gstreamer-omap:mstorsjos-gst-ducati.git
5 years agoMark I frames as keyframes for encoders other than H264 mpeg4enc
Martin Storsjo [Wed, 2 May 2012 17:10:49 +0000 (20:10 +0300)]
Mark I frames as keyframes for encoders other than H264

5 years agoREADME update master
Luciana Fujii Pontello [Fri, 13 Apr 2012 14:13:40 +0000 (10:13 -0400)]
README update

5 years agoducatibufferpool: fix race in gst_ducati_bufferpool_destroy
Alessandro Decina [Wed, 11 Apr 2012 12:08:06 +0000 (14:08 +0200)]
ducatibufferpool: fix race in gst_ducati_bufferpool_destroy

Pop items from ->freelist with the pool lock taken to avoid races with other
threads unreffing outstanding buffers.

5 years agoducatividdec: propagate downstream flow errors
Alessandro Decina [Wed, 11 Apr 2012 12:04:56 +0000 (14:04 +0200)]
ducatividdec: propagate downstream flow errors

5 years agoducatividdec: reset outArgs in flush()
Alessandro Decina [Wed, 11 Apr 2012 12:01:18 +0000 (14:01 +0200)]
ducatividdec: reset outArgs in flush()

Reset outArgs in flush() so that it isn't left in an inconsistent state if
flush() is called from codec_process() during error handling.

5 years agoducatividdec: fix getcaps method adding rowstride
Luciana Fujii Pontello [Sun, 1 Apr 2012 02:34:19 +0000 (22:34 -0400)]
ducatividdec: fix getcaps method adding rowstride

For yuv-strided formats we need to add the rowstride field with the
allowed range to make caps intersect properly.

5 years agoducatividdec: Implement src pad getcaps function
Luciana Fujii Pontello [Thu, 22 Mar 2012 01:42:30 +0000 (21:42 -0400)]
ducatividdec: Implement src pad getcaps function

The caps of the src pad should not change with upstream negotiation
except between regular yuv and rowstrided.

5 years agoducatividdec: Fix indentation
Luciana Fujii Pontello [Thu, 22 Mar 2012 01:41:32 +0000 (21:41 -0400)]
ducatividdec: Fix indentation

5 years agoducatividenc: err.. mark keyframes on IVIDEO_IDR_FRAME frames
Alessandro Decina [Wed, 28 Mar 2012 14:03:26 +0000 (16:03 +0200)]
ducatividenc: err.. mark keyframes on IVIDEO_IDR_FRAME frames

5 years agoducatividenc: add cropping support
Alessandro Decina [Tue, 27 Mar 2012 17:19:08 +0000 (19:19 +0200)]
ducatividenc: add cropping support

5 years agoducatividenc: mark keyframes
Alessandro Decina [Thu, 22 Mar 2012 14:43:28 +0000 (15:43 +0100)]
ducatividenc: mark keyframes

5 years agoducatih264enc: set output caps
Alessandro Decina [Thu, 22 Mar 2012 14:42:34 +0000 (15:42 +0100)]
ducatih264enc: set output caps

5 years agoducatividenc: don't memcpy input if it's already tiler backed
Alessandro Decina [Thu, 22 Mar 2012 14:41:11 +0000 (15:41 +0100)]
ducatividenc: don't memcpy input if it's already tiler backed

5 years agoducatih264dec: implement LATENCY query
Alessandro Decina [Thu, 22 Mar 2012 14:32:38 +0000 (15:32 +0100)]
ducatih264dec: implement LATENCY query

5 years agoducatividdec: store upstream framerate
Alessandro Decina [Thu, 22 Mar 2012 14:31:59 +0000 (15:31 +0100)]
ducatividdec: store upstream framerate

5 years agoducatih264dec: det displayDelay to IVIDDEC3_DISPLAY_DELAY_1
Alessandro Decina [Fri, 17 Feb 2012 11:28:51 +0000 (12:28 +0100)]
ducatih264dec: det displayDelay to IVIDDEC3_DISPLAY_DELAY_1

5 years agoducatividdec: flush the codec on caps changes
Alessandro Decina [Tue, 28 Feb 2012 17:07:33 +0000 (18:07 +0100)]
ducatividdec: flush the codec on caps changes

Flush the codec when caps change. Needed for downstream sinks that provide a
buffer pool and need to destroy all the outstanding buffers before they can
negotiate new caps (hello v4l2sink).

