Original commit message from CVS:
* gst/deinterlace2/tvtime/greedyh.asm:
Always use the C implementation if width is not a multiple of 4. The
assembly optimized version only handle this and calling the C
implementation for the remaining part doesn't work because it needs
previous calculations.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedyh.asm:
* gst/deinterlace2/tvtime/greedyh.c:
* gst/deinterlace2/tvtime/greedyhmacros.h:
Some cleanup, use 3DNOW instead of TDNOW in macros.
* gst/deinterlace2/tvtime/tomsmocomp.c:
(gst_deinterlace_method_tomsmocomp_class_init):
* gst/deinterlace2/tvtime/tomsmocomp/TomsMoCompAll.inc:
* gst/deinterlace2/tvtime/tomsmocomp/tomsmocompmacros.h:
The SSE method in fact only needs MMXEXT, declare it as such.
Original commit message from CVS:
* ext/spc/gstspc.c: (spc_setup):
* gst/deinterlace2/tvtime/tomsmocomp/SearchLoopTop.inc:
Don't use declarations after statements in the remaining code.
Original commit message from CVS:
* gst/deinterlace2/tvtime/tomsmocomp/TomsMoCompAll2.inc:
Mark internal processing functions as static inline for quite some
speedup as they're used only once and need to get many local variables
passed as parameter.
Original commit message from CVS:
* gst/deinterlace2/gstdeinterlace2.c:
(gst_deinterlace_method_get_latency),
(gst_deinterlace2_set_method), (gst_deinterlace2_class_init),
(gst_deinterlace2_push_history), (gst_deinterlace2_chain),
(gst_deinterlace2_setcaps), (gst_deinterlace2_src_query):
* gst/deinterlace2/gstdeinterlace2.h:
Include latency of the method in the returned latency.
Fix outputting of all fields, i.e. doubling of the framerate.
Original commit message from CVS:
* gst/deinterlace2/Makefile.am:
* gst/deinterlace2/gstdeinterlace2.c:
(gst_deinterlace_method_class_init), (gst_deinterlace_method_init),
(gst_deinterlace_method_deinterlace_frame),
(gst_deinterlace_method_get_fields_required),
(gst_deinterlace2_methods_get_type), (_do_init),
(gst_deinterlace2_set_method), (gst_deinterlace2_class_init),
(gst_deinterlace2_child_proxy_get_child_by_index),
(gst_deinterlace2_child_proxy_get_children_count),
(gst_deinterlace2_child_proxy_interface_init),
(gst_deinterlace2_init), (gst_deinterlace2_finalize),
(gst_deinterlace2_chain), (gst_deinterlace2_src_query):
* gst/deinterlace2/gstdeinterlace2.h:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_c),
(deinterlace_greedy_packed422_scanline_mmx),
(deinterlace_greedy_packed422_scanline_mmxext),
(deinterlace_frame_di_greedy),
(gst_deinterlace_method_greedy_l_set_property),
(gst_deinterlace_method_greedy_l_get_property),
(gst_deinterlace_method_greedy_l_class_init),
(gst_deinterlace_method_greedy_l_init):
* gst/deinterlace2/tvtime/greedyh.asm:
* gst/deinterlace2/tvtime/greedyh.c: (greedyDScaler_C),
(deinterlace_frame_di_greedyh),
(gst_deinterlace_method_greedy_h_set_property),
(gst_deinterlace_method_greedy_h_get_property),
(gst_deinterlace_method_greedy_h_class_init),
(gst_deinterlace_method_greedy_h_init):
* gst/deinterlace2/tvtime/greedyh.h:
* gst/deinterlace2/tvtime/plugins.h:
* gst/deinterlace2/tvtime/tomsmocomp.c:
(gst_deinterlace_method_tomsmocomp_set_property),
(gst_deinterlace_method_tomsmocomp_get_property),
(gst_deinterlace_method_tomsmocomp_class_init),
(gst_deinterlace_method_tomsmocomp_init):
* gst/deinterlace2/tvtime/tomsmocomp.h:
* gst/deinterlace2/tvtime/tomsmocomp/TomsMoCompAll.inc:
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_frame_vfir),
(gst_deinterlace_method_vfir_class_init),
(gst_deinterlace_method_vfir_init):
Use a GstObject subtype for the deinterlacing methods and export
the different settings for each deinterlacing method via GObject
properties.
Implement GstChildProxy interface to allow access to the used
deinterlacing method and to allow adjusting the different settings.
