svn path=/trunk/theora/; revision=17876

This gives CIF encoding a 37.5% speed up on a Cortex A8.

svn path=/trunk/theora/; revision=17874

The final argument to TIFFReadRGBAImage is 'stopOnError'. That is,
if it's not true then the function will return success even if
loading the actual image data failed. As such, copying the example
code out of the documentation wasn't really what we wanted.

svn path=/trunk/theora/; revision=17856

This is a modified version of png2theora, using libtiff
instead of libpng to read a sequence of tiff image files
and convert them to an Ogg Theora video.

Originally added because the Sintel film project was
giving us tiff frames and waiting for conversion to
png before encoding was tedious. In fact, it nearly
took longer than creating this program.

svn path=/trunk/theora/; revision=17844

svn path=/trunk/theora/; revision=17843

I've always found two different API links for libtheoradec hard
to find on the main documentation page. This reformats them to
be more obvious, and shifts the emphasis further away from the
legacy api.

I've also added notes about which library to link with to the
API groups in codec.h.

svn path=/trunk/theora/; revision=17833

svn path=/trunk/theora/; revision=17832

We normally ask doxygen to generate latex source, but don't
actually run it through latex as part of the standard build.
If desired, a 'refman.pdf' can be created manually with
'make -C doc/libtheora/latex'.

Using latex+dvipdf turns out not to work, and pdflatex generally
does a better job anyway.

svn path=/trunk/theora/; revision=17831

in the previous commit.

svn path=/trunk/theora/; revision=17830

svn path=/trunk/theora/; revision=17829

We try to build the doxygen api documentation if the tool is
available, so it's more appropriate to document this option
as a --disable than an --enable.

svn path=/trunk/theora/; revision=17828

spec.bib was already in $(SPEC_SRCS) and doesn't need to be
listed explicitly.

svn path=/trunk/theora/; revision=17827

svn path=/trunk/theora/; revision=17825

svn path=/trunk/theora/; revision=17824

svn path=/trunk/theora/; revision=17823

svn path=/trunk/theora/; revision=17822

This change inserts a new speed level OC_SP_LEVEL_NOSATD (3)
between OC_SP_LEVEL_FAST_ANALYSIS (still 2) and OC_SP_LEVEL_NOMC
(now 4). The new speed level further accelerates analysis by
using the cheaper SAD (sum of absolute differences) metric
for measuring distortion instead of the more expensive SATD
(sum of absolute transformed differences).

The new level is intermediate in both quality and encoding
speed between the adjacent speed levels, allowing smoother
adaptation to realtime encoding contraints. In particular
it's much closer to realtime 720p30 on current Core2 cpus.

The cheaper SAD metric is now also used for the fastest NOMC
speed level. This is a reduction in quality for this speed level
as well, but it also about 10% faster. Since the NOMC mode is
already very poor quality, used as a last resort when performance
is critical, we deemed this a better option than adding a further
NOMC_SATD mode or keeping the old NOMC behaviour.

To support the new speed level, parallel copies of the OC_MODE_RD
table are kept, one for each metric. The encoder context keeps
track of which on to use based ont eh current speed level. Analysis
code chooses which metric to calculate and use on the fly. The extra
conditionals did not measurably affect encoding time in benchmarks.

The metrics collection and fitting code used to generate modedec.h
is also updated to track both SAD and SATD measurements.

Thanks to Tim Terriberry and Greg Maxwell for their comments and
and contributions in reviewing this patch. The implementation
of oc_enc_frag_intra_sad is based on one from earlier development
work by Greg Maxwell.

svn path=/trunk/theora/; revision=17821

This is helpful for casual benchmarking and performance comparision.
We use the very portable time.h clock() call and the beginning and
end of the main function and calculate the average rate in millions
of pixels per second and fraction of realtime. Unfortunately clock()
isn't accurate enough to be useful for per-frame statistics.

svn path=/trunk/theora/; revision=17820

added in c17760.

