mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-24 02:31:03 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/subprojects/gst-plugins-bad/gst/bayer/gstbayerorc.orc

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

531 lines
7.9 KiB
Text
Raw Normal View History

bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_horiz_upsample_unaligned
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 2 d0 guint8
.dest 2 d1 guint8
.source 2 s guint8
.temp 2 t
.temp 1 b
.temp 1 c
.temp 1 d
.temp 1 e
splitwb c, b, s
loadoffw t, s, 1
splitwb e, d, t
avgub e, c, e
mergebw d0, c, e
avgub b, b, d
mergebw d1, b, d
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_horiz_upsample
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 2 d0 guint8
.dest 2 d1 guint8
.source 2 s guint8
.temp 2 t
.temp 1 b
.temp 1 c
.temp 1 d
.temp 1 e
loadoffw t, s, -1
bayer: fix unused variable
2014-05-20 09:42:35 +00:00
select1wb b, t
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
splitwb d, c, s
loadoffw t, s, 1
bayer: fix unused variable
2014-05-20 09:42:35 +00:00
select0wb e, t
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
avgub e, c, e
mergebw d0, c, e
avgub b, b, d
mergebw d1, b, d
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_bg_bgra
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 g0 guint8
.source 2 r0 guint8
.source 2 b1 guint8
.source 2 g1 guint8
.source 2 g2 guint8
.source 2 r2 guint8
.temp 4 ra
.temp 4 bg
.temp 2 r
.temp 2 g
.temp 2 t
x2 avgub r, r0, r2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 255
andw t, t, 65280
orw g, t, g
x2 mergebw bg, b1, g
x2 mergebw ra, r, 255
x2 mergewl d, bg, ra
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_gr_bgra
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 b0 guint8
.source 2 g0 guint8
.source 2 g1 guint8
.source 2 r1 guint8
.source 2 b2 guint8
.source 2 g2 guint8
.temp 4 ra
.temp 4 bg
.temp 2 b
.temp 2 g
.temp 2 t
x2 avgub b, b0, b2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 65280
andw t, t, 255
orw g, t, g
x2 mergebw bg, b, g
x2 mergebw ra, r1, 255
x2 mergewl d, bg, ra
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_bg_abgr
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 g0 guint8
.source 2 r0 guint8
.source 2 b1 guint8
.source 2 g1 guint8
.source 2 g2 guint8
.source 2 r2 guint8
.temp 4 ab
.temp 4 gr
.temp 2 r
.temp 2 g
.temp 2 t
x2 avgub r, r0, r2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 255
andw t, t, 65280
orw g, t, g
x2 mergebw ab, 255, b1
x2 mergebw gr, g, r
x2 mergewl d, ab, gr
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_gr_abgr
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 b0 guint8
.source 2 g0 guint8
.source 2 g1 guint8
.source 2 r1 guint8
.source 2 b2 guint8
.source 2 g2 guint8
.temp 4 ab
.temp 4 gr
.temp 2 b
.temp 2 g
.temp 2 t
x2 avgub b, b0, b2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 65280
andw t, t, 255
orw g, t, g
x2 mergebw ab, 255, b
x2 mergebw gr, g, r1
x2 mergewl d, ab, gr
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_bg_rgba
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 g0 guint8
.source 2 r0 guint8
.source 2 b1 guint8
.source 2 g1 guint8
.source 2 g2 guint8
.source 2 r2 guint8
.temp 4 rg
.temp 4 ba
.temp 2 r
.temp 2 g
.temp 2 t
x2 avgub r, r0, r2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 255
andw t, t, 65280
orw g, t, g
x2 mergebw rg, r, g
x2 mergebw ba, b1, 255
x2 mergewl d, rg, ba
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_gr_rgba
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 b0 guint8
.source 2 g0 guint8
.source 2 g1 guint8
.source 2 r1 guint8
.source 2 b2 guint8
.source 2 g2 guint8
.temp 4 rg
.temp 4 ba
.temp 2 b
.temp 2 g
.temp 2 t
x2 avgub b, b0, b2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 65280
andw t, t, 255
orw g, t, g
x2 mergebw rg, r1, g
x2 mergebw ba, b, 255
x2 mergewl d, rg, ba
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_bg_argb
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 g0 guint8
.source 2 r0 guint8
.source 2 b1 guint8
.source 2 g1 guint8
.source 2 g2 guint8
.source 2 r2 guint8
.temp 4 ar
.temp 4 gb
.temp 2 r
.temp 2 g
.temp 2 t
x2 avgub r, r0, r2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 255
andw t, t, 65280
orw g, t, g
x2 mergebw ar, 255, r
x2 mergebw gb, g, b1
x2 mergewl d, ar, gb
bayer: prefix orc functions with bayer_orc Add bayer orc to the unit test
2012-07-24 08:32:51 +00:00
.function bayer_orc_merge_gr_argb
bayer2rgb: Convert to Orc Seriously faster. Algorithm is nearly the same as bilinear, which given the speed of this code, should be considered the baseline of quality. Speed appears to be limited by memory bandwidth, so I didn't bother trying to make it any faster.
