summaryrefslogtreecommitdiff
path: root/vp9/encoder/vp9_quantize.c
AgeCommit message (Collapse)Author
2013-07-15Inline vp9_quantize() in xform_quant().Ronald S. Bultje
Cycle times: 4x4: 151 to 131 cycles (15% faster) 8x8: 334 to 306 cycles (9% faster) 16x16: 1401 to 1368 cycles (2.5% faster) 32x32: 7403 to 7367 cycles (0.5% faster) Total encode time of first 50 frames of bus @ 1500kbps (speed 0) goes from 1min39.2 to 1min38.6, i.e. a 0.67% overall speedup. Change-Id: I799a49460e5e3fcab01725564dd49c629bfe935f
2013-07-11Moving segmentation related vars into separate struct.Dmitry Kovalev
Adding segmentation struct to vp9_seg_common.h. Struct members are from macroblockd and VP9Common structs. Moving segmentation related constants and enums to vp9_seg_common.h. Change-Id: I23fabc33f11a359249f5f80d161daf569d02ec03
2013-07-01Update quantize SSSE3 SIMD to cover 32x32 transform case also.Ronald S. Bultje
Encode time of bus (speed 0) 50 frames @ 1500kbps goes from 2min14.4 to 2min10.1, i.e. a 2.3% overall speed increase. Change-Id: I3699580e74ec26c7d24e03681bc47ba25ee1ee87
2013-07-01Quantize (64-bit only, for now) SSSE3 SIMD.Ronald S. Bultje
Total encoding time for first 50 frames of bus (speed 0) @ 1500kbps goes 2min34.8 to 2min14.4, i.e. a 10.4% overall speedup. The code is x86-64 only, it needs some minor modifications to be 32bit compatible, because it uses 15 xmm registers, whereas 32bit only has 8. Change-Id: I2df53770c2e850813ffa713e1a91b45b0082b904
2013-06-28Make coefficient skip condition an explicit RD choice.Ronald S. Bultje
This commit replaces zrun_zbin_boost, a method of biasing non-zero coefficients following runs of zero-coefficients to be rounded towards zero, with an explicit skip-block choice in the RD loop. The logic is basically that if individual coefficients should be rounded towards zero (from a RD point of view), the trellis/optimize loop should take care of it. If whole blocks should be zero (from a RD point of view), a single RD check is much more efficient than a complete serialization of the quantization loop. Quality change: derf +0.5% psnr, +1.6% ssim; yt +0.6% psnr, +1.1% ssim. SIMD for quantize will follow in a separate patch. Results for other test sets pending. Change-Id: Ife5fa641163ac5150ac428011e87188f1937c1f4
2013-06-27Inline quantize so idiv instruction gets removed from inner loop.Ronald S. Bultje
Encoding time of first 50 frames of bus @ 1500kbps (speed 0) goes from 3min15.0 to 3min10.9, i.e. 2.1% faster overall. Change-Id: If592ee99be09bcd34a7c8498347f44e7305e982c
2013-06-19Add two-pass quantizationYunqing Wang
Optimized the quantization function by making it a two-pass process. The first pass does a quick checking of the transform coefficients against the base ZBIN, and only keep the good enough set of coefficients for quantization. A skipping check is added. If all coefficients are within the base ZBIN, no quantization is needed. The second pass is the actual quantization pass, which only processes the coefficient subset determined in first pass. This reduces the computation. Furthermore, an alternitive method is used for large transform size, which often has sparse nonzero quantized coefficients. Overall, the encoder speedup is about 4%. The quantization function itself gets 20% faster. Change-Id: I3a9dd0da6db030260b6d9c314a9fa48ecae89f22
2013-05-23Merge Scatter Scan experiment.Paul Wilkins
Removal from under configure flag. A bit renaming Change-Id: I2213229dfe852001dfec16b149f47c52ce88f3aa
2013-05-16Initial version of alpha channel supportJohn Koleszar
