2014-02-03 14:29:09 -08:00
|
|
|
/*
|
2014-07-10 00:56:05 +02:00
|
|
|
* This file is part of FFmpeg.
|
2014-02-03 14:29:09 -08:00
|
|
|
*
|
2014-07-10 00:56:05 +02:00
|
|
|
* FFmpeg is free software; you can redistribute it and/or
|
2014-02-03 14:29:09 -08:00
|
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
|
|
|
* License as published by the Free Software Foundation; either
|
|
|
|
|
* version 2.1 of the License, or (at your option) any later version.
|
|
|
|
|
*
|
2014-07-10 00:56:05 +02:00
|
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
2014-02-03 14:29:09 -08:00
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
|
* Lesser General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU Lesser General Public
|
2014-07-10 00:56:05 +02:00
|
|
|
* License along with FFmpeg; if not, write to the Free Software
|
2014-02-03 14:29:09 -08:00
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <stdint.h>
|
|
|
|
|
|
|
|
|
|
#include "config.h"
|
|
|
|
|
#include "libavutil/attributes.h"
|
2015-09-29 18:34:28 -07:00
|
|
|
#include "libavutil/intreadwrite.h"
|
2014-02-03 14:29:09 -08:00
|
|
|
#include "pixblockdsp.h"
|
|
|
|
|
|
2024-03-10 17:14:55 +01:00
|
|
|
static void get_pixels_16_c(int16_t *restrict block, const uint8_t *pixels,
|
2017-03-20 13:47:29 +01:00
|
|
|
ptrdiff_t stride)
|
2024-07-25 18:38:32 +03:00
|
|
|
{
|
|
|
|
|
for (int i = 0; i < 8; i++)
|
|
|
|
|
AV_COPY128(block + i * 8, pixels + i * stride);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void get_pixels_unaligned_16_c(int16_t *restrict block,
|
|
|
|
|
const uint8_t *pixels, ptrdiff_t stride)
|
2015-09-29 18:34:28 -07:00
|
|
|
{
|
2017-03-20 13:47:29 +01:00
|
|
|
AV_COPY128U(block + 0 * 8, pixels + 0 * stride);
|
|
|
|
|
AV_COPY128U(block + 1 * 8, pixels + 1 * stride);
|
|
|
|
|
AV_COPY128U(block + 2 * 8, pixels + 2 * stride);
|
|
|
|
|
AV_COPY128U(block + 3 * 8, pixels + 3 * stride);
|
|
|
|
|
AV_COPY128U(block + 4 * 8, pixels + 4 * stride);
|
|
|
|
|
AV_COPY128U(block + 5 * 8, pixels + 5 * stride);
|
|
|
|
|
AV_COPY128U(block + 6 * 8, pixels + 6 * stride);
|
|
|
|
|
AV_COPY128U(block + 7 * 8, pixels + 7 * stride);
|
2015-09-29 18:34:28 -07:00
|
|
|
}
|
|
|
|
|
|
2024-03-10 17:14:55 +01:00
|
|
|
static void get_pixels_8_c(int16_t *restrict block, const uint8_t *pixels,
|
2017-03-20 13:47:29 +01:00
|
|
|
ptrdiff_t stride)
|
2015-09-29 18:34:28 -07:00
|
|
|
{
|
|
|
|
|
int i;
|
2014-02-03 14:29:09 -08:00
|
|
|
|
2015-09-29 18:34:28 -07:00
|
|
|
/* read the pixels */
|
|
|
|
|
for (i = 0; i < 8; i++) {
|
|
|
|
|
block[0] = pixels[0];
|
|
|
|
|
block[1] = pixels[1];
|
|
|
|
|
block[2] = pixels[2];
|
|
|
|
|
block[3] = pixels[3];
|
|
|
|
|
block[4] = pixels[4];
|
|
|
|
|
block[5] = pixels[5];
|
|
|
|
|
block[6] = pixels[6];
|
|
|
|
|
block[7] = pixels[7];
|
2017-03-20 13:47:29 +01:00
|
|
|
pixels += stride;
|
2015-09-29 18:34:28 -07:00
|
|
|
block += 8;
|
|
|
|
|
}
|
|
|
|
|
}
|
2014-02-03 14:29:09 -08:00
|
|
|
|
2024-03-10 17:14:55 +01:00
|
|
|
static void diff_pixels_c(int16_t *restrict block, const uint8_t *s1,
|
2016-09-06 16:06:12 +02:00
|
|
|
const uint8_t *s2, ptrdiff_t stride)
|
2014-02-03 14:29:09 -08:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
/* read the pixels */
|
|
|
|
|
for (i = 0; i < 8; i++) {
|
|
|
|
|
block[0] = s1[0] - s2[0];
|
|
|
|
|
block[1] = s1[1] - s2[1];
|
|
|
|
|
block[2] = s1[2] - s2[2];
|
|
|
|
|
block[3] = s1[3] - s2[3];
|
|
|
|
|
block[4] = s1[4] - s2[4];
|
|
|
|
|
block[5] = s1[5] - s2[5];
|
|
|
|
|
block[6] = s1[6] - s2[6];
|
|
|
|
|
block[7] = s1[7] - s2[7];
|
|
|
|
|
s1 += stride;
|
|
|
|
|
s2 += stride;
|
|
|
|
|
block += 8;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2025-05-27 19:13:45 +02:00
|
|
|
av_cold void ff_pixblockdsp_init(PixblockDSPContext *c, int bits_per_raw_sample)
|
2014-02-03 14:29:09 -08:00
|
|
|
{
|
2025-05-27 19:13:45 +02:00
|
|
|
const unsigned high_bit_depth = bits_per_raw_sample > 8 &&
|
|
|
|
|
bits_per_raw_sample <= 16;
|
2014-02-03 14:29:09 -08:00
|
|
|
|
2017-08-19 23:38:58 +02:00
|
|
|
c->diff_pixels_unaligned =
|
2014-02-03 14:29:09 -08:00
|
|
|
c->diff_pixels = diff_pixels_c;
|
|
|
|
|
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
if (high_bit_depth) {
|
2024-07-25 18:38:32 +03:00
|
|
|
c->get_pixels_unaligned = get_pixels_unaligned_16_c;
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
c->get_pixels = get_pixels_16_c;
|
|
|
|
|
} else {
|
|
|
|
|
c->get_pixels_unaligned =
|
|
|
|
|
c->get_pixels = get_pixels_8_c;
|
2014-02-03 14:29:09 -08:00
|
|
|
}
|
|
|
|
|
|
2022-06-12 05:51:12 +02:00
|
|
|
#if ARCH_AARCH64
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_aarch64(c, high_bit_depth);
|
2022-06-12 05:51:12 +02:00
|
|
|
#elif ARCH_ARM
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_arm(c, high_bit_depth);
|
2022-06-12 05:51:12 +02:00
|
|
|
#elif ARCH_PPC
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_ppc(c, high_bit_depth);
|
2022-09-26 17:52:24 +03:00
|
|
|
#elif ARCH_RISCV
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_riscv(c, high_bit_depth);
|
2022-06-12 05:51:12 +02:00
|
|
|
#elif ARCH_X86
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_x86(c, high_bit_depth);
|
2022-06-12 05:51:12 +02:00
|
|
|
#elif ARCH_MIPS
|
avcodec/pixblockdsp: Improve 8 vs 16 bit check
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2025-05-27 17:09:34 +02:00
|
|
|
ff_pixblockdsp_init_mips(c, high_bit_depth);
|
2022-06-12 05:51:12 +02:00
|
|
|
#endif
|
2014-02-03 14:29:09 -08:00
|
|
|
}
|