Move global variables of the tomsmocomp deinterlacing method into
function local variables to make it possible to use this deinterlacing
method from different instances.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedyh.asm:
Support widths that are not a multiply of 4 when using the assembly
optimized greedyh implementations.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedyh.c:
(deinterlace_frame_di_greedyh):
Only build the assembly optimized implementations on x86.
Original commit message from CVS:
* gst/deinterlace2/Makefile.am:
* gst/deinterlace2/tvtime/tomsmocomp.c: (tomsmocomp_init),
(tomsmocomp_filter_mmx), (tomsmocomp_filter_3dnow),
(tomsmocomp_filter_sse), (deinterlace_frame_di_tomsmocomp):
* gst/deinterlace2/tvtime/tomsmocomp.h:
Remove useless file and mark everything possible as static.
* gst/deinterlace2/tvtime/greedy.c:
* gst/deinterlace2/tvtime/greedyh.c:
Use "_stdint.h" instead of <stdint.h>.
Original commit message from CVS:
* gst/deinterlace2/Makefile.am:
* gst/deinterlace2/gstdeinterlace2.c: (gst_deinterlace2_init):
* gst/deinterlace2/tvtime/greedy.c: (deinterlace_frame_di_greedy):
* gst/deinterlace2/tvtime/greedyh.c:
(deinterlace_frame_di_greedyh):
* gst/deinterlace2/tvtime/speedtools.h:
* gst/deinterlace2/tvtime/speedy.c:
* gst/deinterlace2/tvtime/speedy.h:
* gst/deinterlace2/tvtime/tomsmocomp.c: (Fieldcopy):
* gst/deinterlace2/tvtime/tomsmocomp/TomsMoCompAll.inc:
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_frame_vfir):
Get rid of speedy.[ch] as we don't use most of it's code anyway
and it doesn't seem to be relicensed to LGPL. Use memcpy() instead
of the speedy memcpy everywhere instead.
* gst/deinterlace2/gstdeinterlace2.h:
Remove many unused declarations.
Original commit message from CVS:
* gst/deinterlace2/gstdeinterlace2.c: (gst_deinterlace2_src_query):
Divide latency be 2 to convert from fields to frames.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_c),
(deinterlace_greedy_packed422_scanline_mmx),
(deinterlace_greedy_packed422_scanline_mmxext),
(deinterlace_frame_di_greedy):
Don't use scanlines function from gstdeinterlace2 as it's
not appropiate for this method. Instead implement deinterlace_frame
function by taking the one from greedyh.
* gst/deinterlace2/tvtime/greedyh.c: (greedyDScaler_C):
Small fix for the C implementation.
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_frame_vfir):
Don't use the scanlines function from gstdeinterlace2 as it's only
used for this method and will be removed. Instead implement
deinterlace_frame function and make it a bit more efficient.
* gst/deinterlace2/gstdeinterlace2.c:
(gst_deinterlace2_class_init), (gst_deinterlace2_set_method),
(gst_deinterlace2_push_history), (gst_deinterlace2_chain),
(gst_deinterlace2_setcaps), (gst_deinterlace2_sink_event),
(gst_deinterlace2_change_state), (gst_deinterlace2_src_event),
(gst_deinterlace2_src_query):
Fix coding style and remove scanlines function as it's unused now.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedyh.asm:
* gst/deinterlace2/tvtime/greedyh.c: (greedyDScaler_C),
(deinterlace_frame_di_greedyh), (dscaler_greedyh_get_method):
* gst/deinterlace2/tvtime/greedyhmacros.h:
Add a C implementation for the greedyh deinterlacing method, clean
up the code a bit and mark the SSE version as MMXEXT as it doesn't
require any SSE instructions.
Original commit message from CVS:
* gst/deinterlace2/gstdeinterlace2.c:
(gst_deinterlace2_set_property), (gst_deinterlace2_chain),
(gst_deinterlace2_setcaps):
If we're outputting all fields the framerate has to be doubled.
Set duration on the outgoing buffers.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_mmx),
(deinterlace_greedy_packed422_scanline_mmxext):
Optimize MMX/MMXEXT implementations a bit by requiring two less
memory accesses and fix the workaround for the missing right shift
on bytes to unset the highest bit of every byte.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_mmxext):
Remove sfence instruction as it's not needed and actually is an SSE
instruction.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_mmx),
(deinterlace_greedy_packed422_scanline):
Add plain MMX implementation for the greedyl method.
Original commit message from CVS:
* gst/deinterlace2/Makefile.am:
Move the assembly includes to noinst_HEADERS where they belong.
