coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2026-10-28 18:07:12 +08:00
Code Issues Packages Projects Releases Wiki Activity
libyuv/source/convert_to_argb.cc
Frank Barchard f249eaa72f [libyuv] Fix top 20 bugs and performance issues 
1. Fix multi-bit TestCpuFlag(A | B) checks in AYUVToNV12, AYUVToNV21,
   RAWToRGBA, NV12ToRGB24Matrix, NV21ToRGB24Matrix, MergeRGBPlane, and
   MergeXR30Plane to compare == (A | B) so AVX-512 VBMI/BW kernels are
   not dispatched when AVX512VBMI or AVX512VL is absent.
2. Fix 2-byte SVE2 heap buffer overflow on odd widths in
   I422ToARGB1555Row_SVE_SC and I422ToARGB4444Row_SVE_SC using
   trn1/trn2 + st2b under .b predicates.
3. Fix ARGBPolynomialRow_SSE2 to use packssdw before packuswb, and add
   missing vzeroupper to SplitRGBRow_AVX2.
4. Fix CalcFrameSsim to return 0.0 instead of 0.0 / 0 (NaN) when
   width <= 8 || height <= 8 so I420Ssim stays finite on small frames.
5. Fix negative-height vertical inversion in ConvertFromI420,
   SplitARGBPlane (when dst_a is NULL), I422Rotate, I010Rotate, and
   I210Rotate.
6. Fix horizontal flip (negative x/dx) in ScaleFilterCols_NEON (signed
   sbfx #16, #16 offset extraction) and ScaleAddCols1_16_C (shift
   x >>= 16 before indexing src_ptr).
7. Fix ConvertToARGB odd crop_x chroma alignment in NV12/NV21, src_u
   offset in I420/YV12/J420/H420/U420, and FOURCC_U422 calling
   U422ToARGB instead of H422ToARGB.
8. Remove dead 14-instruction block in ScaleRowUp2_Bilinear_16_SSE2.
9. Fix UVScale_16 for src_width == 1 (P010ToP410/P210ToP410), enable
   and fix column order in TransposeWx4_32_SVE2, fix ByteToFloatRow
   NEON Any BPP (3 -> 4), fix AB64ToARGBRow_Any_*, ScaleColsUp2_SSE2,
   ScaleUVRowDown2Box_SME, and UYVYToARGBRow_SME macro guards, fix SVE2
   loop tail branches (b.lt/b.ge), and remove duplicate HAS_* macros.
10. Fix ConvertFP16ToFP32Column_NEON loop condition and unsigned b.lo
    underflow on widths < 8 or not multiples of 8.
11. Fix UYVYToNV12 IS_ALIGNED(awidth, ...) checks and WASM SIMD
    IS_ALIGNED(width, 16) checks in I420ToRGB24Matrix /
    I422ToRGB24Matrix.
12. Wire SplitUVRow_SVE2, SplitUVRow_LSX, and SplitUVRow_LASX in
    NV24ToARGB.
13. Fix MJpegDecoder heap buffer overflows in LoadFrame (reallocate
    databuf_[i] when scanlines_size changes), DecodeToBuffers, and
    DecodeToCallback (clamp scanlines_to_copy to lines_left and align
    skip to max_v_samp_factor).
14. Replace width & (vl - 1) tail calculations with post-loop remainder
    across 18 SVE2 functions in row_sve.h for non-power-of-2 SVE VL,
    and remove duplicate label 2: in UYVYToARGBRow_SVE_SC.
15. Handle width & 15 tail pixels in DetileToYUY2_C and validate
    IS_POWEROFTWO(tile_height) in DetileSplitUVPlane and DetileToYUY2.
16. Avoid signed 32-bit overflow in DivideRow_16_C and C16TO8, fix
    ClampFloatToZero_SSE2 pointer increment (lea 0x4) and loop counter
    (sub $0x1), and initialize xmm3 in ARGBCopyYToAlphaRow_SSE2.
17. Check max leaf 7 subleaf and max extended CPUID leaf in GetCpuFlags
    before querying subleaf 7:1 and extended leaf 0x80000021.
18. Free plane_u on error in I422ToNV21, and fix 4x/2x scratch row
    over-allocation in ScaleARGBBilinearDown and ScaleUVBilinearDown.
