mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-12-19 06:46:38 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/gst/videoconvert/gstvideoconvertorc.orc

730 lines
10 KiB
Text
Raw Normal View History

[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_memcpy_2d
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 d1 guint8
.source 1 s1 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
copyb d1, s1
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_I420_UYVY
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 d1 guint8
.dest 4 d2 guint8
.source 2 y1 guint8
.source 2 y2 guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
mergebw uv, u, v
x2 mergebw d1, uv, y1
x2 mergebw d2, uv, y2
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_I420_YUY2
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 d1 guint8
.dest 4 d2 guint8
.source 2 y1 guint8
.source 2 y2 guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
mergebw uv, u, v
x2 mergebw d1, y1, uv
x2 mergebw d2, y2, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_I420_AYUV
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 d1 guint8
.dest 4 d2 guint8
.source 1 y1 guint8
.source 1 y2 guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 1 c255 255
.temp 2 uv
.temp 2 ay
.temp 1 tu
.temp 1 tv
loadupdb tu, u
loadupdb tv, v
mergebw uv, tu, tv
mergebw ay, c255, y1
mergewl d1, ay, uv
mergebw ay, c255, y2
mergewl d2, ay, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_YUY2_I420
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y1 guint8
.dest 2 y2 guint8
.dest 1 u guint8
.dest 1 v guint8
.source 4 yuv1 guint8
.source 4 yuv2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 t1
.temp 2 t2
.temp 2 ty
x2 splitwb t1, ty, yuv1
storew y1, ty
x2 splitwb t2, ty, yuv2
storew y2, ty
x2 avgub t1, t1, t2
splitwb v, u, t1
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_UYVY_YUY2
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 yuy2 guint8
.source 4 uyvy guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
x2 swapw yuy2, uyvy
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_420_422
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 d1 guint8
.dest 1 d2 guint8
.source 1 s guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
copyb d1, s
copyb d2, s
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_420_444
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 d1 guint8
.dest 2 d2 guint8
.source 1 s guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 t
splatbw t, s
storew d1, t
storew d2, t
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_422_444
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 d1 guint8
.source 1 s guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 t
splatbw t, s
storew d1, t
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_444_422
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 d guint8
.source 2 s guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 1 t1
.temp 1 t2
splitwb t1, t2, s
avgub d, t1, t2
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_444_420
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 d guint8
.source 2 s1 guint8
.source 2 s2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 t
.temp 1 t1
.temp 1 t2
x2 avgub t, s1, s2
splitwb t1, t2, t
avgub d, t1, t2
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_planar_chroma_422_420
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 d guint8
.source 1 s1 guint8
.source 1 s2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
avgub d, s1, s2
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_YUY2_AYUV
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 8 ayuv guint8
.source 4 yuy2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 2 c255 0xff
.temp 2 yy
.temp 2 uv
.temp 4 ayay
.temp 4 uvuv
x2 splitwb uv, yy, yuy2
x2 mergebw ayay, c255, yy
mergewl uvuv, uv, uv
x2 mergewl ayuv, ayay, uvuv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_UYVY_AYUV
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 8 ayuv guint8
.source 4 uyvy guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 2 c255 0xff
.temp 2 yy
.temp 2 uv
.temp 4 ayay
.temp 4 uvuv