5 years agoducatividdec: send a crop event every time caps change
Alessandro Decina [Tue, 28 Feb 2012 17:04:59 +0000 (18:04 +0100)]
ducatividdec: send a crop event every time caps change

5 years agoducatih264dec: don't leak on decode errors
Alessandro Decina [Fri, 17 Feb 2012 11:28:51 +0000 (12:28 +0100)]
ducatih264dec: don't leak on decode errors

5 years agoducatih264dec: handle IH264VDEC_ERR_NOSLICE errors
Alessandro Decina [Thu, 16 Feb 2012 17:13:13 +0000 (18:13 +0100)]
ducatih264dec: handle IH264VDEC_ERR_NOSLICE errors

Handle IH264VDEC_ERR_NOSLICE errors by flushing and skipping to the next
keyframe.

5 years agoducatividdec: make codec_flush non-static so it can be used by derived classes
Alessandro Decina [Thu, 16 Feb 2012 17:11:05 +0000 (18:11 +0100)]
ducatividdec: make codec_flush non-static so it can be used by derived classes

Rename codec_flush to gst_ducati_viddec_codec_flush and make it non-static so
derived elements can use it (for example during error recovering).

5 years agoducatividdec: add var to skip buffers until the next keyframe
Alessandro Decina [Thu, 16 Feb 2012 17:08:15 +0000 (18:08 +0100)]
ducatividdec: add var to skip buffers until the next keyframe

Add wait_keyframe variable that is set to TRUE whenever we need to skip buffers
until the next keyframe.

5 years agoducatih264dec: implement B frame dropping strategy similar to ducativc1dec
Alessandro Decina [Tue, 14 Feb 2012 16:41:40 +0000 (17:41 +0100)]
ducatih264dec: implement B frame dropping strategy similar to ducativc1dec

5 years agoducatividdec: don't discard input timestamps at stream start
Alessandro Decina [Tue, 14 Feb 2012 15:07:27 +0000 (16:07 +0100)]
ducatividdec: don't discard input timestamps at stream start

Don't discard input timestamps when forcing the allocation of buffers from the
internal pool (see 9471fe2). Fixes sync glitches with interlaced streams.

5 years agoducatividdec: workaround for the mpeg2 decoder giving wrong crop_height for interlace...
Alessandro Decina [Tue, 29 Nov 2011 19:37:30 +0000 (20:37 +0100)]
ducatividdec: workaround for the mpeg2 decoder giving wrong crop_height for interlaced streams

5 years agoducatividdec: don't renegotiate while flushing
Alessandro Decina [Tue, 29 Nov 2011 19:54:47 +0000 (20:54 +0100)]
ducatividdec: don't renegotiate while flushing

Fixes a crash introduced by calling codec_flush in PAUSED=>READY

5 years agoducatividdec: flush the codec going to READY
Alessandro Decina [Thu, 2 Feb 2012 10:39:20 +0000 (11:39 +0100)]
ducatividdec: flush the codec going to READY

Flush the codec going to READY so we don't leak buffers if state is changed to
<= READY before EOS.

5 years agobufferpool: reset ->orig to NULL in _finalize
Alessandro Decina [Thu, 2 Feb 2012 10:35:39 +0000 (11:35 +0100)]
bufferpool: reset ->orig to NULL in _finalize

Reset ->orig to NULL on _finalize so that if a buffer is recycled we don't
_unref invalid memory.

