2024-06-11 14:28:55 +02:00
|
|
|
/*
|
|
|
|
|
* This file is part of FFmpeg.
|
|
|
|
|
*
|
|
|
|
|
* FFmpeg is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
|
* (at your option) any later version.
|
|
|
|
|
*
|
|
|
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License along
|
|
|
|
|
* with FFmpeg; if not, write to the Free Software Foundation, Inc.,
|
|
|
|
|
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <string.h>
|
|
|
|
|
|
|
|
|
|
#include "libavutil/common.h"
|
|
|
|
|
#include "libavutil/intreadwrite.h"
|
|
|
|
|
#include "libavutil/mem.h"
|
|
|
|
|
#include "libavutil/mem_internal.h"
|
|
|
|
|
|
|
|
|
|
#include "libswscale/swscale.h"
|
|
|
|
|
#include "libswscale/swscale_internal.h"
|
|
|
|
|
|
|
|
|
|
#include "checkasm.h"
|
|
|
|
|
|
2024-09-18 16:43:34 +02:00
|
|
|
static const enum AVPixelFormat pixel_formats[] = {
|
|
|
|
|
AV_PIX_FMT_YUV444P,
|
|
|
|
|
AV_PIX_FMT_YUV444P9,
|
|
|
|
|
AV_PIX_FMT_YUV444P10,
|
|
|
|
|
AV_PIX_FMT_YUV444P12,
|
|
|
|
|
AV_PIX_FMT_YUV444P14,
|
|
|
|
|
AV_PIX_FMT_YUV444P16,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static void randomize_buffers(int16_t *buf0, int16_t *buf1, int bit_depth, int width)
|
|
|
|
|
{
|
|
|
|
|
int32_t *buf0_32 = (int32_t *) buf0;
|
|
|
|
|
int32_t *buf1_32 = (int32_t *) buf1;
|
|
|
|
|
int mask = (1 << bit_depth) - 1;
|
|
|
|
|
int src_shift = bit_depth <= 14 ? 15 - bit_depth : 19 - bit_depth;
|
|
|
|
|
for (int i = 0; i < width; i++) {
|
|
|
|
|
int32_t r = rnd() & mask;
|
|
|
|
|
if (bit_depth == 16) {
|
|
|
|
|
buf0_32[i] = r << src_shift;
|
|
|
|
|
buf1_32[i] = r << src_shift;
|
|
|
|
|
} else {
|
|
|
|
|
buf0[i] = r << src_shift;
|
|
|
|
|
buf1[i] = r << src_shift;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-11 14:28:55 +02:00
|
|
|
static void check_lumConvertRange(int from)
|
|
|
|
|
{
|
|
|
|
|
const char *func_str = from ? "lumRangeFromJpeg" : "lumRangeToJpeg";
|
2024-09-18 16:35:23 +02:00
|
|
|
#define LARGEST_INPUT_SIZE 1920
|
|
|
|
|
static const int input_sizes[] = {8, LARGEST_INPUT_SIZE};
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
SwsContext *sws;
|
|
|
|
|
SwsInternal *c;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
2024-09-18 16:43:34 +02:00
|
|
|
LOCAL_ALIGNED_32(int16_t, dst0, [LARGEST_INPUT_SIZE * 2]);
|
|
|
|
|
LOCAL_ALIGNED_32(int16_t, dst1, [LARGEST_INPUT_SIZE * 2]);
|
2024-12-01 17:09:40 +01:00
|
|
|
int32_t *dst0_32 = (int32_t *) dst0;
|
|
|
|
|
int32_t *dst1_32 = (int32_t *) dst1;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
declare_func(void, int16_t *dst, int width,
|
|
|
|
|
uint32_t coeff, int64_t offset);
|
2024-06-11 14:28:55 +02:00
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
sws = sws_alloc_context();
|
|
|
|
|
if (sws_init_context(sws, NULL, NULL) < 0)
|
2024-06-11 14:28:55 +02:00
|
|
|
fail();
|
|
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
c = sws_internal(sws);
|
2024-10-24 12:12:08 +02:00
|
|
|
sws->src_range = from;
|
|
|
|
|
sws->dst_range = !from;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
2024-09-18 16:43:34 +02:00
|
|
|
for (int pfi = 0; pfi < FF_ARRAY_ELEMS(pixel_formats); pfi++) {
|
|
|
|
|
enum AVPixelFormat pix_fmt = pixel_formats[pfi];
|
|
|
|
|
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
|
|
|
|
|
int bit_depth = desc->comp[0].depth;
|
|
|
|
|
int sample_size = bit_depth == 16 ? sizeof(int32_t) : sizeof(int16_t);