x2 splitwb yy, uv, uyvy
x2 mergebw ayay, c255, yy
mergewl uvuv, uv, uv
x2 mergewl ayuv, ayay, uvuv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_YUY2_Y42B
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y guint8
.dest 1 u guint8
.dest 1 v guint8
.source 4 yuy2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
x2 splitwb uv, y, yuy2
splitwb v, u, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_UYVY_Y42B
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y guint8
.dest 1 u guint8
.dest 1 v guint8
.source 4 uyvy guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
x2 splitwb y, uv, uyvy
splitwb v, u, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_YUY2_Y444
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y guint8
.dest 2 uu guint8
.dest 2 vv guint8
.source 4 yuy2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
.temp 1 u
.temp 1 v
x2 splitwb uv, y, yuy2
splitwb v, u, uv
splatbw uu, u
splatbw vv, v
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_UYVY_Y444
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y guint8
.dest 2 uu guint8
.dest 2 vv guint8
.source 4 uyvy guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
.temp 1 u
.temp 1 v
x2 splitwb y, uv, uyvy
splitwb v, u, uv
splatbw uu, u
splatbw vv, v
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_UYVY_I420
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y1 guint8
.dest 2 y2 guint8
.dest 1 u guint8
.dest 1 v guint8
.source 4 yuv1 guint8
.source 4 yuv2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 t1
.temp 2 t2
.temp 2 ty
x2 splitwb ty, t1, yuv1
storew y1, ty
x2 splitwb ty, t2, yuv2
storew y2, ty
x2 avgub t1, t1, t2
splitwb v, u, t1
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_I420
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y1 guint8
.dest 2 y2 guint8
.dest 1 u guint8
.dest 1 v guint8
.source 8 ayuv1 guint8
.source 8 ayuv2 guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 4 ay
.temp 4 uv1
.temp 4 uv2
.temp 4 uv
.temp 2 uu
.temp 2 vv
.temp 1 t1
.temp 1 t2
x2 splitlw uv1, ay, ayuv1
x2 select1wb y1, ay
x2 splitlw uv2, ay, ayuv2
x2 select1wb y2, ay
x4 avgub uv, uv1, uv2
x2 splitwb vv, uu, uv
splitwb t1, t2, uu
avgub u, t1, t2
splitwb t1, t2, vv
avgub v, t1, t2
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_YUY2
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 yuy2 guint8
.source 8 ayuv guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 yy
.temp 2 uv1
.temp 2 uv2
.temp 4 ayay
.temp 4 uvuv
x2 splitlw uvuv, ayay, ayuv
splitlw uv1, uv2, uvuv
x2 avgub uv1, uv1, uv2
x2 select1wb yy, ayay
x2 mergebw yuy2, yy, uv1
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_UYVY
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 yuy2 guint8
.source 8 ayuv guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 yy
.temp 2 uv1
.temp 2 uv2
.temp 4 ayay
.temp 4 uvuv
x2 splitlw uvuv, ayay, ayuv
splitlw uv1, uv2, uvuv
x2 avgub uv1, uv1, uv2
x2 select1wb yy, ayay
x2 mergebw yuy2, uv1, yy
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_Y42B
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 2 y guint8
.dest 1 u guint8
.dest 1 v guint8
.source 8 ayuv guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 4 ayay
.temp 4 uvuv
.temp 2 uv1
.temp 2 uv2
x2 splitlw uvuv, ayay, ayuv
splitlw uv1, uv2, uvuv
x2 avgub uv1, uv1, uv2
splitwb v, u, uv1
x2 select1wb y, ayay
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_Y444
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 1 y guint8
.dest 1 u guint8
.dest 1 v guint8
.source 4 ayuv guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 ay
.temp 2 uv
splitlw uv, ay, ayuv
splitwb v, u, uv
select1wb y, ay
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y42B_YUY2
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 yuy2 guint8
.source 2 y guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
mergebw uv, u, v
x2 mergebw yuy2, y, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y42B_UYVY
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 uyvy guint8
.source 2 y guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
mergebw uv, u, v
x2 mergebw uyvy, uv, y
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y42B_AYUV