19. Add ARGBSetRow_RVV and ARGBShuffleRow_RVV in row_rvv.cc.
20. Fix NEON mask (15 -> 7) in ScaleUVRowUp2_Linear_Any_NEON /
    _16_Any_NEON, use SIMD InterpolateRow + SplitUVRow in NV16ToI420,
    wire SSSE3/SSE2 in ARGBToI400Matrix / ARGBToYUY2Matrix /
    ARGBToUYVYMatrix and RGB24Mirror, and split HalfFloatRow multiply
    into (src * scale) * 2^-112 to avoid 50x Intel denormal stalls.

Bug: b/574714081
Test: libyuv_unittest
Change-Id: Iec3ae5c58d52bee3ef13cf9b765ac84b733aa576
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/8601931
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Commit-Queue: Frank Barchard <fbarchard@google.com>
2026-10-19 16:46:40 -07:00

750 lines
26 KiB
C++
Raw Blame History

/*
* Copyright 2011 The LibYuv Project Authors. All rights reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "libyuv/convert_argb.h"
#include <limits.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
#include "libyuv/cpu_id.h"
#ifdef HAVE_JPEG
#include "libyuv/mjpeg_decoder.h"
#endif
#include "libyuv/rotate_argb.h"
#include "libyuv/row.h"
#include "libyuv/video_common.h"
#ifdef __cplusplus
namespace libyuv {
extern "C" {
#endif
static int NV16ToARGB(const uint8_t* src_y,
int src_stride_y,
const uint8_t* src_uv,
int src_stride_uv,
uint8_t* dst_argb,
int dst_stride_argb,
int width,
int height) {
int y;
void (*NV12ToARGBRow)(
const uint8_t* y_buf, const uint8_t* uv_buf, uint8_t* rgb_buf,
const struct YuvConstants* yuvconstants, int width) = NV12ToARGBRow_C;
if (!src_y || !src_uv || !dst_argb || width <= 0 || height == 0 ||
height == INT_MIN) {
return -1;
}
if (height < 0) {
height = -height;
dst_argb = dst_argb + (ptrdiff_t)(height - 1) * dst_stride_argb;
dst_stride_argb = -dst_stride_argb;
}
#if defined(HAS_NV12TOARGBROW_SSSE3)
if (TestCpuFlag(kCpuHasSSSE3)) {
NV12ToARGBRow = NV12ToARGBRow_Any_SSSE3;
if (IS_ALIGNED(width, 8)) {
NV12ToARGBRow = NV12ToARGBRow_SSSE3;
}
}
#endif
#if defined(HAS_NV12TOARGBROW_AVX2)
if (TestCpuFlag(kCpuHasAVX2)) {
NV12ToARGBRow = NV12ToARGBRow_Any_AVX2;
if (IS_ALIGNED(width, 16)) {
NV12ToARGBRow = NV12ToARGBRow_AVX2;
}
}
#endif
#if defined(HAS_NV12TOARGBROW_AVX512BW)
if (TestCpuFlag(kCpuHasAVX512BW)) {
NV12ToARGBRow = NV12ToARGBRow_AVX512BW;
}
#endif
#if defined(HAS_NV12TOARGBROW_NEON)
if (TestCpuFlag(kCpuHasNEON)) {
NV12ToARGBRow = NV12ToARGBRow_Any_NEON;
if (IS_ALIGNED(width, 8)) {
NV12ToARGBRow = NV12ToARGBRow_NEON;
}
}
#endif
#if defined(HAS_NV12TOARGBROW_SVE2)
if (TestCpuFlag(kCpuHasSVE2)) {
NV12ToARGBRow = NV12ToARGBRow_SVE2;
}
#endif
#if defined(HAS_NV12TOARGBROW_SME)
if (TestCpuFlag(kCpuHasSME)) {
NV12ToARGBRow = NV12ToARGBRow_SME;
}
#endif
#if defined(HAS_NV12TOARGBROW_LSX)
if (TestCpuFlag(kCpuHasLSX)) {
NV12ToARGBRow = NV12ToARGBRow_Any_LSX;
if (IS_ALIGNED(width, 8)) {
NV12ToARGBRow = NV12ToARGBRow_LSX;
}
}
#endif
#if defined(HAS_NV12TOARGBROW_LASX)
if (TestCpuFlag(kCpuHasLASX)) {
NV12ToARGBRow = NV12ToARGBRow_Any_LASX;
if (IS_ALIGNED(width, 16)) {
NV12ToARGBRow = NV12ToARGBRow_LASX;
}
}
#endif
#if defined(HAS_NV12TOARGBROW_RVV)
if (TestCpuFlag(kCpuHasRVV)) {
