mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-03 16:09:39 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/ext/cog/cog.orc

563 lines
7.1 KiB
Text
Raw Normal View History

cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
cog: touch cog.orc to force regeneration of cogorc.[ch] .. with new Makefile.am rules, which should fix the build without the need to do a 'manual' make clean after updating.
2010-03-23 09:14:03 +00:00
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.function cogorc_downsample_horiz_cosite_1tap
.dest 1 d1
.source 2 s1
select0wb d1, s1
.function cogorc_downsample_horiz_cosite_3tap
.dest 1 d1
.source 2 s1
.source 2 s2
.temp 1 t1
.temp 1 t2
.temp 1 t3
.temp 2 t4
.temp 2 t5
.temp 2 t6
copyw t4, s1
select0wb t1, t4
select1wb t2, t4
select0wb t3, s2
convubw t4, t1
convubw t5, t2
convubw t6, t3
mullw t5, t5, 2
addw t4, t4, t6
addw t4, t4, t5
addw t4, t4, 2
shrsw t4, t4, 2
convsuswb d1, t4
cog: implement chroma-site Determine chroma site position from the caps and use it when upsampling/downsampling chroma.
2010-03-15 08:26:25 +00:00
.function cogorc_downsample_420_jpeg
.dest 1 d1
.source 2 s1
.source 2 s2
.temp 2 t1
.temp 1 t2
.temp 1 t3
.temp 1 t4
.temp 1 t5
copyw t1, s1
select0wb t2, t1
select1wb t3, t1
avgub t2, t2, t3
copyw t1, s2
select0wb t4, t1
select1wb t5, t1
avgub t4, t4, t5
avgub d1, t2, t4
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.function cogorc_downsample_vert_halfsite_2tap
.dest 1 d1
.source 1 s1
.source 1 s2
avgub d1, s1, s2
.function cogorc_downsample_vert_cosite_3tap
.dest 1 d1
.source 1 s1
.source 1 s2
.source 1 s3
.temp 2 t1
.temp 2 t2
.temp 2 t3
convubw t1, s1
convubw t2, s2
convubw t3, s3
mullw t2, t2, 2
addw t1, t1, t3
addw t1, t1, t2
addw t1, t1, 2
shrsw t1, t1, 2
convsuswb d1, t1
.function cogorc_downsample_vert_halfsite_4tap
.dest 1 d1
.source 1 s1
.source 1 s2
.source 1 s3
.source 1 s4
.temp 2 t1
.temp 2 t2
.temp 2 t3
.temp 2 t4
convubw t1, s1
convubw t2, s2
convubw t3, s3
convubw t4, s4
addw t2, t2, t3
mullw t2, t2, 26
addw t1, t1, t4
mullw t1, t1, 6
addw t2, t2, t1
addw t2, t2, 32
shrsw t2, t2, 6
convsuswb d1, t2
.function cogorc_upsample_horiz_cosite_1tap
.dest 2 d1 uint8_t
.source 1 s1
.temp 1 t1
copyb t1, s1
mergebw d1, t1, t1
.function cogorc_upsample_horiz_cosite
.dest 2 d1 uint8_t
.source 1 s1
.source 1 s2
.temp 1 t1
.temp 1 t2
copyb t1, s1
avgub t2, t1, s2
mergebw d1, t1, t2
.function cogorc_upsample_vert_avgub
.dest 1 d1
.source 1 s1
.source 1 s2
avgub d1, s1, s2
.function orc_unpack_yuyv_y
.dest 1 d1
.source 2 s1
select0wb d1, s1
.function orc_unpack_yuyv_u
.dest 1 d1
.source 4 s1
.temp 2 t1
select0lw t1, s1
select1wb d1, t1
.function orc_unpack_yuyv_v
.dest 1 d1
.source 4 s1
.temp 2 t1
select1lw t1, s1
select1wb d1, t1
.function orc_pack_yuyv
.dest 4 d1
.source 2 s1 uint8_t
.source 1 s2
.source 1 s3
.temp 1 t1
.temp 1 t2
.temp 2 t3
.temp 2 t4
.temp 2 t5
copyw t5, s1
select0wb t1, t5
select1wb t2, t5
mergebw t3, t1, s2
mergebw t4, t2, s3
mergewl d1, t3, t4
.function orc_unpack_uyvy_y
.dest 1 d1
.source 2 s1
select1wb d1, s1
.function orc_unpack_uyvy_u
.dest 1 d1
.source 4 s1
.temp 2 t1
select0lw t1, s1
select0wb d1, t1
.function orc_unpack_uyvy_v
.dest 1 d1
.source 4 s1
.temp 2 t1
select1lw t1, s1
select0wb d1, t1
.function orc_pack_uyvy
.dest 4 d1
.source 2 s1 uint8_t
.source 1 s2
.source 1 s3
.temp 1 t1
.temp 1 t2
.temp 2 t3
.temp 2 t4
.temp 2 t5
copyw t5, s1
select0wb t1, t5
select1wb t2, t5
mergebw t3, s2, t1
mergebw t4, s3, t2
mergewl d1, t3, t4
.function orc_addc_convert_u8_s16
.dest 1 d1
.source 2 s1 int16_t
.temp 2 t1
addw t1, s1, 128
convsuswb d1, t1
.function orc_subc_convert_s16_u8
.dest 2 d1 int16_t
.source 1 s1
.temp 2 t1
convubw t1, s1
subw d1, t1, 128
.function orc_splat_u8_ns
.dest 1 d1
.param 1 p1
copyb d1, p1
.function orc_splat_s16_ns
.dest 2 d1 int16_t
.param 2 p1
copyw d1, p1
.function orc_matrix2_u8
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.param 2 p1
.param 2 p2
