The encoder has never produced files that could be decoded
with any software and there should be no reason to create
such files anyway.

Channel coupling is an optional AC-3 feature that increases quality by
combining high frequency information from multiple channels into a
single channel. The per-channel high frequency information is sent with
less accuracy in both the frequency and time domains. This allows more
bits to be used for lower frequencies while preserving enough
information to reconstruct the high frequencies.

In -std=c99 mode GCC defines __STRICT_ANSI__ to hide non-ANSI interfaces.
This causes declarations for some POSIX functions to be omitted from system
headers, which causes compilation failures.

The sample on rsync was corrupt, this one is now bitexact w.r.t. JM.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Diego Biurrun <diego@biurrun.de>

This fixes the following error on OS/2:
error: segment name `.text align=16' not recognized
Signed-off-by: Diego Biurrun <diego@biurrun.de>

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

Currently, the url_interrupt_cb callback will abort all IO
after the first received signal. This makes the output files
from e.g. the mov muxer to be unreadable if the transcode is
aborted with ctrl+c.

After this patch, the first signal cleanly breaks out of
the transcoding loop, but won't forcibly abort all IO.
After the second signal is received, the url_interrupt_cb
callback will abort all IO.
Signed-off-by: Martin Storsjö <martin@martin.st>

This avoids doing a division by zero if the option wasn't found,
or wasn't an option of an appropriate type.
Signed-off-by: Martin Storsjö <martin@martin.st>

125->104 cycles on penryn (x86_64 only)

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

It's badly documented and does the same thing as -flags global_header,
so it's redundant.

Add missing multiple inclusion guards; clean up #endif comments;
add missing library prefixes; keep guard names consistent.

Else a later buffer resize in ffio_set_buf_size() will ABORT.
Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

If done before, some parameters aren't known yet.

With svq3/rtp, initializing before some parameters are known
can lead to calling av_malloc(0), which on OS X currently returns
broken pointers.

Fixes compilation with mp3lame enabled.

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

Fixes memory leaks which are the result of overwriting already-initialized
MDCT contexts during context reinitialization, e.g. in valgrind
fate-aac-latm_000000001180bc60.
Signed-off-by: Diego Biurrun <diego@biurrun.de>

Some compilers choke on -Wall, so only add the flag after checking it works.

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

This format is still used by e.g. OS/2.
Signed-off-by: Diego Biurrun <diego@biurrun.de>

This explicitly disables threading for encoding as slices are otherwise
automatically activated.  This should be dropped once option resetting
between files is fully implemented.
Signed-off-by: Mans Rullgard <mans@mansr.com>

This adds a comment field to the report header, suitable for
extra information not covered by the automatic fields.
Signed-off-by: Mans Rullgard <mans@mansr.com>

This is useful e.g. for building in a different filesystem than
where the source is kept.
Signed-off-by: Mans Rullgard <mans@mansr.com>

Since a private class is set for this muxer, the callers will
assume that the private data starts with an AVClass pointer.
If no such member exists, the first few bytes of the struct
will be overwritten, and the class pointer may be broken at
any later time.
Signed-off-by: Martin Storsjö <martin@martin.st>

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

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

This avoid needlessly convertion from floating point to fixed point and back.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

When backing up the top-left border, check that the top-left
(rather than left) MB indeed does belong to our slice. If it
doesn't, backing up has no positive effect but may accidentally
interfere with other threads writing in the same space.

Fixes occasional one-off effects when enabling slice-MT.

The prototype should use the same typedefs as the definition, or it
will fail where int32_t is not int (DOS apparently).
Signed-off-by: Mans Rullgard <mans@mansr.com>

This is consistent with how all other table generation programs are named.
Moreover this ensures that the cos table generation program is correctly
deleted when cleaning the tree.