CHANGELOG 13.8 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
2011-08-02 v0.9.7 "Cayuga"
  Our third named release, focused on a faster, higher quality, encoder.

  - Upgrading:
    This release is backwards compatible with Aylesbury (v0.9.5) and
    Bali (v0.9.6). Users of older releases should refer to the Upgrading
    notes in this document for that release.

  - Enhancements:
          Stereo 3D format support for vpxenc
          Runtime detection of available processor cores.
          Allow specifying --end-usage by enum name
          vpxdec: test for frame corruption
          vpxenc: add quantizer histogram display
          vpxenc: add rate histogram display
          Set VPX_FRAME_IS_DROPPABLE
          update configure for ios sdk 4.3
          Avoid text relocations in ARM vp8 decoder
          Generate a vpx.pc file for pkg-config.
          New ways of passing encoded data between encoder and decoder.

  - Speed:
      This release includes across-the-board speed improvements to the
      encoder. On x86, these measure at approximately 11.5% in Best mode,
      21.5% in Good mode (speed 0), and 22.5% in Realtime mode (speed 6).
      On ARM Cortex A9 with Neon extensions, real-time encoding of video
      telephony content is 35% faster than Bali on single core and 48%
      faster on multi-core. On the NVidia Tegra2 platform, real time
      encoding is 40% faster than Bali.

      Decoder speed was not a priority for this release, but improved
      approximately 8.4% on x86.

          Reduce motion vector search on alt-ref frame.
          Encoder loopfilter running in its own thread
          Reworked loopfilter to precalculate more parameters
          SSE2/SSSE3 optimizations for build_predictors_mbuv{,_s}().
          Make hor UV predict ~2x faster (73 vs 132 cycles) using SSSE3.
          Removed redundant checks
          Reduced structure sizes
          utilize preload in ARMv6 MC/LPF/Copy routines
          ARM optimized quantization, dfct, variance, subtract
          Increase chrow row alignment to 16 bytes.
          disable trellis optimization for first pass
          Write SSSE3 sub-pixel filter function
          Improve SSE2 half-pixel filter funtions
          Add vp8_sub_pixel_variance16x8_ssse3 function
          Reduce unnecessary distortion computation
          Use diamond search to replace full search
          Preload reference area in sub-pixel motion search (real-time mode)

  - Quality:
      This release focused primarily on one-pass use cases, including
      video conferencing. Low latency data rate control was significantly
      improved, improving streamability over bandwidth constrained links.
      Added support for error concealment, allowing frames to maintain
      visual quality in the presence of substantial packet loss.

          Add rc_max_intra_bitrate_pct control
          Limit size of initial keyframe in one-pass.
          Improve framerate adaptation
          Improved 1-pass CBR rate control
          Improved KF insertion after fades to still.
          Improved key frame detection.
          Improved activity masking (lower PSNR impact for same SSIM boost)
          Improved interaction between GF and ARFs
          Adding error-concealment to the decoder.
          Adding support for independent partitions
          Adjusted rate-distortion constants


  - Bug Fixes:
          Removed firstpass motion map
          Fix parallel make install
          Fix multithreaded encoding for 1 MB wide frame
          Fixed iwalsh_neon build problems with RVDS4.1
          Fix semaphore emulation, spin-wait intrinsics on Windows
          Fix build with xcode4 and simplify GLOBAL.
          Mark ARM asm objects as allowing a non-executable stack.
          Fix vpxenc encoding incorrect webm file header on big endian


83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
2011-03-07 v0.9.6 "Bali"
  Our second named release, focused on a faster, higher quality, encoder.

  - Upgrading:
    This release is backwards compatible with Aylesbury (v0.9.5). Users
    of older releases should refer to the Upgrading notes in this
    document for that release.

  - Enhancements:
      vpxenc --psnr shows a summary when encode completes
      --tune=ssim option to enable activity masking
      improved postproc visualizations for development
      updated support for Apple iOS to SDK 4.2
      query decoder to determine which reference frames were updated
      implemented error tracking in the decoder
      fix pipe support on windows

  - Speed:
      Primary focus was on good quality mode, speed 0. Average improvement
      on x86 about 40%, up to 100% on user-generated content at that speed.
      Best quality mode speed improved 35%, and realtime speed 10-20%. This
      release also saw significant improvement in realtime encoding speed
      on ARM platforms.

