The element downstream of mp3enc might only accept certain sample rates or channels,
make sure we relay any restrictions that do exist to upstream when it does a
get_caps() on the sink pad. That way upstream elements like audioresample or
audioconvert can pick a sample rate / channel configuration that will be accepted,
instead of just negotiating to the highest, which might then be rejected.
https://bugzilla.gnome.org/show_bug.cgi?id=641151
When variable framerate is disabled in libx264 (which occurs when using
the zerolatency tuning), libx264 ignores timestamps but still uses the
timebase leading to messed up rate control with our nanosecond timebase.
We work around this issue by setting the timebase to the reciprocal of
the framerate and we validate that the framerate is suitable.
This has been fixed upstream in libx264 but there are non-fixed versions
in the wild so this workaround is still needed.
Fixes bug #632861
Position queries are badly handled for DVDs (probably due to the division in
chapters): the time returned was the time since the start of chapter.
Now ask upstream for position queries, fall back to the old code if upstream
cannot answer the query.
In X264_BUILD >= 78, b-pyramid became a non-boolean so passing a boolean
argument to the option string value causes an error. For < 78 we pass the
boolean value, for >= 78 we use the x264_b_pyramid_names[] array which will
result in passing 'none' for false and 'strict' for true. Other modes can be
set through the option-string property for now.
https://bugzilla.gnome.org/show_bug.cgi?id=626577
x264_encoder_encode() should be called with a NULL picture until at least
x264_encoder_delayed_frames() returns 0. This fixes what appeared to be a
regression in make check due to the recent change in defaults which enabled
b-frames and b-pyramid, both of which I believe increase the number of delayed
frames when encoding.
Use of a rate control method (pass, bitrate, quantizer, etc properties), a
preset and possibly a profile and/or tuning are now the recommended way to
configure x264 through x264enc.
If a preset/tuning are specified then these will define the default values and
the property defaults will be ignored. After this the option-string property is
applied, followed by the user-set properties, fast first pass restrictions and
finally the profile restrictions.
Addresses part of bug #607798
- Make defaults append the appropriate default value to a string. This is
needed to differentiate between something user-set and the actual prop
default.
- Add an internal option string to which _set_property () cases append for the
majority of properties.
- Use gst_x264_enc_parse_options () to clean up application of settings. This
will make order of application with respect to the presets and tunings quite
simple.
Addresses part of bug #607798
Uses new x264 API to apply reduced complexity values to the parameters to
increase encoding speed in the first pass of a multi-pass encode. This does
not impact on final quality.
Addresses part of bug #607798
In X264_BUILD >= 86 there is a new API for applying restrictions to an H.264
Profile. This makes it easier to achieve Baseline Profile for example.
Addresses part of bug #607798
If it is enabled, then it will be main profile instead of baseline.
This ensures maximum compatibility of the output stream until the
encoder configuration interface gets an overhaul with explicit
output profile selection.
Fixes#619776.
Fix off-by-one bug when constructing the avcC header chunk: we
wrote wrong profile info into the header. The first byte in the
SPS NAL we get from x264 is the nal type, not the profile_idc.
Also add some debug logging.
Fixes#610089.
lame's 'lgv' variable is only initialized when the caps
is negotiated, whenever a seek happens before that, it would
attempt to call a function on an empty pointer, causing the crash.
Fixes#603515
liba52 in a52dec-0.7.4 does not have any MMX, MMXEXT or 3DNOW based
IMDCT transform acceleration. It does however have a software based
acceleration using the djbfft library (D.J. Bernstein's library for
fourier transforms - Extremely fast library for floating-point
convolution). So allow liba52 to use it through the MM_ACCEL_DJBFFT
flag.
The liba52 copy in MPlayer sources does have SSE, 3dnowext, 3dnow
and AltiVec implementations, but those are checked for first, and
djbfft is chosen only if none of those is available - good in the
case of some distributions including a port of the MPlayer changes
in their system a52dec library.
The down and upmix code in liba52 doesn't seem to be disturbed by
this additional MM_ACCEL flag and will still use MMX, SSE or 3DNOW
versions if passed from oil_cpu_get_flags (SSE currently is not).
Fixes bug #592787.
Remove dvdnav plugin, it has no future. resindvd is the plugin to
use for dvd playback. Note that dvdnav was never really enabled
in -ugly, you had to edit configure.ac to get it built.
Original commit message from CVS:
* ext/x264/gstx264enc.c: (gst_x264_enc_class_init),
(gst_x264_enc_header_buf), (gst_x264_enc_encode_frame):
Use hyphen in property name, perform safety buffer size check
prior to mem access, and some more parentheses in macro.
