- Add support for immediate Direct3D rendering
- Add probing function for Intel Quick Sync and MS encoder/decoder MFT
- Add GPU accelerated video converter
* complete rewrite to add auto detection for thirdparties libraries
* all features are enabled by default unless --without-xxx is used
* adds support for pkg-config to ease integration on other projects (tp, webrtc2sip...)
****** FIRT REVISION KNOW TO FULLY WORK WITH TP ******
- Detect support for h264, h263, theora and mp4v-es at runtime instead of using macros at compile-time
- Allows configuring audio ptime and video fps
- Adds support for congestion control
- Complete support for RTCP-REMB (http://tools.ietf.org/html/draft-alvestrand-rmcat-remb-00)
- Change way the bitrate is computed (use width, height, motion-rank and fps)
-
- Better interop with WebRTC endpoints (better video quality)
- Lock-free on MediaSessionMgr for better performances on both audio and video
- Re-design the video jitter buffer for better CPU prefs and video quality. Request lost frames (RTC-NACK) as many times as required to deal with RTCP-losses. The FPS guesser is smarter.
- Fix issues on RTP timestamps on video pkts
- Update libsrtp binaries on Android and Windows (Use latest CVS)
- Better interop with other h264-rtp implementations (e.g. gstreamer, bria, cisco, polycom, lync...)
- Fix issue 233 (tinyNET does not compile on MAC + fix/patch)
- Fix issue 234 (tinyDAV does not compile on MAC)
- Fix issue 238 (iOS: Bad audio quality when audio/video call uses cpu intensive audio codec (e.g. g729 or speex))
- Fix issue 239 (Adds support for thread priority setting). Timers and audio/video threads now use high priority.
- Fix issue 242 (Hold/Resume fails when audio driver is opensl-es (Android))
- Fix issue 243 (PictureID in VP8 is not correct (only happens when there is overflow on the first 4 bytes))
- Fix issue 244 (Adds callbacks from codecs to session to signal IDR frames decoding)
- Fix issue 245 (Fail to decode h264 buffer)
- FIx issue 246 (Gnu Autotools: Detect support for monotonic timers in configure.ac)
Allow changing video size after the decoding process start
Allow rotating without scaling to keep ratio
Use right values to compute DirectShow display ratio
- 1080p (Full HD): all platforms supports full HD video negotiation. Off course it depends on your CPU and network bandwidth. The preferred video size could be changed from the QoS/QoS screen.
- Adaptive video jitter buffer: A video jitter buffer with advanced features like error correction, packet loss retransmission, delay recovery...
- RTP/AVPF profile as per RFC 4585
- RTCP: Full support for RTCP (3550) and many extensions such as: PLI (RFC 4585), SLI (RFC 4585), RPSI (RFC 4585), FIR (RFC 5104), NACK (4585), TMMBN (RFC 5104)...
- rtcp-mux as per 5761
- Negotiation of Generic Image Attributes in the SDP as per RFC 6236
- Source-Specific Media Attributes in SDP as per draft-lennox-mmusic-sdp-source-attributes-01
- Explicit Call Transfer as per 3GPP TS 24.629
- Begin adding support for video jitter buffer (will be used to give feedbacks for packet loss-FEC-)
- Move video flipping code to the converter (refactoring)
- Fix issue 62, issue 41 and issue 66
- Fix issues (workaround) on VP8 (frame corruption)
- Update contribution list
bossiel