1 ┌───────────────────────────────────────────┐
2 │ RELEASE NOTES for FFmpeg 2.3 "Mandelbrot" │
3 └───────────────────────────────────────────┘
5 The FFmpeg Project proudly presents FFmpeg 2.3 "Mandelbrot", a major
6 release with all the great features committed during the three-month period
7 since the release of FFmpeg 2.2.
9 In this release, there are lots of internal overhauls that make FFmpeg a
10 more accessible project for new developers. Many important new
11 optimizations and features like QTKit and AVFoundation input devices are
12 committed. Contributions done by Libav such as a new native Opus decoder
15 Because of the increasing difficulty to maintain and lack of maintainers,
16 we are very sorry to say that we have removed all Blackfin and SPARC
17 architecture assembly optimizations with the cleanups done. If you are
18 interested in maintaining optimization for these two architecture, feel
19 free to contact us and we will restore the code!
21 Oh, and since this release, this modern-looking release note is provided in
22 addition to the old-style Changelog file, to make it easier for you to
23 focus on the most important features in this release.
27 ┌────────────────────────────┐
29 └────────────────────────────┘
31 FFmpeg 2.3 is completely source-compatible to the FFmpeg 2.2 series. There
32 are however some API deprecations that you need to take care of. Use `git
33 diff n2.2 n2.3 doc/APIchanges` to show the list of added and deprecated
34 APIs. FFmpeg 2.3 includes the following library versions:
37 • libavcodec 55.69.100
38 • libavformat 55.48.100
39 • libavdevice 55.13.102
40 • libavfilter 4.11.100
42 • libswresample 0.19.100
43 • libpostproc 52. 3.100
45 Please refer to the doc/APIChanges file for more information.
47 ┌────────────────────────────┐
49 └────────────────────────────┘
51 We are excited to announce that we have committed new x86 assembly
52 optimization for HEVC, and FFmpeg's audio resampler libswresample. ARM
53 users will get a boost in MLP/TrueHD decoding thanks to new optimization.
54 Decoding Huffyuv also got a major boost from optimization on the C code.
56 Of special interest for Microsoft Visual Studio users, we have also
57 converted some preexisting x86 assembly to NASM/Yasm format compatible
58 with MSVC setup, especially in the area of audio resampling.
60 Another major feature in this release is the introduction of AArch64
61 (ARMv8) assembly optimization. AArch64 is another name for the first
62 64-bit ARM architecture, used by Apple A7 SoC inside iPhone 5S. Some
63 32-bit ARM assembly has already been ported to AArch64, but more work is
66 ┌────────────────────────────┐
67 │ Native Opus decoder │
68 └────────────────────────────┘
70 Opus is an open audio format jointly developed by Xiph.Org, Mozilla,
71 Skype/Microsoft, and Broadcom. It combines the features of the Skype Cilk
72 speech codec and the Xiph.Org CELT music codec into one low-latency
73 codec. Decoding Opus is already possible since FFmpeg 1.0 using the
74 libopus library, but the new Opus native decoder brings a higher level of
77 ┌────────────────────────────┐
78 │ QTKit and AVFoundation │
79 └────────────────────────────┘
81 For OS X users, the new QTKit and AVFoundation devices allow you to use
82 the integrated camera on Macs. AVFoundation is a newer API only available
83 on OS X 10.7 "Lion" or newer. For users with older OS X systems, the
84 QTKit device using the older OS X API is for you.
86 ┌────────────────────────────┐
88 └────────────────────────────┘
90 In this release, stream side data are introduced as AVStream.side_data as
91 a way to store miscellaneous stream-wide information. The format is
92 similar to the previously anonymous structure AVPacket.side_data (now
93 named as AVPacketSideData). With this change, audio ReplayGain
94 information and video rotation matrix are now exported through this API,
95 if available in the demuxer.
97 We also have improved libswresample's Doxygen API documentation, so new
98 developers wishing to use FFmpeg's excellent libraries can get started
99 more easily and faster.
101 ┌────────────────────────────┐
102 │ Last But Not Least │
103 └────────────────────────────┘
105 Other interesting new features including hqx video filter, a pixel art
106 scaling filter; a fixed-point AC-3 decoder contributed by Imagination
107 Technologies; an On2 TrueMotion VP7 video decoder; an HTML5 WebVTT
108 subtitle encoder that allows creation of WebVTT from any text-based
109 subtitles; and an 1-bit Direct Stream Digital audio decoder.
111 ┌────────────────────────────┐
112 │ ★ List of New Features │
113 └────────────────────────────┘
115 ┌────────────────────────────┐
116 │ Command line tools │
117 └────────────────────────────┘
119 • Support for decoding through DXVA2 in ffmpeg
121 ┌────────────────────────────┐
123 └────────────────────────────┘
125 • AC3 fixed-point decoding
127 • Alias PIX image encoder and decoder
128 • Improvements to the BRender PIX image decoder
129 • Improvements to the XBM decoder
130 • Improvements to OpenEXR image decoder
131 • Support decoding 16-bit RLE SGI images
132 • Direct Stream Digital (DSD) decoder
133 • On2 AVC (Audio for Video) decoder
134 • Native Opus decoder
137 ┌────────────────────────────┐
139 └────────────────────────────┘
142 • GDI screen grabbing for Windows
143 • AVFoundation input device
145 ┌────────────────────────────┐
147 └────────────────────────────┘
150 • Phantom Cine demuxer
151 • Alternative rendition support for HTTP Live Streaming
152 • Magic Lantern Video (MLV) demuxer
153 • Image format auto-detection
154 • LRC lyric file demuxer and muxer
155 • Samba protocol (via libsmbclient)
156 • WebM DASH Manifest muxer
158 ┌────────────────────────────┐
160 └────────────────────────────┘
162 • shuffleplanes filter
163 • libbs2b-based stereo-to-binaural audio filter
164 • showcqt multimedia filter
167 • hqx filter (hq2x, hq3x, hq4x)
169 • libfribidi support in drawtext
171 ┌────────────────────────────┐
172 │ ⚠ Behaviour changes │
173 └────────────────────────────┘
175 • libx264 reference frames count is now limited depending on level chosen
176 • Because of the new image format auto-detection feature, you don't need to
177 specify image format when decoding an image with no extension.