This is a mostly-working implementation of an extra channel in the bitstream. Configure with --enable-alpha to test. Notable TODOs: - Add extra channel to all mismatch tests, PSNR, SSIM, etc - Configurable subsampling - Variable number of planes (currently always uses all 4) - Loop filtering - Per-plane lossless quantizer - ARNR support This implementation just uses the same contents as the Y channel for the A channel, due to lack of content and general pain in playing back 4 channel content. A later patch will use the actual alpha channel passed in from outside the codec. Change-Id: Ibf81f023b1c570bd84b3064e9b4b8ae52e087592
2013-05-07Renaming Y1 and UV quant prefixes to y_ and uv_.Dmitry Kovalev
Change-Id: If57e360c187a475fc90edb8c7170f498efcb31a5
2013-05-07Adjust q rangePaul Wilkins
Skip Q values between the q.0 mode and a real q of 2.0 as these are not valuable from an RD perspective. Change-Id: I110c4858c57f97315953f4d88a2596d4764360df
2013-05-07Merge SB8X8 into the codebaseJingning Han
Pull sb8x8 out of experimental list. verified via borg run tests. Fixed unit test failures. Change-Id: I12a4bbd17395930580c048ab68becad1ffe46e76
2013-05-03Merge "Separate transform and quant from vp9_encode_sb" into experimentalJohn Koleszar
2013-05-03Separate transform and quant from vp9_encode_sbJohn Koleszar
This allows removing a large number of transform size specific functions, as well as supporting 444/alpha by routing all code through the subsampling-aware path. Change-Id: Ieb085cebe9f37f24fc24de179898b22abfda08a4
2013-05-03Merge "Create common vp9_encode_sb{,y}" into experimentalJohn Koleszar
2013-05-03Merge "Reduced y_dequant, uv_dequant size" into experimentalScott LaVarnway
2013-05-02Create common vp9_encode_sb{,y}John Koleszar
Creates a common encode (subtract, transform, quantize, optimize, inverse transform, reconstruct) function for all sb sizes, including the old 16x16 path. Change-Id: I964dff1ea7a0a5c378046a069ad83495f54df007
2013-05-01Reduced y_dequant, uv_dequant sizeScott LaVarnway
Currently, only two values are used. Removed the unused values. Change-Id: Idc5b8be354d84ffc68df39ea3e45f9f50d977b35
2013-05-01Merge "Adding vp9_get_qindex function." into experimentalDmitry Kovalev
2013-04-30sb8x8 integration in rd loop.Ronald S. Bultje
Work-in-progress, not yet ready for review. TODO items: - bitstream writing (encoder) and reading (decoder) - decoder reconstruction Change-Id: I5afb7284e7e0480847b47cd0097cb469433c9081
2013-04-30Adding vp9_get_qindex function.Dmitry Kovalev
Moving common code from encoder and decoder to vp9_get_qindex function. Also moving quant-related constants from vp9_onyxc_int.h to vp9_quant_common.h. Change-Id: I70c5bfbaa1c8bf00fde0bfc459d077f88b6d46c8
2013-04-26Consistent names for quant-related functions and variables.Dmitry Kovalev
Change-Id: I3a6d601e90e8740b9c26dd0afbfe9d467b75d367
2013-04-26Merge "quantize: make 4x4, 8x8 common with larger transforms" into experimentalJohn Koleszar
2013-04-25quantize: make 4x4, 8x8 common with larger transformsJohn Koleszar
There were 4 variants of the quantize loop in vp9_quantize.c, now there is 1. Change-Id: Ic853393411214b32d46a6ba53769413bd14e1cac
2013-04-25Use b_width/height_log2 instead of mb_ where appropriate.Ronald S. Bultje
Basic assumption: when talking about transform units, use b_; when talking about macroblock indices, use mb_. Change-Id: Ifd163f595d4924ff892de4eb0401ccd56dc81884
2013-04-25Move dequant from BLOCKD to per-plane MACROBLOCKDJohn Koleszar
This data can vary per-plane, but not per-block. Change-Id: I1971b0b2c2e697d2118e38b54ef446e52f63c65a
2013-04-23Move skip_block from BLOCK to MACROBLOCKJohn Koleszar
This data is fixed at the MB level, so move it to the common part of MACROBLOCK. Change-Id: Idd8c87118e501cdf0a202bd84c28b502a8234edf
2013-04-23Move quantizer data from BLOCK to MACROBLOCKJohn Koleszar
Quantizers can vary per plane, but not per block. Move these values to the per-plane part of MACROBLOCK. Change-Id: I320a55e38b7b28b29aec751a4aca5ccd0c9b9326
2013-04-23Merge "Convert coeff to per-plane MACROBLOCK data" into experimentalJohn Koleszar
2013-04-23Convert coeff to per-plane MACROBLOCK dataJohn Koleszar