NV12ToARGBRow = NV12ToARGBRow_RVV;
}
#endif
#if defined(HAS_NV12TOARGBROW_HVX)
if (TestCpuFlag(kCpuHasHVX)) {
NV12ToARGBRow = NV12ToARGBRow_HVX;
}
#endif
#if defined(HAS_NV12TOARGBROW_WASMSIMD)
if (TestCpuFlag(kCpuHasWASMSIMD)) {
NV12ToARGBRow = NV12ToARGBRow_Any_WASMSIMD;
if (IS_ALIGNED(width, 8)) {
NV12ToARGBRow = NV12ToARGBRow_WASMSIMD;
}
}
#endif
for (y = 0; y < height; ++y) {
NV12ToARGBRow(src_y, src_uv, dst_argb, &kYuvI601Constants, width);
dst_argb += dst_stride_argb;
src_y += src_stride_y;
src_uv += src_stride_uv;
}
return 0;
}
static int NV24ToARGB(const uint8_t* src_y,
int src_stride_y,
const uint8_t* src_uv,
int src_stride_uv,
uint8_t* dst_argb,
int dst_stride_argb,
int width,
int height) {
int y;
void (*I444ToARGBRow)(const uint8_t* y_buf, const uint8_t* u_buf,
const uint8_t* v_buf, uint8_t* rgb_buf,
const struct YuvConstants* yuvconstants, int width) =
I444ToARGBRow_C;
void (*SplitUVRow)(const uint8_t* src_uv, uint8_t* dst_u, uint8_t* dst_v,
int width) = SplitUVRow_C;
if (!src_y || !src_uv || !dst_argb || width <= 0 || height == 0 ||
height == INT_MIN) {
return -1;
}
if (height < 0) {
height = -height;
dst_argb = dst_argb + (ptrdiff_t)(height - 1) * dst_stride_argb;
dst_stride_argb = -dst_stride_argb;
}
#if defined(HAS_I444TOARGBROW_SSSE3)
if (TestCpuFlag(kCpuHasSSSE3)) {
I444ToARGBRow = I444ToARGBRow_Any_SSSE3;
if (IS_ALIGNED(width, 8)) {
I444ToARGBRow = I444ToARGBRow_SSSE3;
}
}
#endif
#if defined(HAS_I444TOARGBROW_AVX2)
if (TestCpuFlag(kCpuHasAVX2)) {
I444ToARGBRow = I444ToARGBRow_Any_AVX2;
if (IS_ALIGNED(width, 16)) {
I444ToARGBRow = I444ToARGBRow_AVX2;
}
}
#endif
#if defined(HAS_I444TOARGBROW_AVX512BW)
if (TestCpuFlag(kCpuHasAVX512BW)) {
I444ToARGBRow = I444ToARGBRow_Any_AVX512BW;
if (IS_ALIGNED(width, 32)) {
I444ToARGBRow = I444ToARGBRow_AVX512BW;
}
}
#endif
#if defined(HAS_I444TOARGBROW_NEON)
if (TestCpuFlag(kCpuHasNEON)) {
I444ToARGBRow = I444ToARGBRow_Any_NEON;
if (IS_ALIGNED(width, 8)) {
I444ToARGBRow = I444ToARGBRow_NEON;
}
}
#endif
#if defined(HAS_I444TOARGBROW_SVE2)
if (TestCpuFlag(kCpuHasSVE2)) {
I444ToARGBRow = I444ToARGBRow_SVE2;
}
#endif
#if defined(HAS_I444TOARGBROW_SME)
if (TestCpuFlag(kCpuHasSME)) {
I444ToARGBRow = I444ToARGBRow_SME;
}
#endif
#if defined(HAS_I444TOARGBROW_LSX)
if (TestCpuFlag(kCpuHasLSX)) {
I444ToARGBRow = I444ToARGBRow_Any_LSX;
if (IS_ALIGNED(width, 8)) {
I444ToARGBRow = I444ToARGBRow_LSX;
}
}
#endif
#if defined(HAS_I444TOARGBROW_LASX)
if (TestCpuFlag(kCpuHasLASX)) {
I444ToARGBRow = I444ToARGBRow_Any_LASX;
if (IS_ALIGNED(width, 16)) {
I444ToARGBRow = I444ToARGBRow_LASX;
}
}
#endif
#if defined(HAS_I444TOARGBROW_RVV)
if (TestCpuFlag(kCpuHasRVV)) {
I444ToARGBRow = I444ToARGBRow_RVV;
}
#endif
#if defined(HAS_I444TOARGBROW_HVX)
if (TestCpuFlag(kCpuHasHVX)) {
I444ToARGBRow = I444ToARGBRow_HVX;
}
#endif
#if defined(HAS_SPLITUVROW_SSE2)
if (TestCpuFlag(kCpuHasSSE2)) {
SplitUVRow = SplitUVRow_Any_SSE2;
if (IS_ALIGNED(width, 16)) {
SplitUVRow = SplitUVRow_SSE2;
}
}
#endif
#if defined(HAS_SPLITUVROW_AVX2)
if (TestCpuFlag(kCpuHasAVX2)) {
SplitUVRow = SplitUVRow_Any_AVX2;
