File x265.changes of Package x265 (Revision fe82c33a2ac53f5759bcf5b9431988ce)
Currently displaying revision fe82c33a2ac53f5759bcf5b9431988ce , Show latest
285
1
-------------------------------------------------------------------
2
Sun Aug 28 11:51:23 UTC 2016 - joerg.lorenzen@ki.tng.de
3
4
- Update to version 2.0
5
API and Key Behavior Changes
6
* x265_rc_stats added to x265_picture, containing all RC decision
7
points for that frame.
8
* PTL: high tier is now allowed by default, chosen only if
9
necessary.
10
* multi-pass: First pass now uses slow-firstpass by default,
11
enabling better RC decisions in future passes.
12
* pools: fix behaviour on multi-socketed Windows systems, provide
13
more flexibility in determining thread and pool counts.
14
* ABR: improve bits allocation in the first few frames, abr reset,
15
vbv and cutree improved.
16
New Features
17
* uhd-bd: Enforce Ultra-HD Blu-ray Disc parameters
18
(overrides any other settings).
19
* rskip: Enables skipping recursion to analyze lower CU sizes
20
using heuristics at different rd-levels. Provides good visual
21
quality gains at the highest quality presets.
22
* rc-grain: Enables a new rate control mode specifically for
23
grainy content. Strictly prevents QP oscillations within and
24
between frames to avoid grain fluctuations.
25
* tune grain: A fully refactored and improved option to encode
26
film grain content including QP control as well as analysis
27
options.
28
* asm: ARM assembly is now enabled by default, native or cross
29
compiled builds supported on armv6 and later systems.
30
Misc
31
* An SSIM calculation bug was corrected
32
- soname bump to 87.
33
- Fixed arm.patch.
34
- Added libnuma-devel as buildrequires for arch x86_64 (except
35
for openSUSE 13.1 because libnuma-devel >= 2.0.9 is required).
36
37
-------------------------------------------------------------------
38
Wed Feb 3 13:22:42 UTC 2016 - idonmez@suse.com
39
40
- Update to version 1.9
41
API Changes:
42
* x265_frame_stats returns many additional fields: maxCLL, maxFALL,
43
residual energy, scenecut and latency logging
44
* --qpfile now supports frametype 'K"
45
* x265 now allows CRF ratecontrol in pass N (N greater than or equal to 2)
46
* Chroma subsampling format YUV 4:0:0 is now fully supported and tested
47
New Features:
48
* Quant offsets: This feature allows block level quantization offsets
49
to be specified for every frame. An API-only feature.
50
* --intra-refresh: Keyframes can be replaced by a moving column
51
of intra blocks in non-keyframes.
52
* --limit-modes: Intelligently restricts mode analysis.
53
* --max-luma and --min-luma for luma clipping, optional for HDR use-cases
54
* Emergency denoising is now enabled by default in very low bitrate,
55
VBV encodes
56
Presets and Performance:
57
* Recently added features lookahead-slices, limit-modes, limit-refs
58
have been enabled by default for applicable presets.
59
* The default psy-rd strength has been increased to 2.0
60
* Multi-socket machines now use a single pool of threads that can
61
work cross-socket.
62
63
-------------------------------------------------------------------
64
Fri Nov 27 18:21:04 UTC 2015 - aloisio@gmx.com
65
66
- Update to version 1.8:
67
API Changes:
68
* Experimental support for Main12 is now enabled. Partial
69
assembly support exists.
70
* Main12 and Intra/Still picture profiles are now supported.
71
Still picture profile is detected based on
72
x265_param::totalFrames.
73
* Three classes of encoding statistics are now available
74
through the API.
75
+ x265_stats - contains encoding statistics, available
76
through x265_encoder_get_stats()
77
+ x265_frame_stats and x265_cu_stats - contains frame
78
encoding statistics, available through recon x265_picture
79
* --csv
80
* x265_encoder_log() is now deprecated
81
* x265_param::csvfn is also deprecated
82
* --log-level now controls only console logging, frame
83
level console logging has been removed.
84
* Support added for new color transfer characteristic ARIB
85
STD-B67
86
New Features:
87
* limit-refs
88
+ This feature limits the references analysed for
89
individual CUS.
90
+ Provides a nice tradeoff between efficiency and
91
performance.