5 years agoducativc1dec: when dropping frames, drop only bframes
Alessandro Decina [Tue, 31 Jan 2012 12:03:34 +0000 (13:03 +0100)]
ducativc1dec: when dropping frames, drop only bframes

5 years agoducatividdec: make it possible for subclasses to participate in qos handling
Alessandro Decina [Tue, 31 Jan 2012 11:59:06 +0000 (12:59 +0100)]
ducatividdec: make it possible for subclasses to participate in qos handling

5 years agoducatimpeg4dec: don't leak input buffers
Alessandro Decina [Tue, 24 Jan 2012 16:35:34 +0000 (17:35 +0100)]
ducatimpeg4dec: don't leak input buffers

5 years agoducatividdec: destroy the internal pool once we have an output buffer
Alessandro Decina [Fri, 20 Jan 2012 10:55:57 +0000 (11:55 +0100)]
ducatividdec: destroy the internal pool once we have an output buffer

We use the internal pool until we get the first output buffer, then we switch to
downstream provided buffers. Destroy the internal pool once we get the first
output buffer so that we release memory. The pool is then recreated when/if
needed.

5 years agoducatividdec: update the internal pool caps when switching interlaced mode
Alessandro Decina [Fri, 20 Jan 2012 10:32:23 +0000 (11:32 +0100)]
ducatividdec: update the internal pool caps when switching interlaced mode

5 years agoset encoders' ranks to PRIMARY + 1
Luciana Fujii Pontello [Fri, 20 Jan 2012 00:45:17 +0000 (19:45 -0500)]
set encoders' ranks to PRIMARY + 1

Increase encoders' ranks to make them be chosen before other
primary encoders.

5 years agogstducati: Fix indentation
Luciana Fujii Pontello [Fri, 20 Jan 2012 00:44:43 +0000 (19:44 -0500)]
gstducati: Fix indentation

5 years agoducatividdec: workaround for interlaced clips that are detected as progressive from...
Alessandro Decina [Thu, 19 Jan 2012 16:07:31 +0000 (17:07 +0100)]
ducatividdec: workaround for interlaced clips that are detected as progressive from upstream

Sometimes upstream thinks that some input is progressive (usually if interlaced
isn't marked correctly at the container level) when it's actually interlaced.
Detect such cases with help from the codec and adjust caps (and buffers)
accordingly.

5 years agoducatividdec: fix regression with clips having outBufsInUseFlag==TRUE
Alessandro Decina [Wed, 11 Jan 2012 16:00:34 +0000 (17:00 +0100)]
ducatividdec: fix regression with clips having outBufsInUseFlag==TRUE

5 years agoducatimpeg4dec: don't forget to push the codec data buffer
Alessandro Decina [Mon, 9 Jan 2012 16:16:54 +0000 (17:16 +0100)]
ducatimpeg4dec: don't forget to push the codec data buffer

5 years agoducatividdec: auto-detect when input ts are PTS
Rob Clark [Tue, 27 Dec 2011 22:03:50 +0000 (16:03 -0600)]
ducatividdec: auto-detect when input ts are PTS

When input timestamps are in ascending order, and output timestamps
are not, that must mean the parser/demuxser is giving us the timestamps
in presentation order (PTS), not decode order (DTS).

5 years agoducatividdec: some minor trace cleanups
Rob Clark [Tue, 27 Dec 2011 21:38:15 +0000 (15:38 -0600)]
ducatividdec: some minor trace cleanups

5 years agoducatimpeg2dec: we want parsed stream
Rob Clark [Tue, 27 Dec 2011 21:36:49 +0000 (15:36 -0600)]
ducatimpeg2dec: we want parsed stream

5 years agoducativc1dec: workaround to not abort decoding on broken frames
Alessandro Decina [Fri, 23 Dec 2011 12:14:16 +0000 (13:14 +0100)]
ducativc1dec: workaround to not abort decoding on broken frames

When on some broken frames the codec returns fatal error 0x00409000, pretend it
isn't fatal and just carry on as the codec seems to be able to continue
decoding subsequent valid frames correctly.

5 years agoducatividdec: make it possible for decoders to implement their own error handling
Alessandro Decina [Fri, 23 Dec 2011 12:10:22 +0000 (13:10 +0100)]
ducatividdec: make it possible for decoders to implement their own error handling

Useful to implement decoder specific hacks to workaround broken error reporting
from codecs.

5 years agompeg4dec: don't enable deblocking filter. Fixes codec creation failures.
Alessandro Decina [Wed, 21 Dec 2011 13:35:30 +0000 (14:35 +0100)]
mpeg4dec: don't enable deblocking filter. Fixes codec creation failures.

