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

svn path=/trunk/theora/; revision=17708

svn path=/trunk/theora/; revision=17707

Invoking the build with 'scons collect_metrics=1' will
compile the library -DOC_COLLECT_METRICS, which enables
recording training metrics.

svn path=/trunk/theora/; revision=17692

This makes it a little nice to call from a batch job with parallel
executation. The messages indentifying the input file parameters
and error messages are still printed.

svn path=/trunk/theora/; revision=17660

svn path=/trunk/theora/; revision=17649

This was ultimately named OC_ARM_ASM_MEDIA instead.
It shouldn't have mattered in this case, as it shouldn't be possible to get
 OC_ARM_ASM_MEDIA without OC_ARM_ASM_EDSP, but clean it up anyway.

svn path=/trunk/theora/; revision=17646

The SAD and SADT metrics have different ranges, so it is
necessary to use different scale factors when comparing
them.

Originally we used SAD, and when SATD was added, the
scale factor was updated without changing the name.
This commit finally does so. However, it is intended
that the code switch back to the SAD metric at higher
speed levels, we also keep OC_SAD_SHIFT with its
appropriate value, 6 according to Tim.

Likewise, OC_SAD_BINS is renamed OC_COMP_BINS in
expectation of supporting mode tables for both
metrics.

svn path=/trunk/theora/; revision=17645

to the control codes passed.

svn path=/trunk/theora/; revision=17639

Thanks to Basil Gohar for pointing this out.

svn path=/trunk/theora/; revision=17632

Don't change the block qii flags if no ac coefficients will be coded in the selected mode. The code previously tried to do this but failed, instead it desynchronized the qii costing with the coded flags. 0.065dB SSIM improvement on 312kbit/sec foreman_cif.y4m.

svn path=/trunk/theora/; revision=17624

On non-glibc systems (Android), the system headers won't do it for us.
This also simplifies its use, since we can guarantee it's always #define'd
 (i.e., as 0 on non-gcc builds).

svn path=/trunk/theora/; revision=17578

svn path=/trunk/theora/; revision=17577

Signed/unsigned comparison and integer conversion overflow.

svn path=/trunk/theora/; revision=17576

sse2trans.h was including encoder headers, when it can be used by the decoder
 alone.

svn path=/trunk/theora/; revision=17575

sse2idct.c was needlessly including x86enc.h when it is part of the decoder.
collect.c had a trailing space.

svn path=/trunk/theora/; revision=17574

Remove a few unused variables (that were assigned to, but never used), and
 add an explicit cast to convert -OC_LOTS_OF_BITS to unsigned, since the C64x
 compiler complains about the sign change (arguably rightly so).

svn path=/trunk/theora/; revision=17565

svn path=/trunk/theora/; revision=17564

We don't actually use this for anything yet, but it may help calling
 applications (e.g., doing software YUV2RGB conversion).
Also, change ref_frame_data to point directly to the desired reference frame,
 rather than require a lookup through ref_frame_idx first.
This saves an indirection and gives a 0.7% speed-up at 720p on a Cortex A8.
It should have an even bigger benefit on C64x, though it wasn't benchmarked
 there.

svn path=/trunk/theora/; revision=17563

svn path=/trunk/theora/; revision=17550

mcenc.c: Skip the 4MV SATD in speed level 2.  2% speedup.  Avoid duplicating the first set b candidate (correctly adding the surrounding accelerators hurt efficiency).

svn path=/trunk/theora/; revision=17507

This removes some useless clamping and only computes set B predictors when
 they're actually going to be used.

svn path=/trunk/theora/; revision=17506

Coverity thinks we're overrunning this static array by passing it to memset.
In reality, it's a multidimensional array, and guaranteed to be contiguous, but
 we might as well silence this warning.
Reported in https://bugzilla.mozilla.org/show_bug.cgi?id=559343

svn path=/trunk/theora/; revision=17505

This should have had a _c suffix, which r17503 exposed.

svn path=/trunk/theora/; revision=17504