when decoding MR7_BT_B.h264" bug.https://bugs.freedesktop.org/show_bug.cgi?id=64629
A new bug was created for "low ssim value with MR6_BT_B.h264 " bug.https://bugs.freedesktop.org/show_bug.cgi?id=64630
ibva: (staging) 968a9bf006679ad8d752fa1ddcd4bb80fc8c5e51 Intel-driver: (staging) bb24c8a81e512d19aad0359d81f7247e6f20cc29 Gst-vaapi: (0.4) 3e7aa4a9372393369a60b012305e19cbd7bb533f below files are failed by ssim low, and they are passed with mplayer. CABAST3_Sony_E.jsv CABASTBR3_Sony_B.jsv CVFC1_Sony_C.jsv FM1_FT_E.264 FRExt_MMCO4_Sony_B.264 MR6_BT_B.h264 MR7_BT_B.h264 MR8_BT_B.h264
Fixed in git master branch (differently). Thanks.
commit bd5ae1b2205edaaf5581463cab200d15d80e7fd5 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com> Date: Mon Dec 9 12:46:45 2013 +0100 decoder: h264: fix decoding of scaling lists. The GStreamer codecparser layer now parses the scaling lists in zigzag scan order, as expected, so that to match the original bitstream layout and specification. However, further convert the scaling lists into raster scan order to fit the existing practice in most VA drivers. https://bugzilla.gnome.org/show_bug.cgi?id=706406 commit f870469b0a4451f885de5a1d0e81f31ac7687150 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com> Date: Mon Dec 9 12:07:28 2013 +0100 codecparsers: update to gst-vaapi-branch commit 177c73b. a7e3255 add H.265 (HEVC) bitstream parser 177c73b h264: fix picture level scaling lists derivation (rule B) 14733f1 h264: fix parsing of VCL HRD parameters 59a0b47 h264: store quantization matrices in zig-zag order ffb6e26 h264: add helpers to convert quantization matrices c78a504 mpeg2: also initialize debug category in parse_sequence_header() 719d1b0 mpeg2: turn internal consistency check into a g_assert() 5241d8e all: remove some unused functions 18eb312 all: fix for GST_DISABLE_GST_DEBUG 963c04a all: make warnings more meaningful
It works for the most files listed above, but some clips reproduced with low SSIM value, they are MR6_BT_B.h264 MR8_BT_B.h264 MR7_BT_B.h264 FRExt1_Panasonic.avc FRExt2_Panasonic.avc FRExt4_Panasonic.avc FRExt3_Panasonic.avc, In addition to these, the following files BA1_FT_C.264 CABA3_TOSHIBA_E.264 CABAST3_Sony_E.jsv CABASTBR3_Sony_B.jsv CACQP3_Sony_D.jsv cama1_vtc_c.avc cama2_vtc_b.avc cama3_vtc_b.avc CANL1_TOSHIBA_G.264 CAWP1_TOSHIBA_E.264 CI1_FT_B.264 CVCANLMA2_Sony_C.jsv CVWP1_TOSHIBA_E.264 CVWP2_TOSHIBA_E.264 CVWP3_TOSHIBA_E.264 RExt_MMCO4_Sony_B.264 also have this problem.
Bug 724518(https://bugzilla.gnome.org/show_bug.cgi?id=724518) - [gst-vaapi-master ]Low ssim value when decoding some AVC files was filed on gnome, as bug 68009, close it.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.