Set outloopDeBlocking to FALSE as the codec doesn't seem to support that being
TRUE on recent ducati.

5 years agoducatimpeg2dec: fix mpeg2 playback with 1d buffers
Rob Clark [Sun, 18 Dec 2011 22:59:29 +0000 (16:59 -0600)]
ducatimpeg2dec: fix mpeg2 playback with 1d buffers

5 years agoducatimpeg4dec: fix compiler warnings on glp1.4
Alessandro Decina [Fri, 16 Dec 2011 09:56:14 +0000 (10:56 +0100)]
ducatimpeg4dec: fix compiler warnings on glp1.4

5 years agoducatividdec: fix memory leak with bufferpool buffers
Rob Clark [Mon, 12 Dec 2011 00:02:07 +0000 (18:02 -0600)]
ducatividdec: fix memory leak with bufferpool buffers

If we don't have an input frame, and unref the outbuf, and the original
buffer was replaced by a bufferpool buffer, we need to unref the new
buffer, not the original.

5 years agoducatimpeg4dec: enabling workarounds for XviD too
Rob Clark [Sat, 10 Dec 2011 18:15:20 +0000 (12:15 -0600)]
ducatimpeg4dec: enabling workarounds for XviD too

5 years agoducatividdec: fix seek issue with xvid (ts_is_pts) clips
Rob Clark [Thu, 8 Dec 2011 03:43:35 +0000 (21:43 -0600)]
ducatividdec: fix seek issue with xvid (ts_is_pts) clips

5 years agoducatimpeg4dec/ducatividdec: implement DTS queue
Rob Clark [Sun, 4 Dec 2011 06:57:12 +0000 (00:57 -0600)]
ducatimpeg4dec/ducatividdec: implement DTS queue

When packed frames are used, the timestamps we get from demuxer are
actually PTS, not DTS.  Detect this and fix up the timestamps on
the outgoing buffers.

This commit implements detection of packed frames (PB-frames) in
ducatimpeg4dec, and a DTS queue in the base ducatividdec (which
could be used by other decoders).  When packed frames are detected,
we assume the demuxer doesn't properly understand B-frames, and
therefore timestamps from demuxer must actually be PTS timestamps.
Because the codec is unaware of this, and re-ordering frames to
display order, if we preserve the original timestamps the video
sink ends up seeing timestamps out of order.  To work around this
we pop the DTS from the demuxer out of a queue and use that
instead.

5 years agoducatimpeg4dec: xvid support
Rob Clark [Sun, 4 Dec 2011 05:20:15 +0000 (23:20 -0600)]
ducatimpeg4dec: xvid support

Handle xvid streams with B-frames encoded in containers that don't
support B-frames natively (ie. PB-frames).  Packed frames are split
and passed to the codec in successive _process() calls, and non-coded
frames are discarded.

5 years agoRevert "ducatimpeg4dec: use base VIDDEC3_* structs in allocate_params"
Rob Clark [Sun, 4 Dec 2011 05:12:05 +0000 (23:12 -0600)]
Revert "ducatimpeg4dec: use base VIDDEC3_* structs in allocate_params"

This reverts commit cad4102f744ea191dbce0d2a0d4e637320b387e8.

This commit broke KungFuPanda_1280x720_orig.mp4 which makes me a sad panda.

5 years agoducatividdec: fix NPE if outcaps are EMPTY
Rob Clark [Sat, 3 Dec 2011 23:49:39 +0000 (17:49 -0600)]
ducatividdec: fix NPE if outcaps are EMPTY

5 years agoducatividdec: reduce debug verbosity
Rob Clark [Sun, 4 Dec 2011 05:08:45 +0000 (23:08 -0600)]
ducatividdec: reduce debug verbosity

5 years agoducatividdec: add support for outBufsInUseFlag
Alessandro Decina [Fri, 9 Dec 2011 10:30:35 +0000 (11:30 +0100)]
ducatividdec: add support for outBufsInUseFlag

When outArgs->outBufsInUseFlag is set, the codec needs the current output buffer
to be passed for the next input frame.