        Improved encoder threading
        Dont pick encoder filter level when loopfilter is disabled.
        Avoid double copying of key frames into alt and golden buffer
        FDCT optimizations.
        x86 sse2 temporal filter
        SSSE3 version of fast quantizer
        vp8_rd_pick_best_mbsegmentation code restructure
        Adjusted breakout RD for SPLITMV
        Changed segmentation check order
        Improved rd_pick_intra4x4block
        Adds armv6 optimized variance calculation
        ARMv6 optimized sad16x16
        ARMv6 optimized half pixel variance calculations
        Full search SAD function optimization in SSE4.1
        Improve MV prediction accuracy to achieve performance gain
        Improve MV prediction in vp8_pick_inter_mode() for speed>3

  - Quality:
      Best quality mode improved PSNR 6.3%, and SSIM 6.1%. This release
      also includes support for "activity masking," which greatly improves
      SSIM at the expense of PSNR. For now, this feature is available with
      the --tune=ssim option. Further experimentation in this area
      is ongoing. This release also introduces a new rate control mode
      called "CQ," which changes the allocation of bits within a clip to
      the sections where they will have the most visual impact.

        Tuning for the more exact quantizer.
        Relax rate control for last few frames
        CQ Mode
        Limit key frame quantizer for forced key frames.
        KF/GF Pulsing
        Add simple version of activity masking.
        make rdmult adaptive for intra in quantizer RDO
        cap the best quantizer for 2nd order DC
        change the threshold of DC check for encode breakout

  - Bug Fixes:
      Fix crash on Sparc Solaris.
      Fix counter of fixed keyframe distance
      ARNR filter pointer update bug fix
      Fixed use of motion percentage in KF/GF group calc
      Changed condition for using RD in Intra Mode
      Fix encoder real-time only configuration.
      Fix ARM encoder crash with multiple token partitions
      Fixed bug first cluster timecode of webm file is wrong.
      Fixed various encoder bugs with odd-sized images
      vp8e_get_preview fixed when spatial resampling enabled
      quantizer: fix assertion in fast quantizer path
      Allocate source buffers to be multiples of 16
      Fix for manual Golden frame frequency
      Fix drastic undershoot in long form content


John Koleszar's avatar
John Koleszar committed
160
2010-10-28 v0.9.5 "Aylesbury"
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
  Our first named release, focused on a faster decoder, and a better encoder.

  - Upgrading:
    This release incorporates backwards-incompatible changes to the
    ivfenc and ivfdec tools. These tools are now called vpxenc and vpxdec.

    vpxdec
      * the -q (quiet) option has been removed, and replaced with
        -v (verbose). the output is quiet by default. Use -v to see
        the version number of the binary.

      * The default behavior is now to write output to a single file
        instead of individual frames. The -y option has been removed.
        Y4M output is the default.

      * For raw I420/YV12 output instead of Y4M, the --i420 or --yv12
        options must be specified.

          $ ivfdec -o OUTPUT INPUT
          $ vpxdec --i420 -o OUTPUT INPUT

      * If an output file is not specified, the default is to write
        Y4M to stdout. This makes piping more natural.

          $ ivfdec -y -o - INPUT | ...
          $ vpxdec INPUT | ...

      * The output file has additional flexibility for formatting the
        filename. It supports escape characters for constructing a
        filename from the width, height, and sequence number. This
        replaces the -p option. To get the equivalent:

          $ ivfdec -p frame INPUT
          $ vpxdec --i420 -o frame-%wx%h-%4.i420 INPUT

    vpxenc
      * The output file must be specified with -o, rather than as the
        last argument.

          $ ivfenc <options> INPUT OUTPUT
          $ vpxenc <options> -o OUTPUT INPUT

      * The output defaults to webm. To get IVF output, use the --ivf
        option.

          $ ivfenc <options> INPUT OUTPUT.ivf
          $ vpxenc <options> -o OUTPUT.ivf --ivf INPUT


  - Enhancements:
      ivfenc and ivfdec have been renamed to vpxenc, vpxdec.
      vpxdec supports .webm input
      vpxdec writes .y4m by default
      vpxenc writes .webm output by default
      vpxenc --psnr now shows the average/overall PSNR at the end
      ARM platforms now support runtime cpu detection
      vpxdec visualizations added for motion vectors, block modes, references
      vpxdec now silent by default
      vpxdec --progress shows frame-by-frame timing information
      vpxenc supports the distinction between --fps and --timebase
      NASM is now a supported assembler
      configure: enable PIC for shared libs by default
      configure: add --enable-small
      configure: support for ppc32-linux-gcc
      configure: support for sparc-solaris-gcc