2011-05-31 06:43:39 +00:00
.dest 8 d guint8
.source 2 b0 guint8
.source 2 g0 guint8
.source 2 g1 guint8
.source 2 r1 guint8
.source 2 b2 guint8
.source 2 g2 guint8
.temp 4 ar
.temp 4 gb
.temp 2 b
.temp 2 g
.temp 2 t
x2 avgub b, b0, b2
x2 avgub g, g0, g2
copyw t, g1
x2 avgub g, g, t
andw g, g, 65280
andw t, t, 255
orw g, t, g
x2 mergebw ar, 255, r1
x2 mergebw gb, g, b
x2 mergewl d, ar, gb
bayer2rgb: Support video/x-bayer 10/12/14/16 bit depths Add support for 10/12/14/16 bit depths . This consists of multiple parts. First is the parsing of caps, which pulls out the bitness and endianness from the video/x-bayer format. Second, gst_bayer2rgb_split_and_upsample_horiz() is split into two similar functions, one for 8bit bayer handling and another for 16bit bayer handling. The content is basically identical, except one uses 8bpp and the other 16bpp inputs and outputs, and they each use different ORC code to match. The 16bpp variant also handles endian swapping. There is now a wrapper called gst_bayer2rgb_split_and_upsample_horiz() which selects the correct function based on bpp from the parser. Third, gst_bayer2rgb_process() is extended to handle both 8bit and 16bit bayer data. Yet again there are matching ORC functions to handle the 16bit data. This time however the 16bit handling of data is slightly special. The ORC is not able to emit opcodes for 'x2 mergelq', so the trick here is to store the BG and GR longs into separate 'dtmp' temporary buffer, and then do one more ORC post-processing step, compensate for the less-than-16bpp bitness using left shift, and reorder them into the destination frame using 'mergelq' . Example usage: ``` $ gst-launch-1.0 videotestsrc ! \ video/x-bayer,width=512,height=512,format=bggr16le ! \ bayer2rgb ! \ video/x-raw,format=RGBA64_LE ! \ videoconvert ! \ autovideosink ``` Example usage: ``` $ gst-launch-1.0 videotestsrc ! \ video/x-raw,width=512,height=512,format=ARGB ! \ rgb2bayer ! \ video/x-bayer,format=bggr12le ! \ bayer2rgb ! \ video/x-raw,format=RGBA64_LE ! \ videoconvert ! \ autovideosink ``` Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4686>
2023-05-20 17:33:27 +00:00
# 10..16 bit bayer handling
.function bayer16_orc_horiz_upsample_le
.dest 4 d0 guint16
.dest 4 d1 guint16
.source 4 s guint16
.temp 4 t
.temp 2 b
.temp 2 c
.temp 2 d
.temp 2 e
splitlw c, b, s
loadoffl t, s, 1
splitlw e, d, t
avguw e, c, e
mergewl d0, c, e
avguw b, b, d
mergewl d1, b, d
.function bayer16_orc_horiz_upsample_be
.dest 4 d0 guint16
.dest 4 d1 guint16
.source 4 s guint16
.temp 4 t
.temp 2 b
.temp 2 c
.temp 2 d
.temp 2 e
splitlw c, b, s
swapw b, b
swapw c, c
loadoffl t, s, 1
splitlw e, d, t
swapw d, d
swapw e, e
avguw e, c, e
mergewl d0, c, e
avguw b, b, d
mergewl d1, b, d
.function bayer16_orc_merge_bg_bgra
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 g0 guint8
.source 4 r0 guint8
.source 4 b1 guint8
.source 4 g1 guint8
.source 4 g2 guint8
.source 4 r2 guint8
.temp 4 r
.temp 4 g
.temp 4 t
x2 avguw r, r0, r2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 65535
andl t, t, 4294901760
orl g, t, g
x2 mergewl d1, b1, g
x2 mergewl d2, r, 65535
.function bayer16_orc_merge_gr_bgra
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 b0 guint8
.source 4 g0 guint8
