
This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
vpx Multi-Format Codec SDK README - 21 June 2012 Welcome to the WebM VP8 Codec SDK! COMPILING THE APPLICATIONS/LIBRARIES: The build system used is similar to autotools. Building generally consists of "configuring" with your desired build options, then using GNU make to build the application. 1. Prerequisites * All x86 targets require the Yasm[1] assembler be installed. * All Windows builds require that Cygwin[2] be installed. * Building the documentation requires PHP[3] and Doxygen[4]. If you do not have these packages, you must pass --disable-install-docs to the configure script. * Downloading the data for the unit tests requires curl[5] and sha1sum. sha1sum is provided via the GNU coreutils, installed by default on many *nix platforms, as well as MinGW and Cygwin. If coreutils is not available, a compatible version of sha1sum can be built from source[6]. These requirements are optional if not running the unit tests. [1]: http://www.tortall.net/projects/yasm [2]: http://www.cygwin.com [3]: http://php.net [4]: http://www.doxygen.org [5]: http://curl.haxx.se [6]: http://www.microbrew.org/tools/md5sha1sum/ 2. Out-of-tree builds Out of tree builds are a supported method of building the application. For an out of tree build, the source tree is kept separate from the object files produced during compilation. For instance: $ mkdir build $ cd build $ ../libvpx/configure <options> $ make 3. Configuration options The 'configure' script supports a number of options. The --help option can be used to get a list of supported options: $ ../libvpx/configure --help 4. Cross development For cross development, the most notable option is the --target option. The most up-to-date list of supported targets can be found at the bottom of the --help output of the configure script. As of this writing, the list of available targets is: armv5te-android-gcc armv5te-linux-rvct armv5te-linux-gcc armv6-darwin-gcc armv6-linux-rvct armv6-linux-gcc armv7-android-gcc armv7-linux-rvct armv7-linux-gcc mips32-linux-gcc ppc32-darwin8-gcc ppc32-darwin9-gcc ppc64-darwin8-gcc ppc64-darwin9-gcc ppc64-linux-gcc x86-darwin8-gcc x86-darwin8-icc x86-darwin9-gcc x86-darwin9-icc x86-linux-gcc x86-linux-icc x86-solaris-gcc x86-win32-vs7 x86-win32-vs8 x86_64-darwin9-gcc x86_64-linux-gcc x86_64-solaris-gcc x86_64-win64-vs8 universal-darwin8-gcc universal-darwin9-gcc generic-gnu The generic-gnu target, in conjunction with the CROSS environment variable, can be used to cross compile architectures that aren't explicitly listed, if the toolchain is a cross GNU (gcc/binutils) toolchain. Other POSIX toolchains will likely work as well. For instance, to build using the mipsel-linux-uclibc toolchain, the following command could be used (note, POSIX SH syntax, adapt to your shell as necessary): $ CROSS=mipsel-linux-uclibc- ../libvpx/configure In addition, the executables to be invoked can be overridden by specifying the environment variables: CC, AR, LD, AS, STRIP, NM. Additional flags can be passed to these executables with CFLAGS, LDFLAGS, and ASFLAGS. 5. Configuration errors If the configuration step fails, the first step is to look in the error log. This defaults to config.err. This should give a good indication of what went wrong. If not, contact us for support. SUPPORT This library is an open source project supported by its community. Please please email webm-discuss@webmproject.org for help.
Description
Languages
C
80%
C++
9%
Assembly
6.7%
Makefile
1.5%
Shell
1.3%
Other
1.5%