virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-15 14:13:16 +02:00
Code Issues Releases Activity

lavc/aarch64: clean-up sao band 8x8 function formatting

Browse Source 
Signed-off-by: J. Dekker <jdek@itanimul.li>
This commit is contained in:
J. Dekker
2021-12-15 20:06:20 +01:00
parent c97ffc1a77
commit 89a2ed4a8b
1 changed files with 25 additions and 40 deletions
Download Patch File Download Diff File Expand all files Collapse all files

65
libavcodec/aarch64/hevcdsp_sao_neon.S
View File

@@ -30,24 +30,21 @@
// int width, int height) // int width, int height)
function ff_hevc_sao_band_filter_8x8_8_neon, export=1 function ff_hevc_sao_band_filter_8x8_8_neon, export=1
sub sp, sp, #64 sub sp, sp, #64
stp xzr, xzr, [sp] stp xzr, xzr, [sp]
stp xzr, xzr, [sp, #16] stp xzr, xzr, [sp, #16]
stp xzr, xzr, [sp, #32] stp xzr, xzr, [sp, #32]
stp xzr, xzr, [sp, #48] stp xzr, xzr, [sp, #48]
mov w8, #4 mov w8, #4
0: 0: ldrsh x9, [x4, x8, lsl #1] // sao_offset_val[k+1]
ldrsh x9, [x4, x8, lsl #1] // x9 = sao_offset_val[k+1]
subs w8, w8, #1 subs w8, w8, #1
add w10, w8, w5 // x10 = k + sao_left_class add w10, w8, w5 // k + sao_left_class
and w10, w10, #0x1F and w10, w10, #0x1F
strh w9, [sp, x10, lsl #1] strh w9, [sp, x10, lsl #1]
bne 0b bne 0b
ld1 {v16.16b-v19.16b}, [sp], #64 ld1 {v16.16b-v19.16b}, [sp], #64
movi v20.8h, #1 movi v20.8h, #1
1: // beginning of line 1: mov w8, w6 // beginning of line
mov w8, w6 2: // Simple layout for accessing 16bit values
2:
// Simple layout for accessing 16bit values
// with 8bit LUT. // with 8bit LUT.
// //
// 00 01 02 03 04 05 06 07 // 00 01 02 03 04 05 06 07
@@ -55,33 +52,21 @@ function ff_hevc_sao_band_filter_8x8_8_neon, export=1
// |xDE#xAD|xCA#xFE|xBE#xEF|xFE#xED|.... // |xDE#xAD|xCA#xFE|xBE#xEF|xFE#xED|....
// +-----------------------------------> // +----------------------------------->
// i-0 i-1 i-2 i-3 // i-0 i-1 i-2 i-3
// dst[x] = av_clip_pixel(src[x] + offset_table[src[x] >> shift]); ld1 {v2.8b}, [x1] // dst[x] = av_clip_pixel(src[x] + offset_table[src[x] >> shift]);
ld1 {v2.8b}, [x1] uxtl v0.8h, v2.8b // load src[x]
// load src[x] ushr v2.8h, v0.8h, #3 // >> BIT_DEPTH - 3
uxtl v0.8h, v2.8b shl v1.8h, v2.8h, #1 // low (x2, accessing short)
// >> shift add v3.8h, v1.8h, v20.8h // +1 access upper short
ushr v2.8h, v0.8h, #3 // BIT_DEPTH - 3 sli v1.8h, v3.8h, #8 // shift insert index to upper byte
// x2 (access lower short) tbx v2.16b, {v16.16b-v19.16b}, v1.16b // table
shl v1.8h, v2.8h, #1 // low (x2, accessing short) add v1.8h, v0.8h, v2.8h // src[x] + table
// +1 access upper short sqxtun v4.8b, v1.8h // clip + narrow
add v3.8h, v1.8h, v20.8h st1 {v4.8b}, [x0] // store
// shift insert index to upper byte subs w8, w8, #8 // done 8 pixels
sli v1.8h, v3.8h, #8
// table
tbx v2.16b, {v16.16b-v19.16b}, v1.16b
// src[x] + table
add v1.8h, v0.8h, v2.8h
// clip + narrow
sqxtun v4.8b, v1.8h
// store
st1 {v4.8b}, [x0]
// done 8 pixels
subs w8, w8, #8
bne 2b bne 2b
// finished line subs w7, w7, #1 // finished line, prep. new
subs w7, w7, #1 add x0, x0, x2 // dst += stride_dst
add x0, x0, x2 // dst += stride_dst add x1, x1, x3 // src += stride_src
add x1, x1, x3 // src += stride_src
bne 1b bne 1b
ret ret
endfunc endfunc