.source 4 g1 guint8
.source 4 r1 guint8
.source 4 b2 guint8
.source 4 g2 guint8
.temp 4 b
.temp 4 g
.temp 4 t
x2 avguw b, b0, b2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 4294901760
andl t, t, 65535
orl g, t, g
x2 mergewl d1, b, g
x2 mergewl d2, r1, 65535
.function bayer16_orc_merge_bg_abgr
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 g0 guint8
.source 4 r0 guint8
.source 4 b1 guint8
.source 4 g1 guint8
.source 4 g2 guint8
.source 4 r2 guint8
.temp 4 r
.temp 4 g
.temp 4 t
x2 avguw r, r0, r2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 65535
andl t, t, 4294901760
orl g, t, g
x2 mergewl d1, 65535, b1
x2 mergewl d2, g, r
.function bayer16_orc_merge_gr_abgr
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 b0 guint8
.source 4 g0 guint8
.source 4 g1 guint8
.source 4 r1 guint8
.source 4 b2 guint8
.source 4 g2 guint8
.temp 4 b
.temp 4 g
.temp 4 t
x2 avguw b, b0, b2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 4294901760
andl t, t, 65535
orl g, t, g
x2 mergewl d1, 65535, b
x2 mergewl d2, g, r1
.function bayer16_orc_merge_bg_rgba
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 g0 guint8
.source 4 r0 guint8
.source 4 b1 guint8
.source 4 g1 guint8
.source 4 g2 guint8
.source 4 r2 guint8
.temp 4 r
.temp 4 g
.temp 4 t
x2 avguw r, r0, r2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 65535
andl t, t, 4294901760
orl g, t, g
x2 mergewl d1, r, g
x2 mergewl d2, b1, 65535
.function bayer16_orc_merge_gr_rgba
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 b0 guint8
.source 4 g0 guint8
.source 4 g1 guint8
.source 4 r1 guint8
.source 4 b2 guint8
.source 4 g2 guint8
.temp 4 b
.temp 4 g
.temp 4 t
x2 avguw b, b0, b2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 4294901760
andl t, t, 65535
orl g, t, g
x2 mergewl d1, r1, g
x2 mergewl d2, b, 65535
.function bayer16_orc_merge_bg_argb
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 g0 guint8
.source 4 r0 guint8
.source 4 b1 guint8
.source 4 g1 guint8
.source 4 g2 guint8
.source 4 r2 guint8
.temp 4 r
.temp 4 g
.temp 4 t
x2 avguw r, r0, r2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 65535
andl t, t, 4294901760
orl g, t, g
x2 mergewl d1, 65535, r
x2 mergewl d2, g, b1
.function bayer16_orc_merge_gr_argb
.dest 8 d1 guint16
.dest 8 d2 guint16
.source 4 b0 guint8
.source 4 g0 guint8
.source 4 g1 guint8
.source 4 r1 guint8
.source 4 b2 guint8
.source 4 g2 guint8
.temp 4 b
.temp 4 g
.temp 4 t
x2 avguw b, b0, b2
x2 avguw g, g0, g2
copyl t, g1
x2 avguw g, g, t
andl g, g, 4294901760
andl t, t, 65535
orl g, t, g
x2 mergewl d1, 65535, r1
x2 mergewl d2, g, b
.function bayer16to16_orc_reorder
.dest 8 d guint8
.source 4 s1 guint32
.source 4 s2 guint32
.param 4 shift
.temp 4 u
.temp 4 v
.temp 8 q
x2 muluwl q, s1, 0xffff
x2 shrul q, q, shift
x2 convuuslw u, q
x2 muluwl q, s2, 0xffff
x2 shrul q, q, shift
x2 convuuslw v, q
mergelq d, u, v
.function bayer16to8_orc_reorder
.dest 4 d guint8
.source 4 s1 guint32
.source 4 s2 guint32
.param 4 shift
.temp 2 u
.temp 2 v
.temp 4 l
x2 shruw l, s1, shift
x2 convuuswb u, l
x2 shruw l, s2, shift
x2 convuuswb v, l
mergewl d, u, v
.function bayer8to16_orc_reorder
.dest 8 d guint8
.source 4 s guint32
x4 splatbw d, s
Reference in a new issue Copy permalink