if (IS_ALIGNED(width, 32)) {
SplitUVRow = SplitUVRow_AVX2;
}
}
#endif
#if defined(HAS_SPLITUVROW_AVX512BW)
if (TestCpuFlag(kCpuHasAVX512BW)) {
SplitUVRow = SplitUVRow_Any_AVX512BW;
if (IS_ALIGNED(width, 64)) {
SplitUVRow = SplitUVRow_AVX512BW;
}
}
#endif
#if defined(HAS_SPLITUVROW_NEON)
if (TestCpuFlag(kCpuHasNEON)) {
SplitUVRow = SplitUVRow_Any_NEON;
if (IS_ALIGNED(width, 16)) {
SplitUVRow = SplitUVRow_NEON;
}
}
#endif
#if defined(HAS_SPLITUVROW_SVE2)
if (TestCpuFlag(kCpuHasSVE2)) {
SplitUVRow = SplitUVRow_SVE2;
}
#endif
#if defined(HAS_SPLITUVROW_LSX)
if (TestCpuFlag(kCpuHasLSX)) {
SplitUVRow = SplitUVRow_Any_LSX;
if (IS_ALIGNED(width, 16)) {
SplitUVRow = SplitUVRow_LSX;
}
}
#endif
#if defined(HAS_SPLITUVROW_LASX)
if (TestCpuFlag(kCpuHasLASX)) {
SplitUVRow = SplitUVRow_Any_LASX;
if (IS_ALIGNED(width, 32)) {
SplitUVRow = SplitUVRow_LASX;
}
}
#endif
#if defined(HAS_SPLITUVROW_RVV)
if (TestCpuFlag(kCpuHasRVV)) {
SplitUVRow = SplitUVRow_RVV;
}
#endif
#if defined(HAS_SPLITUVROW_HVX)
if (TestCpuFlag(kCpuHasHVX)) {
SplitUVRow = SplitUVRow_HVX;
}
#endif
align_buffer_64(row_u, width);
align_buffer_64(row_v, width);
if (!row_u || !row_v) {
free_aligned_buffer_64(row_u);
free_aligned_buffer_64(row_v);
return 1;
}
for (y = 0; y < height; ++y) {
SplitUVRow(src_uv, row_u, row_v, width);
I444ToARGBRow(src_y, row_u, row_v, dst_argb, &kYuvI601Constants, width);
dst_argb += dst_stride_argb;
src_y += src_stride_y;
src_uv += src_stride_uv;
}
free_aligned_buffer_64(row_u);
free_aligned_buffer_64(row_v);
return 0;
}
// Convert camera sample to ARGB with cropping, rotation and vertical flip.
// src_width is used for source stride computation
// src_height is used to compute location of planes, and indicate inversion
// sample_size is measured in bytes and is the size of the frame.
// With MJPEG it is the compressed size of the frame.
// TODO(fbarchard): Add the following:
// H010ToARGB
// I010ToARGB
LIBYUV_API
int ConvertToARGB(const uint8_t* sample,
size_t sample_size,
uint8_t* dst_argb,
int dst_stride_argb,
int crop_x,
int crop_y,
int src_width,
int src_height,
int crop_width,
int crop_height,
enum RotationMode rotation,
uint32_t fourcc) {
if (src_height == INT_MIN || crop_height == INT_MIN) {
return -1;
}
int abs_src_height = (src_height < 0) ? -src_height : src_height;
int abs_crop_height = (crop_height < 0) ? -crop_height : crop_height;
if (dst_argb == NULL || sample == NULL || src_width <= 0 ||
src_width > INT_MAX / 4 || crop_width <= 0 || crop_width > INT_MAX / 4 ||
src_height == 0 || crop_height == 0 || crop_x < 0 || crop_y < 0 ||
crop_width > src_width || crop_x > src_width - crop_width ||
abs_crop_height > abs_src_height ||
crop_y > abs_src_height - abs_crop_height) {
return -1;
}
uint32_t format = CanonicalFourCC(fourcc);
int aligned_src_width = (src_width + 1) & ~1;
const uint8_t* src;
const uint8_t* src_uv;
int r = 0;
// One pass rotation is available for some formats. For the rest, convert
// to ARGB (with optional vertical flipping) into a temporary ARGB buffer,
// and then rotate the ARGB to the final destination buffer.