5 years agoducativc1dec: remove WMV1/WMV2 support
Alessandro Decina [Wed, 7 Dec 2011 11:00:50 +0000 (12:00 +0100)]
ducativc1dec: remove WMV1/WMV2 support

The codec can only do WMV3/VC1.

5 years agoducatih264dec: specify in template caps that the decoder can do up to profile=high...
Alessandro Decina [Tue, 6 Dec 2011 11:16:01 +0000 (12:16 +0100)]
ducatih264dec: specify in template caps that the decoder can do up to profile=high,level=4.2

5 years agoducatividdec: emit a STREAM DECODE error when codec_process fails
Alessandro Decina [Tue, 6 Dec 2011 10:09:16 +0000 (11:09 +0100)]
ducatividdec: emit a STREAM DECODE error when codec_process fails

5 years agoducatividdec: implement QoS handling
Alessandro Decina [Sun, 4 Dec 2011 17:13:13 +0000 (18:13 +0100)]
ducatividdec: implement QoS handling

Helps a lot with large resolution clips where occasionally, especially at
the beginning, the codecs can't keep up decoding in realtime.

5 years agoducatividdec: set unaligned width and height in the crop event
Alessandro Decina [Tue, 29 Nov 2011 19:13:41 +0000 (20:13 +0100)]
ducatividdec: set unaligned width and height in the crop event

Set unaligned width and height in the crop event. Fixes artifacts at bottom of
some clips.

5 years agobufferpool: alloc 1d buffers if the pool is created with null caps
Alessandro Decina [Mon, 28 Nov 2011 13:38:10 +0000 (14:38 +0100)]
bufferpool: alloc 1d buffers if the pool is created with null caps

5 years agobufferpool: allocate 1d or 2d according to caps
Rob Clark [Thu, 24 Nov 2011 19:34:06 +0000 (13:34 -0600)]
bufferpool: allocate 1d or 2d according to caps

5 years agoducatividdec: propagate pixel-aspect-ratio in caps
Luciana Fujii Pontello [Thu, 17 Nov 2011 01:55:08 +0000 (20:55 -0500)]
ducatividdec: propagate pixel-aspect-ratio in caps

Parse pixel-aspect-ratio in sink caps and set it on source caps.

5 years agogst-indent gstducatividdec.c
Alessandro Decina [Wed, 16 Nov 2011 16:43:59 +0000 (17:43 +0100)]
gst-indent gstducatividdec.c

5 years agoducatividdec: don't log an ERROR when pad_alloc returns wrong-state
Alessandro Decina [Wed, 16 Nov 2011 16:41:31 +0000 (17:41 +0100)]
ducatividdec: don't log an ERROR when pad_alloc returns wrong-state

...as it's a pretty common condition while seeking or shutting down the pipeline

5 years agoducatividdec: don't error out on non fatal errors
Alessandro Decina [Tue, 15 Nov 2011 15:55:05 +0000 (16:55 +0100)]
ducatividdec: don't error out on non fatal errors

If VIDDEC3_process returns a non fatal error, log it but don't return FLOW_ERROR
upstream.

5 years agoducatividdec: add support for video/x-raw-yuv-strided
Alessandro Decina [Mon, 14 Nov 2011 16:08:00 +0000 (17:08 +0100)]
ducatividdec: add support for video/x-raw-yuv-strided

Add support video/x-raw-yuv-strided and rework negotiation a bit so that we can
support upstream negotiation for the rowstride field (v4l2sink likes to do
that).

5 years agoducatividenc: fix typo basevideocoder => basevideoencoder
Alessandro Decina [Thu, 3 Nov 2011 11:41:23 +0000 (12:41 +0100)]
ducatividenc: fix typo basevideocoder => basevideoencoder

5 years agoducatividenc: fix outbuf size calculation
Alessandro Decina [Mon, 31 Oct 2011 15:05:09 +0000 (16:05 +0100)]
ducatividenc: fix outbuf size calculation

5 years agoducatividenc: fix crash going PLAYING->READY->PLAYING
Alessandro Decina [Mon, 31 Oct 2011 14:13:44 +0000 (15:13 +0100)]
ducatividenc: fix crash going PLAYING->READY->PLAYING