svn path=/trunk/theora/; revision=17819

svn path=/trunk/theora/; revision=17787

This was only used when a stream started without a keyframe, and was
 initializing the wrong region of memory.
This was harmless (the region it initialized was still safely
 allocated, it just didn't cover all of the frame we wanted to use
 as a reference frame), but lead to unpredictable results and
 valgrind errors.
See https://bugzilla.mozilla.org/show_bug.cgi?id=625773

svn path=/trunk/theora/; revision=17780

svn path=/trunk/theora/; revision=17771

This example program allocates a minimal encoder context so
it can query settings through th_encode_ctl(). As of r17760
setting the framerate to a non-zero/zero value is required.

svn path=/trunk/theora/; revision=17761

The specification says that the frame rate numerator and
denominator MUST both be greater than zero. This wasn't
checked as part of the structure validation code, and in
fact passing a zero value here caused a segfault.

The code now checks for this in oc_state_init() and returns
TH_EINVAL, which is causes th_encode_alloc() and oc_decode_alloc()
to return NULL.

A check for that NULL return is also added to encoder_example.c
to show that it may fail when an invalid th_info parameter is
passed.

svn path=/trunk/theora/; revision=17760

svn path=/trunk/theora/; revision=17756

Right now we still immediately compute abs(dc), but we will need the signed
 value to do real DC costing.
This commit should not change the output of the encoder.

svn path=/trunk/theora/; revision=17755

svn path=/trunk/theora/; revision=17753

The changes in r17745 made PROC and ENDP directives emit multiple
 lines, leading to mis-matched line numbers between the original RVCT
 source and the GNU translation.
gas allows multiple statements per line, so use them to keep the line
 numbers in sync.
Also, contrary to the commit message in r17745, function names actually
 do show up correctly in valgrind; I was testing the wrong binary.

svn path=/trunk/theora/; revision=17752

Also knock another instruction off the PMOVMSKB emulation.

svn path=/trunk/theora/; revision=17751

The wrong offset was being used for the shift factor in
 oc_enc_enquant_table_fixup_neon(), and the wrong value was being returned from
 oc_enc_quantize_neon() when the entire block was quantized to zero.

svn path=/trunk/theora/; revision=17750

svn path=/trunk/theora/; revision=17749

PROC now marks the corresponding symbol as a function, and ENDP emits
 an appropriate .size directive for it.
These should help tools identify these functions, though valgrind still
 can't associate the function name with the corresponding code.

svn path=/trunk/theora/; revision=17745

I'd found these helpful for script debugging, but really it's
better for scripts to print their own version of this information.
The -q option should really mean no non-error output.

svn path=/trunk/theora/; revision=17744

Removed _BIND_TO_CURRENT_CRT_VERSION preprocessor directive for VS2008 project files.
Now debug version of dump_video and encoder_example work.

svn path=/trunk/theora/; revision=17739

r17728 now expects this to be cleared to zero before entering
 oc_enc_block_transform_quantize(), but we weren't doing so.

svn path=/trunk/theora/; revision=17738

It helps to actually rename the function after changing its declaration.

svn path=/trunk/theora/; revision=17737

Convert references to the stack buffer in the MMX fDCT to use esp-relative
 offsets, saving a register.
Move the MSVC MMX fDCT into the MMXEXT section (as was done for the gcc one),
 since it now requires pshufw for the zig-zagging.

svn path=/trunk/theora/; revision=17736

Fixed one GCC to Visual C asm instruction call. Now the optimized version produces the same output as the C version.

svn path=/trunk/theora/; revision=17734

This removes one of the transposes from the fDCT, and avoids several zig-zag
 lookups during tokenization.
This change also makes the encoder iDCT clear the input buffer like the
 decoder, which can be re-used for the next block, avoiding the need for a
 memcpy or memset in the tokenizer.
This gives a 1.3% speed-up at the default speed-level (1), and a 3.1% speed-up
 at speed-level 2 (for 480p, on x86-64).

svn path=/trunk/theora/; revision=17728

This program shows how to query a few version-dependent library
paramenters, such as the vendor string and maximum speed level,
and prints a summary to stdout in 'header: value' format.

I mostly wanted this as a utility to get the maximum speed level
for benchmark scripts, but thought it structuring it as an example
was better than stuffing it in tools.

It would be reasonable to extend this to return decoder parameters
like the maximum preprocessing level.

svn path=/trunk/theora/; revision=17709