.param 2 p3
.temp 2 t1
.temp 2 t2
convubw t1, s1
mullw t1, t1, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
addw t1, t1, p3
shrsw t1, t1, 6
convsuswb d1, t1
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.function orc_matrix2_11_u8
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.param 2 p1
.param 2 p2
.temp 2 t1
.temp 2 t2
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.temp 2 t3
.temp 2 t4
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
convubw t1, s1
subw t1, t1, 16
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
mullw t3, t1, p1
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
convubw t2, s2
subw t2, t2, 128
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
mullw t4, t2, p2
addw t3, t3, t4
cog: convert some params to consts in orc code
2009-11-20 19:55:15 +00:00
addw t3, t3, 128
shrsw t3, t3, 8
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
addw t3, t3, t1
addw t3, t3, t2
convsuswb d1, t3
.function orc_matrix2_12_u8
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.param 2 p1
.param 2 p2
.temp 2 t1
.temp 2 t2
.temp 2 t3
.temp 2 t4
convubw t1, s1
subw t1, t1, 16
mullw t3, t1, p1
convubw t2, s2
subw t2, t2, 128
mullw t4, t2, p2
addw t3, t3, t4
cog: convert some params to consts in orc code
2009-11-20 19:55:15 +00:00
addw t3, t3, 128
shrsw t3, t3, 8
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
addw t3, t3, t1
addw t3, t3, t2
addw t3, t3, t2
convsuswb d1, t3
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.function orc_matrix3_u8
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.source 1 s3 uint8_t
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.temp 2 t1
.temp 2 t2
convubw t1, s1
mullw t1, t1, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
convubw t2, s3
mullw t2, t2, p3
addw t1, t1, t2
addw t1, t1, p4
shrsw t1, t1, 6
convsuswb d1, t1
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.function orc_matrix3_100_u8
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.source 1 s3 uint8_t
.param 2 p1
.param 2 p2
.param 2 p3
.temp 2 t1
.temp 2 t2
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.temp 2 t3
#.temp 2 t4
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
convubw t1, s1
subw t1, t1, 16
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
mullw t3, t1, p1
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
convubw t2, s2
subw t2, t2, 128
mullw t2, t2, p2
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
addw t3, t3, t2
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
convubw t2, s3
subw t2, t2, 128
mullw t2, t2, p3
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
addw t3, t3, t2
cog: convert some params to consts in orc code
2009-11-20 19:55:15 +00:00
addw t3, t3, 128
shrsw t3, t3, 8
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
addw t3, t3, t1
convsuswb d1, t3
cogcolorspace: implement color-matrix handling
2010-03-15 01:56:06 +00:00
.function orc_matrix3_100_offset_u8
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.source 1 s3 uint8_t
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
#.param 2 p6
.temp 2 t1
.temp 2 t2
.temp 2 t3
#.temp 2 t3
#.temp 2 t4
convubw t3, s1
mullw t1, t3, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
convubw t2, s3
mullw t2, t2, p3
addw t1, t1, t2
addw t1, t1, p4
shrsw t1, t1, p5
#addw t1, t1, p6
addw t1, t1, t3
convsuswb d1, t1
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.function orc_matrix3_000_u8
.dest 1 d1 uint8_t
.source 1 s1 uint8_t
.source 1 s2 uint8_t
.source 1 s3 uint8_t
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.param 2 p5
cog: Add color matrix to RGB->YCbCr conversion
2009-10-08 22:18:04 +00:00
#.param 2 p6
cog: Improve RGB<->YCbCr matrixing YCbCr->RGB now has a 'bits' setting, which controls the number of bits used for the matrix coefficients. Choices are 6 or 8. 6 is low quality, but faster. RGB->YCbCr now works. Always works in 8 bit, since it's the same speed as 6 bit.