6 years agoducatimpeg4enc: use less custom parameters
Alessandro Decina [Fri, 21 Oct 2011 10:32:54 +0000 (12:32 +0200)]
ducatimpeg4enc: use less custom parameters

6 years agoducatimpeg4enc: add mpeg4 encoder
Alessandro Decina [Fri, 21 Oct 2011 08:36:52 +0000 (10:36 +0200)]
ducatimpeg4enc: add mpeg4 encoder

6 years agoFactor out a base encoder class from ducatih264enc
Alessandro Decina [Thu, 20 Oct 2011 13:59:56 +0000 (15:59 +0200)]
Factor out a base encoder class from ducatih264enc

6 years agoducatih264enc: use base VIDENC2 structures where possible
Alessandro Decina [Wed, 19 Oct 2011 14:16:06 +0000 (16:16 +0200)]
ducatih264enc: use base VIDENC2 structures where possible

6 years agoducatih264enc: add h264 encoder
Alessandro Decina [Fri, 14 Oct 2011 16:14:00 +0000 (18:14 +0200)]
ducatih264enc: add h264 encoder

6 years agoviddec: workaround for the vc1 codec
Alessandro Decina [Thu, 29 Sep 2011 21:44:28 +0000 (23:44 +0200)]
viddec: workaround for the vc1 codec

Set memType=XDM_MEMTYPE_RAW on outbuf descs when the underlying memory is
actually of type XDM_MEMTYPE_TILEDPAGE. This should be reverted once the codec
is fixed.

6 years agoUndo a change in gst_ducati_get_mem_type.
Alessandro Decina [Thu, 29 Sep 2011 21:40:47 +0000 (23:40 +0200)]
Undo a change in gst_ducati_get_mem_type.

Make it return -1 when the virtual address doesn't match any of the known zones.

6 years agoviddec: set the size in bytes OR pixels for outbufs, not both
Alessandro Decina [Thu, 29 Sep 2011 20:53:03 +0000 (22:53 +0200)]
viddec: set the size in bytes OR pixels for outbufs, not both

6 years agoviddec: more fixes to support 1d buffers
Alessandro Decina [Thu, 29 Sep 2011 20:47:35 +0000 (22:47 +0200)]
viddec: more fixes to support 1d buffers

Don't detect TILEDPAGE memory as RAW. Set outbuf descs sizes in bytes when the
underlying memory is RAW or TILEDPAGE.

6 years agoviddec: use 1d buffers
Alessandro Decina [Thu, 29 Sep 2011 17:11:06 +0000 (19:11 +0200)]
viddec: use 1d buffers

Expect and allocate 1d buffers. Support for 2d buffers will be added again at
some point in the near future.

6 years agoducativc1dec: set frameLayerDataPresentFlag to FALSE
Alessandro Decina [Thu, 29 Sep 2011 16:50:38 +0000 (18:50 +0200)]
ducativc1dec: set frameLayerDataPresentFlag to FALSE

6 years agoducativc1dec: add debugging code to optionally forge frame headers
Alessandro Decina [Mon, 26 Sep 2011 16:08:31 +0000 (18:08 +0200)]
ducativc1dec: add debugging code to optionally forge frame headers

Add code to inject frame headers if params->frameLayerDataPresentFlag is set to
TRUE. Disabled by default, useful for debugging.

6 years agoducativc1dec: try hard to produce a valid sequence layer header
Alessandro Decina [Mon, 26 Sep 2011 16:01:04 +0000 (18:01 +0200)]
ducativc1dec: try hard to produce a valid sequence layer header

Make sure the codec data in the sequence layer is always 4 bytes as per STRUCT_C
definition. Some files seem to come with larger, 0 padded, codec_data buffers...

Ensure that the Reserved6 bit of STRUCT_C is always set to 1 as the codec seems
to be strict about this and some files seem to have it set to 0...

6 years agoducativc1dec: change param settings
Alessandro Decina [Mon, 26 Sep 2011 15:58:29 +0000 (17:58 +0200)]
ducativc1dec: change param settings

Set lateAcquireArg to -1 so that the codec starts at all.. enable error
concealment and configure codec to not expect frame headers.