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
int src_shift = bit_depth <= 14 ? 15 - bit_depth : 19 - bit_depth;
|
|
|
|
|
int mpeg_min = 16 << (bit_depth - 8);
|
|
|
|
|
int mpeg_max = 235 << (bit_depth - 8);
|
|
|
|
|
int jpeg_max = (1 << bit_depth) - 1;
|
2024-10-24 12:12:08 +02:00
|
|
|
sws->src_format = pix_fmt;
|
|
|
|
|
sws->dst_format = pix_fmt;
|
2024-09-18 16:43:34 +02:00
|
|
|
c->dstBpc = bit_depth;
|
|
|
|
|
ff_sws_init_scale(c);
|
2024-09-23 14:03:43 +02:00
|
|
|
for (int dstWi = 0; dstWi < FF_ARRAY_ELEMS(input_sizes); dstWi++) {
|
|
|
|
|
int width = input_sizes[dstWi];
|
|
|
|
|
if (check_func(c->lumConvertRange, "%s%d_%d", func_str, bit_depth, width)) {
|
|
|
|
|
randomize_buffers(dst0, dst1, bit_depth, width);
|
2024-12-01 17:09:40 +01:00
|
|
|
if (bit_depth == 16) {
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
dst1_32[0] = dst0_32[0] = mpeg_min << src_shift;
|
|
|
|
|
dst1_32[1] = dst0_32[1] = mpeg_max << src_shift;
|
|
|
|
|
}
|
2024-12-01 17:09:40 +01:00
|
|
|
dst1_32[2] = dst0_32[2] = -1;
|
|
|
|
|
} else {
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
dst1[0] = dst0[0] = mpeg_min << src_shift;
|
|
|
|
|
dst1[1] = dst0[1] = mpeg_max << src_shift;
|
|
|
|
|
}
|
2024-12-01 17:09:40 +01:00
|
|
|
dst1[2] = dst0[2] = -1;
|
|
|
|
|
}
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
call_ref(dst0, width,
|
|
|
|
|
c->lumConvertRange_coeff, c->lumConvertRange_offset);
|
|
|
|
|
call_new(dst1, width,
|
|
|
|
|
c->lumConvertRange_coeff, c->lumConvertRange_offset);
|
2024-09-23 14:03:43 +02:00
|
|
|
if (memcmp(dst0, dst1, width * sample_size))
|
|
|
|
|
fail();
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
/* check that the mpeg range is respected */
|
|
|
|
|
if (bit_depth == 16) {
|
|
|
|
|
if ((dst1_32[0] >> src_shift) > 0 || (dst1_32[1] >> src_shift) != jpeg_max)
|
|
|
|
|
fail();
|
|
|
|
|
} else {
|
|
|
|
|
if ((dst1[0] >> src_shift) > 0 || (dst1[1] >> src_shift) != jpeg_max)
|
|
|
|
|
fail();
|
|
|
|
|
}
|
|
|
|
|
}
|
2024-09-23 14:03:43 +02:00
|
|
|
if (width == LARGEST_INPUT_SIZE && (bit_depth == 8 || bit_depth == 16))
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
bench_new(dst1, width,
|
|
|
|
|
c->lumConvertRange_coeff, c->lumConvertRange_offset);
|
2024-09-23 14:03:43 +02:00
|
|
|
}
|
2024-06-11 14:28:55 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
sws_freeContext(sws);
|
2024-06-11 14:28:55 +02:00
|
|
|
}
|
|
|
|
|
#undef LARGEST_INPUT_SIZE
|
|
|
|
|
|
|
|
|
|
static void check_chrConvertRange(int from)
|
|
|
|
|
{
|
|
|
|
|
const char *func_str = from ? "chrRangeFromJpeg" : "chrRangeToJpeg";
|
2024-09-18 16:35:23 +02:00
|
|
|
#define LARGEST_INPUT_SIZE 1920
|
|
|
|
|
static const int input_sizes[] = {8, LARGEST_INPUT_SIZE};
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
SwsContext *sws;
|
|
|
|
|
SwsInternal *c;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
2024-09-18 16:43:34 +02:00
|
|
|
LOCAL_ALIGNED_32(int16_t, dstU0, [LARGEST_INPUT_SIZE * 2]);
|
|
|
|
|
LOCAL_ALIGNED_32(int16_t, dstV0, [LARGEST_INPUT_SIZE * 2]);
|
|
|
|
|
LOCAL_ALIGNED_32(int16_t, dstU1, [LARGEST_INPUT_SIZE * 2]);
|
|
|
|
|
LOCAL_ALIGNED_32(int16_t, dstV1, [LARGEST_INPUT_SIZE * 2]);
|
2024-12-01 17:09:40 +01:00
|
|
|
int32_t *dstU0_32 = (int32_t *) dstU0;
|
|
|
|
|
int32_t *dstU1_32 = (int32_t *) dstU1;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