92
+ aq-mode 3
93
* A new aq-mode that provides additional biasing for
94
low-light conditions.
95
* An improved scene cut detection logic that allows
96
ratecontrol to manage visual quality at fade-ins and
97
fade-outs better.
98
Preset and Tune Options:
99
* tune grain
100
+ Increases psyRdoq strength to 10.0, and rdoq-level to 2.
101
+ qg-size
102
* Default value changed to 32.
103
- soname bump to 68
104
- Reworked arm.patch for 1.8
105
106
-------------------------------------------------------------------
107
Fri May 29 09:11:02 UTC 2015 - aloisio@gmx.com
108
109
- soname bump to 59
110
- Update to version 1.7
111
* large amount of assembly code optimizations
112
* some preliminary support for high dynamic range content
113
* improvements for multi-library support
114
* some new quality features
115
(full documentation at: http://x265.readthedocs.org/en/1.7)
116
* This release simplifies the multi-library support introduced
117
in version 1.6. Any libx265 can now forward API requests to
118
other installed libx265 libraries (by name) so applications
119
like ffmpeg and the x265 CLI can select between 8bit and 10bit
120
encodes at runtime without the need of a shim library or
121
library load path hacks. See --output-depth, and
122
http://x265.readthedocs.org/en/1.7/api.html#multi-library-interface
123
* For quality, x265 now allows you to configure the quantization
124
group size smaller than the CTU size (for finer grained AQ
125
adjustments). See --qg-size.
126
* x265 now supports limited mid-encode reconfigure via a new public
127
method: x265_encoder_reconfig()
128
* For HDR, x265 now supports signaling the SMPTE 2084 color transfer
129
function, the SMPTE 2086 mastering display color primaries, and the
130
content light levels. See --master-display, --max-cll
131
* x265 will no longer emit any non-conformant bitstreams unless
132
--allow-non-conformance is specified.
133
* The x265 CLI now supports a simple encode preview feature. See
134
--recon-y4m-exec.
135
* The AnnexB NAL headers can now be configured off, via x265_param.bAnnexB
136
This is not configurable via the CLI because it is a function of the
137
muxer being used, and the CLI only supports raw output files. See
138
--annexb
139
Misc:
140
* --lossless encodes are now signaled as level 8.5
141
* --profile now has a -P short option
142
* The regression scripts used by x265 are now public, and can be found at:
143
https://bitbucket.org/sborho/test-harness
144
* x265's cmake scripts now support PGO builds, the test-harness can be
145
used to drive the profile-guided build process.
146
147
-------------------------------------------------------------------
148
Tue Apr 28 20:08:06 UTC 2015 - aloisio@gmx.com
149
150
- soname bumped to 51
151
- Update to stable version 1.6
152
Perfomance changes:
153
* heavy improvements for AVX2 capable platforms
154
(Haswell and later Intel CPUs) and work efficiency
155
improvements for multiple-socket machines.
156
157
API changes:
158
* --threads N replaced by --pools N,N and --lookahead-slices N
159
* --[no-]rdoq-level N - finer control over RDOQ effort
160
* --min-cu-size N - trade-off compression for performance
161
* --max-tu-size N - trade-off compression for performance
162
* --[no-]temporal-layers - code unreferenced B frames in temporal
163
layer 1
164
* --[no-]cip aliases added for --[no-]constrained-intra
165
* Added support for new color transfer functions "smpte-st-2084"
166
and "smpte-st-428
167
* --limit-refs N was added, but not yet implemented
168
* Deprecated x265_setup_primitives() was removed from the public
169
API and is no longer exported DLLs
170
171
Threading changes:
172
* The x265 thread pool has been made NUMA aware.
173
* The --threads parameter, which used to specify a global
174
pool size, has been replaced with a --pools parameter which
175
allows you to specify a pool size per NUMA node (aka CPU socket
176
or package). The default is still to allocate one pool worker
177
thread per logical core on the machine, but with --pools one
178
can isolate those threads to a given socket.
179
* Other than socket isolation, the biggest visible change in the
180
NUMA aware thread pools is the increase in work efficiency.
181
The total utilization will generally decrease but the performance
182
will increase since worker threads spend less time context
183
switching. Also, the threading of the lookahead was made more
184
work-efficient. Each lookahead job is a much larger piece of work.