This commit moves the coeff storage from the MACROBLOCK struct to its per-plane part. The next commit will remove the coeff member from the BLOCK structure so that it is consistently accessed per-plane. Also refactors vp9_sb_block_error_c and vp9_sb_uv_block_error_c to be variable subsampling aware. Change-Id: I18c30f87f27c3a012119b6c1970d5fa499804455
2013-04-22Adding get_scan_{4x4, 8x8, 16x16} functions.Dmitry Kovalev
Change-Id: Id4306ef6d65d4a3984aed50b775bdf48d4f6c438
2013-04-22Removes the code_nonzerocount experimentDeb Mukherjee
This patch does not seem to give any benefits. Change-Id: I9d2b4091d6af3dfc0875f24db86c01e2de57f8db
2013-04-16Renaming y1dc_delta_q, uvdc_delta_q, uvac_delta_q fields from VP9Common.Dmitry Kovalev
New names are y_dc_delta_q, uv_dc_delta_q, uv_ac_delta_q. Change-Id: I4acae1fc23a4697ce2c5a5becb8dc28ef0a4b552
2013-04-15Fix width/height switch-up in U/V SB quantize code.Ronald S. Bultje
Change-Id: I697514efd6024e1b4153bbde58ae5e323b030981
2013-04-11Remove unused macroblock versions of reconstruction functions.Ronald S. Bultje
More specifically, remove vp9_quantize_mb*, vp9_optimize_mb*, vp9_inverse_transform_mb* and vp9_transform_mb*. Instead, use the generic _sb* functions that take a size argument, and call them with BLOCK_SIZE_MB16X16. Change-Id: I33024afea95d3a23ffbc1df7da426e4645110f29
2013-04-10Merge "Fixing upper case names." into experimentalDmitry Kovalev
2013-04-09Make SB coding size-independent.Ronald S. Bultje
Merge sb32x32 and sb64x64 functions; allow for rectangular sizes. Code gives identical encoder results before and after. There are a few macros for rectangular block sizes under the sbsegment experiment; this experiment is not yet functional and should not yet be used. Change-Id: I71f93b5d2a1596e99a6f01f29c3f0a456694d728
2013-04-09Fixing upper case names.Dmitry Kovalev
Renaming Y1dequant to y_dequant, UVdequant to uv_dequant, QIndex to qindex. Change-Id: I1c356e5f886deb3f8807dc212de9799b55b09d58
2013-04-04Move EOB to per-plane dataJohn Koleszar
Continue migrating data from BLOCKD/MACROBLOCKD to the per-plane structures. Change-Id: Ibbfa68d6da438d32dcbe8df68245ee28b0a2fa2c
2013-04-04Move qcoeff, dqcoeff from BLOCKD to per-plane dataJohn Koleszar
Start grouping data per-plane, as part of refactoring to support additional planes, and chroma planes with other-than 4:2:0 subsampling. Change-Id: Idb76a0e23ab239180c818025bae1f36f1608bb23
2013-03-26Add col/row-based coefficient scanning patterns for 1D 8x8/16x16 ADSTs.Ronald S. Bultje
These are mostly just for experimental purposes. I saw small gains (in the 0.1% range) when playing with this on derf. Change-Id: Ib21eed477bbb46bddcd73b21c5c708a5b46abedc
2013-03-25Modeling default coef probs with distributionDeb Mukherjee
Replaces the default tables for single coefficient magnitudes with those obtained from an appropriate distribution. The EOB node is left unchanged. The model is represeted as a 256-size codebook where the index corresponds to the probability of the Zero or the One node. Two variations are implemented corresponding to whether the Zero node or the One-node is used as the peg. The main advantage is that the default prob tables will become considerably smaller and manageable. Besides there is substantially less risk of over-fitting for a training set. Various distributions are tried and the one that gives the best results is the family of Generalized Gaussian distributions with shape parameter 0.75. The results are within about 0.2% of fully trained tables for the Zero peg variant, and within 0.1% of the One peg variant. The forward updates are optionally (controlled by a macro) model-based, i.e. restricted to only convey probabilities from the codebook. Backward updates can also be optionally (controlled by another macro) model-based, but is turned off by