2009-10-08 20:24:36 +00:00
.temp 2 t1
.temp 2 t2
#.temp 2 t3
#.temp 2 t4
convubw t1, s1
mullw t1, t1, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
convubw t2, s3
mullw t2, t2, p3
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
addw t1, t1, t2
addw t1, t1, p4
shrsw t1, t1, p5
cog: Add color matrix to RGB->YCbCr conversion
2009-10-08 22:18:04 +00:00
#addw t1, t1, p6
convwb d1, t1
cog: hacking to improve matrix quality
2009-10-07 23:30:26 +00:00
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.function orc_pack_123x
.dest 4 d1 uint32_t
.source 1 s1
.source 1 s2
.source 1 s3
.param 1 p1
.temp 2 t1
.temp 2 t2
mergebw t1, s1, s2
mergebw t2, s3, p1
mergewl d1, t1, t2
.function orc_pack_x123
.dest 4 d1 uint32_t
.source 1 s1
.source 1 s2
.source 1 s3
.param 1 p1
.temp 2 t1
.temp 2 t2
mergebw t1, p1, s1
mergebw t2, s2, s3
mergewl d1, t1, t2
cog: (re-)implement 1/2/4-tap scaling
2009-09-21 16:35:30 +00:00
.function cogorc_combine2_u8
.dest 1 d1
.source 1 s1
.source 1 s2
.param 2 p1
.param 2 p2
.temp 2 t1
.temp 2 t2
convubw t1, s1
mullw t1, t1, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
shruw t1, t1, 8
convuuswb d1, t1
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.function cogorc_combine4_u8
.dest 1 d1
.source 1 s1
.source 1 s2
.source 1 s3
.source 1 s4
.param 2 p1
.param 2 p2
.param 2 p3
.param 2 p4
.temp 2 t1
.temp 2 t2
convubw t1, s1
mullw t1, t1, p1
convubw t2, s2
mullw t2, t2, p2
addw t1, t1, t2
convubw t2, s3
mullw t2, t2, p3
addw t1, t1, t2
convubw t2, s4
mullw t2, t2, p4
addw t1, t1, t2
addw t1, t1, 32
shrsw t1, t1, 6
convsuswb d1, t1
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
.function cogorc_unpack_axyz_0
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.dest 1 d1
.source 4 s1
.temp 2 t1
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
select0lw t1, s1
select0wb d1, t1
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
.function cogorc_unpack_axyz_1
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.dest 1 d1
.source 4 s1
.temp 2 t1
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
select0lw t1, s1
select1wb d1, t1
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
.function cogorc_unpack_axyz_2
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
.dest 1 d1
.source 4 s1
.temp 2 t1
cog: Fix AYUV unpacking
2009-10-08 20:23:53 +00:00
select1lw t1, s1
select0wb d1, t1
.function cogorc_unpack_axyz_3
.dest 1 d1
.source 4 s1
.temp 2 t1
select1lw t1, s1
select1wb d1, t1
cog: autogenerate code, colorspace improvements
2009-09-19 20:32:19 +00:00
Reference in a new issue Copy permalink