185
Before (1.5):
186
disable thread pool: --threads 1
187
default thread pool: --threads 0
188
restrict to 4 threads: --threads 4
189
After (1.6):
190
disable thread pools: --pools 0
191
default thread pools: --pools *
192
restrict to 4 threads: --pools 4
193
restrict to 4 threads on socket 1: --pools -,4
194
restrict to all threads on socket 0: --pools +,-
195
196
Multi-lib interface:
197
* In order to support runtime selection of a libx265
198
shared library, we have introduced an x265_api structure
199
and an x265_api_get() function. Applications which use
200
this interface to acquire the libx265 functional interface
201
will be able to use shim libraries to bind a particular build
202
of libx265 at run time. See the API documentation for full
203
details.
204
205
-------------------------------------------------------------------
206
Sun Feb 22 09:07:11 UTC 2015 - aloisio@gmx.com
207
208
- soname bump
209
- Update to stable version 1.5
210
* improvements in Main10 compression efficiency and performance
211
and psycho-visual optimizations now enabled by default
212
Feature additions:
213
* analysis re-use features have been completed
214
* rate control zones have been introduced
215
* --tune grain introduced
216
* deblocking tC and Beta offsets are now configurable
217
* denoise is seperately configurable for inter and intra CUs
218
* frame based CSV logging has been improved
219
* New support for VTune task profiles
220
Presets and defaults:
221
* ultrafast no longer disables the deblocking loop filter
222
* psy-rd defaults to 0.3 (was 0, disabled)
223
* psy-rdoq defaults to 1.0 (was 0, disabled)
224
* aq-mode defaults to 1 (was 2, auto-variance)
225
* 4:2:2 and 4:4:4 encodes no longer generate compliance warnings
226
API changes:
227
* param.rc.rateTolerance has been removed and replaced with a simpler
228
param.rc.bStrictCbr flag.
229
* --log-level debug is now --log-level 4 instead of --log-level 3.
230
A new 'frame' log level was inserted at level 3 in order to support
231
frame level CSV logging without also enabling frame level console
232
logging. Using the string name 'debug' is unambiguous as its
233
behavior has not changed.
234
235
- version 1.4
236
* large refactoring in the analysis code
237
Feature additions:
238
* --pmode (parallel mode decision)
239
* --pme (parallel motion estimation).
240
Presets and defaults:
241
* --amp is now respected in RD levels 2, 3, and 4 (previously only
242
in 5 and 6).
243
* --b-intra is now respected in all RD levels.
244
* --fast-cbf, which has only ever effective at RD levels 5 and 6,
245
is no longer enabled uselessly in the fastest presets.
246
* --weightb is now enabled by default at presets slower, veryslow,
247
and placebo.
248
* --cu-lossless was changed to only attempt a lossless encode of
249
the best lossy encode method. This made --cu-lossless a much less
250
expensive encode option to have enabled, and hopefully made the
251
feature more robust and maintainable.
252
* The upper threshold for --psy-rdoq was raised to 50 (from 10)
253
since the higher values were found to be beneficial for sources
254
with high frequency noise (film grain).
255
* The default thread pool size logic was updated to account for the
256
addition of --pmode and --pme (if WPP is disabled but --pmode or
257
--pme are enabled, a thread pool is still allocated).
258
259
-------------------------------------------------------------------
260
Mon Dec 8 04:53:31 UTC 2014 - crrodriguez@opensuse.org
261
262
- Ensure we use the proper CXXFLAGS, CFLAGS and therefore
263
debuginfo packages are generated correctly.
264
265
-------------------------------------------------------------------
266
Sat Oct 4 15:24:23 UTC 2014 - aloisio@gmx.com
267
268
- Bumped to version hg20140928
269
270
-------------------------------------------------------------------
271
Thu Jun 5 14:55:45 UTC 2014 - guillaume@opensuse.org
272
273
- Fix ARM build with arm.patch from Arch Linux:
274
https://github.com/archlinuxarm/PKGBUILDs/blob/master/extra/x265/arm.patch
275
276
-------------------------------------------------------------------
277
Thu May 08 00:00:00 UTC 2014 - Manfred.Tremmel@iiv.de
278
279
- added baselibs.conf
280
281
-------------------------------------------------------------------
282
Mon Mar 24 00:00:00 UTC 2014 - Manfred.Tremmel@iiv.de
283
284
- initial build of todays mercurial checkout
285