declare_func(void, int16_t *dstU, int16_t *dstV, int width,
|
|
|
|
|
uint32_t coeff, int64_t offset);
|
2024-06-11 14:28:55 +02:00
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
sws = sws_alloc_context();
|
|
|
|
|
if (sws_init_context(sws, NULL, NULL) < 0)
|
2024-06-11 14:28:55 +02:00
|
|
|
fail();
|
|
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
c = sws_internal(sws);
|
2024-10-24 12:12:08 +02:00
|
|
|
sws->src_range = from;
|
|
|
|
|
sws->dst_range = !from;
|
2024-06-11 14:28:55 +02:00
|
|
|
|
2024-09-18 16:43:34 +02:00
|
|
|
for (int pfi = 0; pfi < FF_ARRAY_ELEMS(pixel_formats); pfi++) {
|
|
|
|
|
enum AVPixelFormat pix_fmt = pixel_formats[pfi];
|
|
|
|
|
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
|
|
|
|
|
int bit_depth = desc->comp[0].depth;
|
|
|
|
|
int sample_size = bit_depth == 16 ? sizeof(int32_t) : sizeof(int16_t);
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
int src_shift = bit_depth <= 14 ? 15 - bit_depth : 19 - bit_depth;
|
|
|
|
|
int mpeg_min = 16 << (bit_depth - 8);
|
|
|
|
|
int mpeg_max = 240 << (bit_depth - 8);
|
|
|
|
|
int jpeg_max = (1 << bit_depth) - 1;
|
2024-10-24 12:12:08 +02:00
|
|
|
sws->src_format = pix_fmt;
|
|
|
|
|
sws->dst_format = pix_fmt;
|
2024-09-18 16:43:34 +02:00
|
|
|
c->dstBpc = bit_depth;
|
|
|
|
|
ff_sws_init_scale(c);
|
2024-09-23 14:03:43 +02:00
|
|
|
for (int dstWi = 0; dstWi < FF_ARRAY_ELEMS(input_sizes); dstWi++) {
|
|
|
|
|
int width = input_sizes[dstWi];
|
|
|
|
|
if (check_func(c->chrConvertRange, "%s%d_%d", func_str, bit_depth, width)) {
|
|
|
|
|
randomize_buffers(dstU0, dstU1, bit_depth, width);
|
|
|
|
|
randomize_buffers(dstV0, dstV1, bit_depth, width);
|
2024-12-01 17:09:40 +01:00
|
|
|
if (bit_depth == 16) {
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
dstU1_32[0] = dstU0_32[0] = mpeg_min << src_shift;
|
|
|
|
|
dstU1_32[1] = dstU0_32[1] = mpeg_max << src_shift;
|
|
|
|
|
}
|
2024-12-01 17:09:40 +01:00
|
|
|
dstU1_32[2] = dstU0_32[2] = -1;
|
|
|
|
|
} else {
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
dstU1[0] = dstU0[0] = mpeg_min << src_shift;
|
|
|
|
|
dstU1[1] = dstU0[1] = mpeg_max << src_shift;
|
|
|
|
|
}
|
2024-12-01 17:09:40 +01:00
|
|
|
dstU1[2] = dstU0[2] = -1;
|
|
|
|
|
}
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
call_ref(dstU0, dstV0, width,
|
|
|
|
|
c->chrConvertRange_coeff, c->chrConvertRange_offset);
|
|
|
|
|
call_new(dstU1, dstV1, width,
|
|
|
|
|
c->chrConvertRange_coeff, c->chrConvertRange_offset);
|
2024-09-23 14:03:43 +02:00
|
|
|
if (memcmp(dstU0, dstU1, width * sample_size) ||
|
|
|
|
|
memcmp(dstV0, dstV1, width * sample_size))
|
|
|
|
|
fail();
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
if (!from) {
|
|
|
|
|
/* check that the mpeg range is respected */
|
|
|
|
|
if (bit_depth == 16) {
|
|
|
|
|
if ((dstU1_32[0] >> src_shift) > 0 || (dstU1_32[1] >> src_shift) != jpeg_max)
|
|
|
|
|
fail();
|
|
|
|
|
} else {
|
|
|
|
|
if ((dstU1[0] >> src_shift) > 0 || (dstU1[1] >> src_shift) != jpeg_max)
|
|
|
|
|
fail();
|
|
|
|
|
}
|
|
|
|
|
}
|
2024-09-23 14:03:43 +02:00
|
|
|
if (width == LARGEST_INPUT_SIZE && (bit_depth == 8 || bit_depth == 16))
|
swscale/range_convert: fix mpeg ranges in yuv range conversion for non-8-bit pixel formats
There is an issue with the constants used in YUV to YUV range conversion,
where the upper bound is not respected when converting to mpeg range.