6 years agoviddec: forward queries that we don't handle directly
Alessandro Decina [Wed, 31 Aug 2011 09:04:26 +0000 (11:04 +0200)]
viddec: forward queries that we don't handle directly

6 years agoducativp6dec: set payloadHeaderPresent to TRUE
Alessandro Decina [Tue, 30 Aug 2011 13:34:41 +0000 (15:34 +0200)]
ducativp6dec: set payloadHeaderPresent to TRUE

6 years agoducativp7dec: set payloadHeaderPresent to TRUE
Alessandro Decina [Tue, 30 Aug 2011 13:34:07 +0000 (15:34 +0200)]
ducativp7dec: set payloadHeaderPresent to TRUE

6 years agoducatimpeg4dec: use base VIDDEC3_* structs in allocate_params
Alessandro Decina [Fri, 26 Aug 2011 12:01:31 +0000 (14:01 +0200)]
ducatimpeg4dec: use base VIDDEC3_* structs in allocate_params

6 years agoducatimpeg2dec: use base VIDDEC3* structs in allocate_params
Alessandro Decina [Fri, 26 Aug 2011 11:47:51 +0000 (13:47 +0200)]
ducatimpeg2dec: use base VIDDEC3* structs in allocate_params

6 years agoducatih264dec: add stream-format and align in caps
Alessandro Decina [Fri, 26 Aug 2011 11:45:37 +0000 (13:45 +0200)]
ducatih264dec: add stream-format and align in caps

6 years agoconfigure: depend on -base for gstvideo
Alessandro Decina [Fri, 26 Aug 2011 09:27:27 +0000 (11:27 +0200)]
configure: depend on -base for gstvideo

6 years agoconfigure: libmemmgr was renamed to libtimemmgr
Alessandro Decina [Fri, 26 Aug 2011 07:52:54 +0000 (09:52 +0200)]
configure: libmemmgr was renamed to libtimemmgr

6 years agoh264dec: enable temporal direct predict mode
Rob Clark [Wed, 26 Jan 2011 17:43:39 +0000 (11:43 -0600)]
h264dec: enable temporal direct predict mode

With out this feature enabled, B-frames that are encoded using this prediction
mode will generate errors.

6 years agoImproved compiler warning flags strictness
Rob Clark [Wed, 26 Jan 2011 16:40:08 +0000 (10:40 -0600)]
Improved compiler warning flags strictness

6 years agoviddec: fix typo
Rob Clark [Wed, 12 Jan 2011 03:20:51 +0000 (21:20 -0600)]
viddec: fix typo

6 years agorvdec: fix compile errors
Rob Clark [Mon, 10 Jan 2011 17:38:15 +0000 (11:38 -0600)]
rvdec: fix compile errors

gstducatirvdec.c: In function 'gst_ducati_rvdec_allocate_params': error: unused variable 'params
gstducatirvdec.c: In function 'gst_ducati_rvdec_allocate_params': error: unused variable 'params

6 years agoviddec: add "version" property
Rob Clark [Sat, 1 Jan 2011 03:59:49 +0000 (21:59 -0600)]
viddec: add "version" property

Use XDM_GETVERSION to get codec version, available as a read-only property.

6 years agofix compile error with latest gst headers
Rob Clark [Sat, 1 Jan 2011 03:56:48 +0000 (21:56 -0600)]
fix compile error with latest gst headers

A conflict between XDM and gst headers results in:

  cc1: warnings being treated as errors
  In file included from gstducati.h:32,
                   from gstducatividdec.h:25,
                   from gstducativp7dec.h:25,
                   from gstducativp7dec.c:40:
  /usr/include/dce/xdc/std.h:172: error: "restrict" redefined
  /usr/include/gstreamer-0.10/gst/gstmacros.h:44: note: this is the location of the previous definition

The solution is to ensure XDM headers are included first.

6 years agoviddec: fix for width/height that is not a multiple of 16
Rob Clark [Thu, 16 Dec 2010 01:03:09 +0000 (19:03 -0600)]
viddec: fix for width/height that is not a multiple of 16

This fixes an issue with latest h264dec codec drop.

6 years agoviddec: some cleanups
Rob Clark [Thu, 16 Dec 2010 01:02:52 +0000 (19:02 -0600)]
viddec: some cleanups