Signed-off-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Diego Biurrun <diego@biurrun.de>

This prevents having to sign-extend on 64-bit systems with 32-bit ints,
such as x86-64. Also fixes crashes on systems where we don't do it and
arguments are not in registers, such as Win64 for all weight functions.

Output packet size should match avctx->block_align. The target output packet
size is 1024 bytes.
Before:
mono   - 1024 samples -> 512 bytes
stereo - 2048 samples -> 2048 bytes
After:
mono   - 2048 samples -> 1024 bytes
stereo - 1024 samples -> 1024 bytes

This will allow parsers to export the duration of the current frame being
output, if known, instead of using AVCodecContext.frame_size.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

Also, only set bit rate for CBR.

also remove unneeded commented-out full frame data debugging

We now require at least libmp3lame 3.98.3.
lame_encode_buffer_interleaved() still doesn't work for mono, but it does not
"die"; it just expects a stereo interleaved buffer.

it is already set in avcodec_alloc_frame()

Set the pict_type/key_frame properties on the coded picture.

This makes switching to encode2() simpler, because it allows us to know
exactly how large should the output buffer be before we start writing
into it.

Red zone usage is not allowed in the Win64 ABI.

Update api so it will compile again.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

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

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

This allows the following commit to extrapolate better dts for the first
frame. Pts difference between the first two frames is reused as the
difference between pts and dts of the first frame.