// For in-place conversion, if destination dst_argb is same as source sample,
// also enable temporary buffer.
int need_buf = (rotation && format != FOURCC_ARGB) || dst_argb == sample;
uint8_t* dest_argb = dst_argb;
int dest_dst_stride_argb = dst_stride_argb;
uint8_t* rotate_buffer = NULL;
int inv_crop_height = (crop_height < 0) ? -crop_height : crop_height;
if (src_height < 0) {
inv_crop_height = -inv_crop_height;
}
if (need_buf) {
const uint64_t rotate_buffer_size =
(uint64_t)crop_width * 4 * abs_crop_height;
#if UINT64_MAX > SIZE_MAX
if (rotate_buffer_size > SIZE_MAX) {
return -1; // Invalid size.
}
#endif
rotate_buffer = (uint8_t*)malloc((size_t)rotate_buffer_size);
if (!rotate_buffer) {
return 1; // Out of memory runtime error.
}
dst_argb = rotate_buffer;
dst_stride_argb = crop_width * 4;
}
switch (format) {
// Single plane formats
case FOURCC_YUY2:
src = sample + ((ptrdiff_t)aligned_src_width * crop_y + crop_x) * 2;
r = YUY2ToARGB(src, aligned_src_width * 2, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
break;
case FOURCC_UYVY:
src = sample + ((ptrdiff_t)aligned_src_width * crop_y + crop_x) * 2;
r = UYVYToARGB(src, aligned_src_width * 2, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
break;
case FOURCC_24BG:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 3;
r = RGB24ToARGB(src, src_width * 3, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_RAW:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 3;
r = RAWToARGB(src, src_width * 3, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_ARGB:
if (!need_buf && !rotation) {
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = ARGBToARGB(src, src_width * 4, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
}
break;
case FOURCC_BGRA:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = BGRAToARGB(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_ABGR:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = ABGRToARGB(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_RGBA:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = RGBAToARGB(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_AR30:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = AR30ToARGB(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_AB30:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = AB30ToARGB(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_RGBP:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = RGB565ToARGB(src, src_width * 2, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
break;
case FOURCC_RGBO:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = ARGB1555ToARGB(src, src_width * 2, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
break;
case FOURCC_R444:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = ARGB4444ToARGB(src, src_width * 2, dst_argb, dst_stride_argb,
crop_width, inv_crop_height);
break;
case FOURCC_I400:
src = sample + (ptrdiff_t)src_width * crop_y + crop_x;
r = I400ToARGB(src, src_width, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
case FOURCC_J400:
src = sample + (ptrdiff_t)src_width * crop_y + crop_x;
r = J400ToARGB(src, src_width, dst_argb, dst_stride_argb, crop_width,
inv_crop_height);
break;
// Biplanar formats
case FOURCC_NV12:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample +
aligned_src_width * ((ptrdiff_t)abs_src_height + crop_y / 2) +
(crop_x & ~1);
r = NV12ToARGB(src, src_width, src_uv, aligned_src_width, dst_argb,
dst_stride_argb, crop_width, inv_crop_height);
break;
case FOURCC_NV21:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample +
aligned_src_width * ((ptrdiff_t)abs_src_height + crop_y / 2) +
(crop_x & ~1);
// Call NV12 but with u and v parameters swapped.