  - Bugs:
      Improve handling of invalid frames
      Fix valgrind errors in the NEON loop filters.
      Fix loopfilter delta zero transitions
      Fix valgrind errors in vp8_sixtap_predict8x4_armv6().
      Build fixes for darwin-icc

  - Speed:
      20-40% (average 28%) improvement in libvpx decoder speed,
      including:
        Rewrite vp8_short_walsh4x4_sse2()
        Optimizations on the loopfilters.
        Miscellaneous improvements for Atom
        Add 4-tap version of 2nd-pass ARMv6 MC filter.
        Improved multithread utilization
        Better instruction choices on x86
        reorder data to use wider instructions
        Update NEON wide idcts
        Make block access to frame buffer sequential
        Improved subset block search
        Bilinear subpixel optimizations for ssse3.
        Decrease memory footprint

      Encoder speed improvements (percentage gain not measured):
        Skip unnecessary search of identical frames
        Add SSE2 subtract functions
        Improve bounds checking in vp8_diamond_search_sadx4()
        Added vp8_fast_quantize_b_sse2

  - Quality:
      Over 7% overall PSNR improvement (6.3% SSIM) in "best" quality
      encoding mode, and up to 60% improvement on very noisy, still
      or slow moving source video

        Motion compensated temporal filter for Alt-Ref Noise Reduction
        Improved use of trellis quantization on 2nd order Y blocks
        Tune effect of motion on KF/GF boost in two pass
        Allow coefficient optimization for good quality speed 0.
        Improved control of active min quantizer for two pass.
        Enable ARFs for non-lagged compress

268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
2010-09-02 v0.9.2
  - Enhancements:
      Disable frame dropping by default
      Improved multithreaded performance
      Improved Force Key Frame Behaviour
      Increased rate control buffer level precision
      Fix bug in 1st pass motion compensation
      ivfenc: correct fixed kf interval, --disable-kf
  - Speed:
      Changed above and left context data layout
      Rework idct calling structure.
      Removed unnecessary MB_MODE_INFO copies
      x86: SSSE3 sixtap prediction
      Reworked IDCT to include reconstruction (add) step
      Swap alt/gold/new/last frame buffer ptrs instead of copying.
      Improve SSE2 loopfilter functions
      Change bitreader to use a larger window.
      Avoid loopfilter reinitialization when possible
  - Quality:
      Normalize quantizer's zero bin and rounding factors
      Add trellis quantization.
      Make the quantizer exact.
      Updates to ARNR filtering algorithm
      Fix breakout thresh computation for golden & AltRef frames
      Redo the forward 4x4 dct
      Improve the accuracy of forward walsh-hadamard transform
      Further adjustment of RD behaviour with Q and Zbin.
  - Build System:
      Allow linking of libs built with MinGW to MSVC
      Fix target auto-detection on mingw32
      Allow --cpu= to work for x86.
      configure: pass original arguments through to make dist
      Fix builds without runtime CPU detection
      msvs: fix install of codec sources
      msvs: Change devenv.com command line for better msys support
      msvs: Add vs9 targets.
      Add x86_64-linux-icc target
  - Bugs:
      Potential crashes on older MinGW builds
      Fix two-pass framrate for Y4M input.
      Fixed simple loop filter, other crashes on ARM v6
      arm: fix missing dependency with --enable-shared
      configure: support directories containing .o
      Replace pinsrw (SSE) with MMX instructions
      apple: include proper mach primatives
      Fixed rate control bug with long key frame interval.
      Fix DSO link errors on x86-64 when not using a version script
      Fixed buffer selection for UV in AltRef filtering


John Koleszar's avatar
John Koleszar committed
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
2010-06-17 v0.9.1
  - Enhancements:
      * ivfenc/ivfdec now support YUV4MPEG2 input and pipe I/O
      * Speed optimizations
  - Bugfixes:
      * Rate control
      * Prevent out-of-bounds accesses on invalid data
  - Build system updates:
      * Detect toolchain to be used automatically for native builds
      * Support building shared libraries
      * Better autotools emulation (--prefix, --libdir, DESTDIR)
  - Updated LICENSE
      * http://webmproject.blogspot.com/2010/06/changes-to-webm-open-source-license.html


John Koleszar's avatar
John Koleszar committed
333
334
335
2010-05-18 v0.9.0
  - Initial open source release. Welcome to WebM and VP8!