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 8 ayuv guint8
.source 2 yy guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 1 c255 255
.temp 2 uv
.temp 2 ay
.temp 4 uvuv
.temp 4 ayay
mergebw uv, u, v
x2 mergebw ayay, c255, yy
mergewl uvuv, uv, uv
x2 mergewl ayuv, ayay, uvuv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y444_YUY2
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 yuy2 guint8
.source 2 y guint8
.source 2 u guint8
.source 2 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
.temp 4 uvuv
.temp 2 uv1
.temp 2 uv2
x2 mergebw uvuv, u, v
splitlw uv1, uv2, uvuv
x2 avgub uv, uv1, uv2
x2 mergebw yuy2, y, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y444_UYVY
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 uyvy guint8
.source 2 y guint8
.source 2 u guint8
.source 2 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 uv
.temp 4 uvuv
.temp 2 uv1
.temp 2 uv2
x2 mergebw uvuv, u, v
splitlw uv1, uv2, uvuv
x2 avgub uv, uv1, uv2
x2 mergebw uyvy, uv, y
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_Y444_AYUV
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 ayuv guint8
.source 1 yy guint8
.source 1 u guint8
.source 1 v guint8
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 1 c255 255
.temp 2 uv
.temp 2 ay
mergebw uv, u, v
mergebw ay, c255, yy
mergewl ayuv, ay, uv
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_ARGB
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 argb guint8
.source 4 ayuv guint8
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 1 a
.temp 1 y
.temp 1 u
.temp 1 v
.temp 2 wy
.temp 2 wu
.temp 2 wv
.temp 2 wr
.temp 2 wg
.temp 2 wb
.temp 1 r
.temp 1 g
.temp 1 b
.temp 4 x
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.const 1 c128 128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
x4 subb x, ayuv, c128
splitlw wv, wy, x
splitwb y, a, wy
splitwb v, u, wv
splatbw wy, y
splatbw wu, u
splatbw wv, v
mulhsw wy, wy, p1
mulhsw wr, wv, p2
addssw wr, wy, wr
mulhsw wb, wu, p3
addssw wb, wy, wb
mulhsw wg, wu, p4
addssw wg, wy, wg
mulhsw wy, wv, p5
addssw wg, wg, wy
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
convssswb r, wr
convssswb g, wg
convssswb b, wb
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
mergebw wr, a, r
mergebw wb, g, b
mergewl x, wr, wb
x4 addb argb, x, c128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_BGRA
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.dest 4 bgra guint8
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.source 4 ayuv guint8
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 1 a
.temp 1 y
.temp 1 u
.temp 1 v
.temp 2 wy
.temp 2 wu
.temp 2 wv
.temp 2 wr
.temp 2 wg
.temp 2 wb
.temp 1 r
.temp 1 g
.temp 1 b
.temp 4 x
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.const 1 c128 128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
x4 subb x, ayuv, c128
splitlw wv, wy, x
splitwb y, a, wy
splitwb v, u, wv
splatbw wy, y
splatbw wu, u
splatbw wv, v
mulhsw wy, wy, p1
mulhsw wr, wv, p2
addssw wr, wy, wr
mulhsw wb, wu, p3
addssw wb, wy, wb
mulhsw wg, wu, p4
addssw wg, wy, wg
mulhsw wy, wv, p5
addssw wg, wg, wy
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
convssswb r, wr
convssswb g, wg
convssswb b, wb
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
mergebw wb, b, g
mergebw wr, r, a
mergewl x, wb, wr
x4 addb bgra, x, c128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_ABGR
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 argb guint8
.source 4 ayuv guint8
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 1 a
.temp 1 y
.temp 1 u
.temp 1 v
.temp 2 wy
.temp 2 wu
.temp 2 wv
.temp 2 wr
.temp 2 wg
.temp 2 wb
.temp 1 r
.temp 1 g
.temp 1 b
.temp 4 x
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.const 1 c128 128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
x4 subb x, ayuv, c128
splitlw wv, wy, x
splitwb y, a, wy
splitwb v, u, wv
splatbw wy, y
splatbw wu, u
splatbw wv, v
mulhsw wy, wy, p1
mulhsw wr, wv, p2
addssw wr, wy, wr
mulhsw wb, wu, p3
addssw wb, wy, wb
mulhsw wg, wu, p4
addssw wg, wy, wg
mulhsw wy, wv, p5
addssw wg, wg, wy
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
