- 18 Jan, 2017 4 commits
-
-
Jerome Jiang authored
Avoid quality loss when copying partition of superblock with large motions. Maximum consecutively copied frames can be set (currently 5). Change-Id: I11c30575514f02194c0f001444cf4021609e5049
-
Peter Boström authored
-
James Zern authored
-
Jerome Jiang authored
Change-Id: I4fa3262e0f1c4018604c954b020ec5d1e3d1465c
-
- 17 Jan, 2017 5 commits
-
-
Jerome Jiang authored
-
Jerome Jiang authored
Also set the flag to 1 when exit early choosing 64x64 block such that skipping new mv for golden works in these scenerios. Change the size of prev_segment_id to the number of superblocks to save memory. Borg test shows quality regression of 0.012% on average PSNR and 0.035% on SSIM. Change-Id: I5014224c8617d439d35c66ece3fed9ae30b31d23
-
Johann Koenig authored
-
Moriyoshi Koizumi authored
This should have been taken into account at 64347a10 Change-Id: Ie8e3ad7cbaab3e5799e04bd50f2639390b0a2428
-
Peter Boström authored
Adds an optional output framestats.csv file that prints comparions per-frame instead of averaged over the entire clip. It prints per-channel and combined metrics for SSIM and PSNR. Change-Id: Id28dfade27bc5775b59a9d83cfe8b37d1d52b686
-
- 14 Jan, 2017 1 commit
-
-
James Zern authored
before calling vp9_apply_encoding_flags() which may crash if the resolution was invalid. this is the same change as: c0523090 vp8e_encode: check validate_config return BUG=https://bugzilla.mozilla.org/show_bug.cgi?id=1315288 Change-Id: Icd2aab322422e83d3a778fca6d7789e5000239d7
-
- 13 Jan, 2017 8 commits
-
-
Marco authored
If enabled will compute source_sad for every superblock on every frame, prior to encoding. Off by default, only on for speed=8 when copy_partition is set. Change-Id: Iab7903180a23dad369135e8234b7f896f20e1231
-
Marco Paniconi authored
-
Marco authored
Change-Id: I4799cb2b67d911ee385e6d6992c61633ca77e69d
-
Jingning Han authored
-
Jingning Han authored
-
Peter Boström authored
Used with --framestats=file.csv. Currently prints raw codec QP (not internal 0-63 range) and bytes per frame. Change-Id: Ifbb90129c218dda869eaf5b810bad12a32ebd82d
-
Marco Paniconi authored
-
Jerome Jiang authored
Avoid many visual artifacts. Compression quality is improved by more than 1%. Encode speed is about 4% for QVGA and 6% for VGA faster on android. Change-Id: I4dd0a81429ddf7efdef1e80a191da5fb8de8e8af
-
- 12 Jan, 2017 7 commits
-
-
Johann authored
-
Jingning Han authored
Use same transpose process as inv_txfm_sse2 does. Change-Id: I2db05f0b254628a11f621c4c09abb89501ba6d3c
-
Jingning Han authored
Use same transpose process as inv_txfm_sse2 does. Change-Id: Ic4827825bd174cba57a0a80e19bf458a648e7d94
-
Peter Boström authored
To be used for frame stats output of vpxdec. Change-Id: I0739a01bd3635c4b3fedd58f3e27363ce8fb1b1e
-
Johann authored
Change-Id: If27447472417c7ed34238295427ddb9da0561725
-
Marco Paniconi authored
-
Johann Koenig authored
-
- 11 Jan, 2017 10 commits
-
-
Peter Boström authored
-
Jerome Jiang authored
-
Johann Koenig authored
-
Peter Boström authored
Renames SSIM to VpxSSIM as an upscaled weighted SSIM metric, then prints Y, U and V channels unweighted as well as a weighted but not scaled SSIM score that's 8/1/1 parts Y/U/V (same as VpxSSIM). Change-Id: Iff800cc8f145314eeb1a9b4af1e11a25bec095ca
-
Jingning Han authored
-
Jerome Jiang authored
For speed 8, it speeds up the encoding on android by 6% for QVGA and 7.4% for VGA with the new threshold. Overall PSNR is improved by 0.667 for rtc. Change-Id: I4a644560b32c0b5b4e9f49ffb953d000413a3732
-
Johann authored
This file is guarded by HAVE_NEON_ASM in the .mk file now. Change-Id: I513a621c234aa90ad52e426c8ed494d8a7d4b74a
-
Johann authored
Demonstrate its use with the IDCT test. Change-Id: Idf87fe048847c180f13818fd4df916ba4500134b
-
hui su authored
Also reduce the number of test frames. Change-Id: Iea6fa93ca6b924535aef7bf8b388db4d0ec84c08
-
Marco authored
If enabled denoiser will only denoise the top spatial layer for now. Added unittest for SVC with denoising. Change-Id: Ifa373771c4ecfa208615eb163cc38f1c22c6664b
-
- 10 Jan, 2017 4 commits
-
-
Jingning Han authored
This commit reworks the SSSE3 implementation of the forward 8x8 2D-DCT. It uses a cyclic rotation approach to the temporary xmm registers. It reduces the average cycles from 158 to 154. The SSE2 version uses 169 cycles. Change-Id: I1b79b9642aae0ed3fb3cefb5b70246e6de5d5caa
-
Marco authored
When aq=3 mode is on and the gf_cbr_boost is set: make sure golden frame is always refreshed, and don't incorporate segement cost in qp setting on the boosted golden frame. Better performance on RTC set with gf_cbr_boost on, for example with gf_cbr_boost=50, gains from ~0.5-3%. Change-Id: Ie811f5e4d444ff3320bd6e2c1745b2c4c09a8460
-
Jerome Jiang authored
-
Jerome Jiang authored
Quality improved by 1.866 and 0.386 for two noisy clips (dark720p and marcooffice720p), respectively. Change-Id: Ib33a7672ae9ca53da156208f7cd13f07b5543e44
-
- 09 Jan, 2017 1 commit
-
-
Jerome Jiang authored
-