With this commit, the constants are calculated at runtime, depending on
the bit depth. This approach also allows us to more easily understand how
the constants are derived.
For bit depths <= 14, the number of fixed point bits has been set to 14
for all conversions, to simplify the code.
For bit depths > 14, the number of fixed points bits has been raised and
set to 18, to allow for the conversion to be accurate enough for the mpeg
range to be respected.
The convert functions now take the conversion constants (coeff and offset)
as function arguments.
For bit depths <= 14, coeff is unsigned 16-bit and offset is 32-bit.
For bit depths > 14, coeff is unsigned 32-bit and offset is 64-bit.
x86_64:
chrRangeFromJpeg8_1920_c: 2127.4 2125.0 (1.00x)
chrRangeFromJpeg16_1920_c: 2325.2 2127.2 (1.09x)
chrRangeToJpeg8_1920_c: 3166.9 3168.7 (1.00x)
chrRangeToJpeg16_1920_c: 2152.4 3164.8 (0.68x)
lumRangeFromJpeg8_1920_c: 1263.0 1302.5 (0.97x)
lumRangeFromJpeg16_1920_c: 1080.5 1299.2 (0.83x)
lumRangeToJpeg8_1920_c: 1886.8 2112.2 (0.89x)
lumRangeToJpeg16_1920_c: 1077.0 1906.5 (0.56x)
aarch64 A55:
chrRangeFromJpeg8_1920_c: 28835.2 28835.6 (1.00x)
chrRangeFromJpeg16_1920_c: 28839.8 32680.8 (0.88x)
chrRangeToJpeg8_1920_c: 23074.7 23075.4 (1.00x)
chrRangeToJpeg16_1920_c: 17318.9 24996.0 (0.69x)
lumRangeFromJpeg8_1920_c: 15389.7 15384.5 (1.00x)
lumRangeFromJpeg16_1920_c: 15388.2 17306.7 (0.89x)
lumRangeToJpeg8_1920_c: 19227.8 19226.6 (1.00x)
lumRangeToJpeg16_1920_c: 15387.0 21146.3 (0.73x)
aarch64 A76:
chrRangeFromJpeg8_1920_c: 6324.4 6268.1 (1.01x)
chrRangeFromJpeg16_1920_c: 6339.9 11521.5 (0.55x)
chrRangeToJpeg8_1920_c: 9656.0 9612.8 (1.00x)
chrRangeToJpeg16_1920_c: 6340.4 11651.8 (0.54x)
lumRangeFromJpeg8_1920_c: 4422.0 4420.8 (1.00x)
lumRangeFromJpeg16_1920_c: 4420.9 5762.0 (0.77x)
lumRangeToJpeg8_1920_c: 5949.1 5977.5 (1.00x)
lumRangeToJpeg16_1920_c: 4446.8 5946.2 (0.75x)
NOTE: all simd optimizations for range_convert have been disabled.
they will be re-enabled when they are fixed for each architecture.
NOTE2: the same issue still exists in rgb2yuv conversions, which is not
addressed in this commit.
2024-09-18 23:46:06 +02:00
|
|
|
bench_new(dstU1, dstV1, width,
|
|
|
|
|
c->chrConvertRange_coeff, c->chrConvertRange_offset);
|
2024-09-23 14:03:43 +02:00
|
|
|
}
|
2024-06-11 14:28:55 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
swscale: rename SwsContext to SwsInternal
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
2024-10-09 19:44:33 +02:00
|
|
|
sws_freeContext(sws);
|
2024-06-11 14:28:55 +02:00
|
|
|
}
|
|
|
|
|
#undef LARGEST_INPUT_SIZE
|
|
|
|
|
|
|
|
|
|
void checkasm_check_sw_range_convert(void)
|
|
|
|
|
{
|
|
|
|
|
check_lumConvertRange(1);
|
|
|
|
|
report("lumRangeFromJpeg");
|
|
|
|
|
check_chrConvertRange(1);
|
|
|
|
|
report("chrRangeFromJpeg");
|
|
|
|
|
check_lumConvertRange(0);
|
|
|
|
|
report("lumRangeToJpeg");
|
|
|
|
|
check_chrConvertRange(0);
|
|
|
|
|
report("chrRangeToJpeg");
|
|
|
|
|
}
|