default. Currently model-based forward updates work about the same as unconstrained updates, but there is a drop in performance with backward-updates being model based. The model based approach also allows the probabilities for the key frames to be adjusted from the defaults based on the base_qindex of the frame. Currently the adjustment function is a placeholder that adjusts the prob of EOB and Zero node from the nominal one at higher quality (lower qindex) or lower quality (higher qindex) ends of the range. The rest of the probabilities are then derived based on the model from the adjusted prob of zero. Change-Id: Iae050f3cbcc6d8b3f204e8dc395ae47b3b2192c9
2013-03-08Merge "Consistent usage of ROUND_POWER_OF_TWO macro." into experimentalDmitry Kovalev
2013-03-07Consistent usage of ROUND_POWER_OF_TWO macro.Dmitry Kovalev
Change-Id: I44660975e9985310d8c654c158ee7a61291b5a08
2013-03-07Re-add support for ADST in superblocks.Ronald S. Bultje
This also changes the RD search to take account of the correct block index when searching (this is required for ADST positioning to work correctly in combination with tx_select). Change-Id: Ie50d05b3a024a64ecd0b376887aa38ac5f7b6af6
2013-03-07Coding con-zero count rather than EOB for coeffsDeb Mukherjee
This patch revamps the entropy coding of coefficients to code first a non-zero count per coded block and correspondingly remove the EOB token from the token set. STATUS: Main encode/decode code achieving encode/decode sync - done. Forward and backward probability updates to the nzcs - done. Rd costing updates for nzcs - done. Note: The dynamic progrmaming apporach used in trellis quantization is not exactly compatible with nzcs. A suboptimal approach has been used instead where branch costs are updated to account for changes in the nzcs. TODO: Training the default probs/counts for nzcs Change-Id: I951bc1e22f47885077a7453a09b0493daa77883d
2013-03-04Make superblocks independent of macroblock code and data.Ronald S. Bultje
Split macroblock and superblock tokenization and detokenization functions and coefficient-related data structs so that the bitstream layout and related code of superblock coefficients looks less like it's a hack to fit macroblocks in superblocks. In addition, unify chroma transform size selection from luma transform size (i.e. always use the same size, as long as it fits the predictor); in practice, this means 32x32 and 64x64 superblocks using the 16x16 luma transform will now use the 16x16 (instead of the 8x8) chroma transform, and 64x64 superblocks using the 32x32 luma transform will now use the 32x32 (instead of the 16x16) chroma transform. Lastly, add a trellis optimize function for 32x32 transform blocks. HD gains about 0.3%, STDHD about 0.15% and derf about 0.1%. There's a few negative points here and there that I might want to analyze a little closer. Change-Id: Ibad7c3ddfe1acfc52771dfc27c03e9783e054430
2013-02-27Move eob from BLOCKD to MACROBLOCKD.Ronald S. Bultje
Consistent with VP8. Change-Id: I8c316ee49f072e15abbb033a80e9c36617891f07
2013-02-22Experimental removal of over quant codePaul Wilkins
The over quant code was added in VP8 post bitstream freeze to allow compression to lower data rates In VP9 the real qualtizer range has been greatly extended anyway. Change-Id: I5d384fa5e9a83ef75a3df34ee30627bd21901526
2013-02-19Use lossless for Q0Yaowu Xu
The commit changes the coding mode to lossless whenever the lowest quantizer is choosen. As expected, test results showed no difference for cif and std-hd set where Q0 is rarely used. For yt and yt-hd set, Q0 is used for a number of clips, where this commit helped a lot in the high end. Average over all clips in the sets: yt: 2.391% 1.017% 1.066% hd: 1.937% .764% .787% Change-Id: I9fa9df8646fd70cb09ffe9e4202b86b67da16765
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-11 05:29:45 +0000