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_line_c),
(deinterlace_line_mmx):
Fix C and MMX implementations a bit more.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_c),
(deinterlace_greedy_packed422_scanline_mmxext),
(deinterlace_greedy_packed422_scanline):
Fix the C implementation to produce correct results and optimize the
MMXEXT implementation.
Handle odd widths and don't read over array boundaries in the MMXEXT
implementation.
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_line_c),
(deinterlace_line_mmx), (deinterlace_scanline_vfir):
Fix a small rounding bug in the MMX implementation, the MMX
implementation doesn't actually need MMXEXT instructions so don't mark
it as such.
Handle odd widths in both implementations.
Original commit message from CVS:
* gst/deinterlace2/tvtime/greedy.c:
(deinterlace_greedy_packed422_scanline_sse),
(deinterlace_greedy_packed422_scanline_c),
(deinterlace_greedy_packed422_scanline):
Implement a C version of the greedy low motion algorithm and mark the
assembly optimized version as SSE as it uses SSE instructions
additional to MMX instructions.
Original commit message from CVS:
* gst/deinterlace2/tvtime/vfir.c: (deinterlace_line_mmxext),
(deinterlace_line_c), (deinterlace_scanline_vfir):
Make it possible to use the vfir method on X86 CPUs without MMXEXT too
but use the MMXEXT optimized code whenever possible.
Original commit message from CVS:
* gst/deinterlace2/gstdeinterlace2.c:
(gst_deinterlace2_class_init), (gst_deinterlace2_init),
(gst_deinterlace2_reset_history), (gst_deinterlace2_reset),
(gst_deinterlace2_finalize), (gst_deinterlace2_chain),
(gst_deinterlace2_sink_event), (gst_deinterlace2_change_state),
(gst_deinterlace2_src_query):
* gst/deinterlace2/gstdeinterlace2.h:
Reset element state on PAUSED->READY properly, don't leak any buffers
when finalizing, allocate buffers with gst_pad_alloc_buffer() and
properly return flow returns from gst_pad_push() instead of ignoring them.
Original commit message from CVS:
* configure.ac:
* gst/deinterlace2/Makefile.am:
* gst/deinterlace2/tvtime/greedyh.asm:
* gst/deinterlace2/tvtime/tomsmocomp/SearchLoopBottom.inc:
Fix compilation on generic x86/amd64 and include deinterlace2 in the
build system. Because of several bugs it's still enabled only
by --enable-experimental.
Flacdec outputs 16-bit samples, so let's check if the value of the first
sample is what we expect rather than just the first byte, which may be
different from what we expect depending on the host's endianness. Fixes
the flacdec unit tests on PPC. Also fix a bunch of leaks in the unit
tests to make valgrind happy. Fixes#582420.
We shouldn't register a new GstTag for every unknown tag
we find as this might lead to conflicts and also those
tags are essentially unknown.
Add mappings for some known tags and also convert string
dates to GDate, as found in many FLV files.
Add support for ECMA arrays in script tags. This fixes
seeking on some files that have the seek table stored
inside an ECMA array instead of the normal array.
Original commit message from CVS:
2008-11-24 Julien Moutte <julien@fluendo.com>
* gst/flv/gstflvdemux.c: (gst_flv_demux_find_offset),
(gst_flv_demux_handle_seek_push),
(gst_flv_demux_handle_seek_pull):
Fix non key unit seeking by always going to the previous
keyframe. Mark
the discont flag when we've moved in the file.
* gst/flv/gstflvparse.c: (gst_flv_parse_audio_negotiate): MP3
streams
are parsed already, makes autoplugged pipelines shorter.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_cleanup),
(gst_flv_demux_loop), (gst_flv_demux_handle_seek_push),
(gst_flv_demux_handle_seek_pull), (gst_flv_demux_sink_event),
(gst_flv_demux_dispose), (gst_flv_demux_init):
* gst/flv/gstflvdemux.h:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video), (gst_flv_parse_tag_timestamp):
Put the GstSegment directly into the instance struct instead of
allocating and free'ing it again.
Push tags already if only one pad was added, no need to wait for
the second one.
When generating our index set has_video and has_audio if we find
video or audio in case the FLV header has incorrect data.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_chain),
(gst_flv_demux_pull_tag), (gst_flv_demux_pull_header),
(gst_flv_demux_create_index):
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_script),
(gst_flv_parse_tag_audio), (gst_flv_parse_tag_video),
(gst_flv_parse_tag_timestamp), (gst_flv_parse_tag_type),
(gst_flv_parse_header):
* gst/flv/gstflvparse.h:
Don't memcpy() all data we want to push downstream, instead just
create subbuffers and push them downstream.
Fix some minor memory leaks.