convssswb r, wr
convssswb g, wg
convssswb b, wb
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
mergebw wb, a, b
mergebw wr, g, r
mergewl x, wb, wr
x4 addb argb, x, c128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_AYUV_RGBA
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.flags 2d
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 argb guint8
.source 4 ayuv guint8
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 1 a
.temp 1 y
.temp 1 u
.temp 1 v
.temp 2 wy
.temp 2 wu
.temp 2 wv
.temp 2 wr
.temp 2 wg
.temp 2 wb
.temp 1 r
.temp 1 g
.temp 1 b
.temp 4 x
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.const 1 c128 128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
x4 subb x, ayuv, c128
splitlw wv, wy, x
splitwb y, a, wy
splitwb v, u, wv
splatbw wy, y
splatbw wu, u
splatbw wv, v
mulhsw wy, wy, p1
mulhsw wr, wv, p2
addssw wr, wy, wr
mulhsw wb, wu, p3
addssw wb, wy, wb
mulhsw wg, wu, p4
addssw wg, wy, wg
mulhsw wy, wv, p5
addssw wg, wg, wy
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
convssswb r, wr
convssswb g, wg
convssswb b, wb
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
mergebw wr, r, g
mergebw wb, b, a
mergewl x, wr, wb
x4 addb argb, x, c128
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
videoconvert: prefix orc functions with video_convert_orc_
2012-07-23 15:19:56 +00:00
.function video_convert_orc_convert_I420_BGRA
[MOVED FROM BAD 32/68] colorspace: Add conversion code Work in progress. Colorspace handles most format conversion using 3-stage getline/matrix/putline process using an AYUV or ARGB intermediate, with most functions handled by Orc. There is also a table of single-pass conversions, all handled by Orc. The plan is to add optional stages for various chroma upsampling/downsampling algorithms, dithering, and float/int16 intermediates, and then have Orc create multi-stage functions at runtime.
2010-09-14 01:49:43 +00:00
.dest 4 argb guint8
.source 1 y guint8
.source 1 u guint8
.source 1 v guint8
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.temp 2 wy
.temp 2 wu
.temp 2 wv
.temp 2 wr
.temp 2 wg
.temp 2 wb
.temp 1 r
.temp 1 g
.temp 1 b
videoconvert: fix I420 to BGRA fast-path some more Calculate alpha value differently so that we can avoid running out of registers.
2014-01-09 17:08:41 +00:00
.temp 4 x
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
.const 1 c128 128
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
subb r, y, c128
splatbw wy, r
videoconvert: don't interpolate chroma in I420 -> RGB Don't try to interpolate the chroma samples, the used algorithm only works for horizontal cositing. Let's switch to a faster and safer version until we handle chroma siting correctly in the fastpaths.
2014-01-13 15:20:55 +00:00
loadupdb r, u
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
subb r, r, c128
splatbw wu, r
videoconvert: don't interpolate chroma in I420 -> RGB Don't try to interpolate the chroma samples, the used algorithm only works for horizontal cositing. Let's switch to a faster and safer version until we handle chroma siting correctly in the fastpaths.
2014-01-13 15:20:55 +00:00
loadupdb r, v
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
subb r, r, c128
splatbw wv, r
mulhsw wy, wy, p1
mulhsw wr, wv, p2
addssw wr, wy, wr
mulhsw wb, wu, p3
addssw wb, wy, wb
mulhsw wg, wu, p4
addssw wg, wy, wg
mulhsw wy, wv, p5
addssw wg, wg, wy
[MOVED FROM BAD 31/68] colorspace: Revive element Now based on Orc.
2010-09-13 19:48:50 +00:00
convssswb r, wr
convssswb g, wg
convssswb b, wb
videoconvert: rework YUV->RGB fastpaths Rework the orc code to be around 10% faster and support arbitrary matrices. Pass the matrix parameters to the YUV->RGB functions to make them work for all matrices. This enables more and faster fastpath conversions. See https://bugzilla.gnome.org/show_bug.cgi?id=721701
2014-01-09 17:12:00 +00:00
mergebw wb, b, g
mergebw wr, r, 127
mergewl x, wb, wr
videoconvert: fix I420 to BGRA fast-path some more Calculate alpha value differently so that we can avoid running out of registers.
2014-01-09 17:08:41 +00:00
x4 addb argb, x, c128
Reference in a new issue Copy permalink