On 12/07/14 7:51 PM, Timothy Gu wrote: > On Sat, Jul 12, 2014 at 3:08 AM, Clément Bœsch <u...@pkh.me> wrote: >> On Fri, Jul 11, 2014 at 09:08:40PM -0700, Timothy Gu wrote: >>> Based on a patch by Clément Bœsch <u...@pkh.me>. >>> >>> Signed-off-by: Timothy Gu <timothyg...@gmail.com> >>> --- >>> >>> So, I rebased Clément's original patch and added (a lot) more stuff to make >>> it >>> a real official modern-looking release note. Some stuff I am not sure about: >>> >>> - Should we use Clément's format with all kinds of Unicode characters or >>> stick to Markdown? >> >> We don't really render Markdown anywhere; here we can just link to the >> raw on the repository and have a decent layout. We don't need complex >> markup, unless we want to link to commit hashes or stuff? > > OK. But the next two questions still stand. > >> >>> - Does the new image auto-detection really work as advertised in "Behaviour >>> changes" section? >>> - Is there any stuff I missed in the "Behaviour changes" section? (I'm sure >>> there are.) > > ----8<---- > > Based on a patch by Clément Bœsch <u...@pkh.me>. > > Signed-off-by: Timothy Gu <timothyg...@gmail.com> > --- > RELEASE_NOTES | 157 > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 157 insertions(+) > create mode 100644 RELEASE_NOTES > > diff --git a/RELEASE_NOTES b/RELEASE_NOTES > new file mode 100644 > index 0000000..5d2830e > --- /dev/null > +++ b/RELEASE_NOTES > @@ -0,0 +1,157 @@ > + ┌─────────────────────────────────┐ > + │ RELEASE NOTES for FFmpeg 2.3 │ > + └─────────────────────────────────┘ > + > + The FFmpeg Project proudly presents FFmpeg 2.3, a new release completely > + source-compatible to the FFmpeg 2.2 series. There are however some API > + deprecations that you need to take care of. Use `git diff n2.2 n2.3 > + doc/APIchanges` to show the list of added and deprecated APIs. > + > + In this release, there are lots of internal overhauls that make FFmpeg a > + more accessible project for new developers. Many important new features > + like native Opus decoder and QTKit and AVFoundation input devices are also > + committed. > + > + Another major feature in this release is the introduction of AArch64 > + (ARMv8) assembly optimization. AArch64 is another name for the first > 64-bit > + ARM architecture, used by Apple A7 SoC inside iPhone 5S. We have added > some > + x86 optimizations to accelerate the decoding of Huffyuv and HEVC, two very > + different codecs in terms of age.
HEVC, Huffyuv and audio resampling with libswresample got the biggest improvements. Afaik, Huffyuv was mostly from Christophe's c code refactoring rather than asm code. DCA and VP9 decoding, and some encoding functions (motion estimation and such) got some minor x86 asm improvements, and MLP/TrueHD got some considerable boost on arm hardware. I'd say that at the very least we should mention libswresample here. Might be worth mentioning as well that a lot of pre-existing inline x86 asm code has been ported to Nasm/Yasm, which is of special interest for MSVC users. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
