Template the function for 8/10-bit and use lowlevel bitstream
macros.

6% faster overall on i7 gcc 4.5.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

High bitdepth H.264 needs 32-bit transform coefficients, whereas
dnxhd does not.  This creates a conflict with the templated
functions operating on DCTELEM data.  This patch adds a field
allowing the caller to choose the element size in dsputil_init()
and adds the required functions.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

On some platforms EINVAL could be positive, ensure we return negative values.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

This matches other dsputil functions and simplifies calls.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

G729_BITEXACT is never set, so the code is all dead code.

Mpegvideo would free frames as soon as they're not the next or prev
picture. This is fine for a single-threading model, but fails miserably
in a system where pictures can be referenced (as e.g. last/prev pic)
in other threads. Keeping track of ownership of pictures keeps image
references (e.g. motion vectors, or the reference of a motion vector)
alive as long as the picture data itself is alive.

This also happens to fix make THREADS=[3-16] fate-vsynth[12]-error.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Use of these has been broken ever since the h264 idct was changed
to always use transposed inputs.  Furthermore, they were only
ever used if some *other* non-default idct was requested.

Signed-off-by: Mans Rullgard <mans@mansr.com>

This test program so full of programming errors it is impossible
to make sense of it.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This checks if the set of selected exponent strategies for all blocks in a
channel are in the frame exponent strategy table, and if so, writes the
table index instead of each strategy. This saves up to 7 bits per channel per
frame, so the overall effect on quality is small.

be used in the E-AC-3 encoder.

The change to LOCAL_ALIGNED means the declared object must be an
array and the subsequent test should not use the & operator.

Noticed by Uoti Urpala.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

DCTELEM can never be changed to anything else, too much code
depends on it.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Fixes artefacts with Neverwinter Nights WOTCLogo.bik
(http://drmccoy.de/zeugs/WOTCLogo.bik).
Fixes trac ticket #352.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

It has no advantage and seems to cause compile issues with
some combinations of compiler/compiler flags/PIC.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Mans Rullgard <mans@mansr.com>

This particular part serves to document the optimized code variant.

Signed-off-by: Martin Storsjö <martin@martin.st>

This commit is plain wrong, theres no reason to error out at this point nor is the
error value guranteed to be negative as it should be

This reverts commit 289c60001f.

Signed-off-by: Mans Rullgard <mans@mansr.com>