r = NV21ToARGB(src, src_width, src_uv, aligned_src_width, dst_argb,
dst_stride_argb, crop_width, inv_crop_height);
break;
case FOURCC_NV16:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample +
aligned_src_width * ((ptrdiff_t)abs_src_height + crop_y) +
(crop_x & ~1);
r = NV16ToARGB(src, src_width, src_uv, aligned_src_width, dst_argb,
dst_stride_argb, crop_width, inv_crop_height);
break;
case FOURCC_NV24:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample + (ptrdiff_t)aligned_src_width * abs_src_height +
(ptrdiff_t)aligned_src_width * 2 * crop_y +
(ptrdiff_t)crop_x * 2;
r = NV24ToARGB(src, src_width, src_uv, aligned_src_width * 2, dst_argb,
dst_stride_argb, crop_width, inv_crop_height);
break;
case FOURCC_P010: {
const uint16_t* src16 = (const uint16_t*)sample;
const uint16_t* src_y16 =
src16 + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint16_t* src_uv16 =
src16 + ((ptrdiff_t)src_width * abs_src_height) +
((ptrdiff_t)(crop_y / 2) * aligned_src_width) + ((crop_x / 2) * 2);
r = P010ToARGBMatrix(src_y16, src_width, src_uv16, aligned_src_width,
dst_argb, dst_stride_argb, &kYuvI601Constants,
crop_width, inv_crop_height);
break;
}
case FOURCC_P210: {
const uint16_t* src16 = (const uint16_t*)sample;
const uint16_t* src_y16 =
src16 + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint16_t* src_uv16 =
src16 + ((ptrdiff_t)src_width * abs_src_height) +
((ptrdiff_t)crop_y * aligned_src_width) + ((crop_x / 2) * 2);
r = P210ToARGBMatrix(src_y16, src_width, src_uv16, aligned_src_width,
dst_argb, dst_stride_argb, &kYuvI601Constants,
crop_width, inv_crop_height);
break;
}
case FOURCC_P410: {
const uint16_t* src16 = (const uint16_t*)sample;
const uint16_t* src_y16 =
src16 + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint16_t* src_uv16 =
src16 + ((ptrdiff_t)src_width * abs_src_height) +
((ptrdiff_t)crop_y * src_width * 2) + crop_x * 2;
r = P410ToARGBMatrix(src_y16, src_width, src_uv16, src_width * 2,
dst_argb, dst_stride_argb, &kYuvI601Constants,
crop_width, inv_crop_height);
break;
}
// Triplanar formats
case FOURCC_I420:
case FOURCC_YV12: {
const uint8_t* src_y = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint8_t* src_u;
const uint8_t* src_v;
int halfwidth = (src_width + 1) / 2;
int halfheight = (abs_src_height + 1) / 2;
if (format == FOURCC_YV12) {
src_v = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * (crop_y / 2) + crop_x / 2;
src_u = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)halfheight + crop_y / 2) + crop_x / 2;
} else {
src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * (crop_y / 2) + crop_x / 2;
src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)halfheight + crop_y / 2) + crop_x / 2;
}
r = I420ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_J420: {
int halfwidth = (src_width + 1) / 2;
int halfheight = (abs_src_height + 1) / 2;
const uint8_t* src_y = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * (crop_y / 2) + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)halfheight + crop_y / 2) +
crop_x / 2;
r = J420ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_H420: {
int halfwidth = (src_width + 1) / 2;
int halfheight = (abs_src_height + 1) / 2;
const uint8_t* src_y = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * (crop_y / 2) + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)halfheight + crop_y / 2) +
crop_x / 2;
r = H420ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_U420: {
int halfwidth = (src_width + 1) / 2;
int halfheight = (abs_src_height + 1) / 2;
const uint8_t* src_y = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * (crop_y / 2) + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)halfheight + crop_y / 2) +
crop_x / 2;
r = U420ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_I422:
case FOURCC_YV16: {
int halfwidth = (src_width + 1) / 2;
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u;
const uint8_t* src_v;
if (format == FOURCC_YV16) {
src_v = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * crop_y + crop_x / 2;
src_u = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)abs_src_height + crop_y) + crop_x / 2;
} else {
src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * crop_y + crop_x / 2;
src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)abs_src_height + crop_y) + crop_x / 2;
}
r = I422ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_J422: {
int halfwidth = (src_width + 1) / 2;
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * crop_y + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)abs_src_height + crop_y) +
crop_x / 2;
r = J422ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_H422: {
int halfwidth = (src_width + 1) / 2;
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * crop_y + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)abs_src_height + crop_y) +
crop_x / 2;
r = H422ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_U422: {
int halfwidth = (src_width + 1) / 2;
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u = sample + (ptrdiff_t)src_width * abs_src_height +
(ptrdiff_t)halfwidth * crop_y + crop_x / 2;
const uint8_t* src_v = sample + (ptrdiff_t)src_width * abs_src_height +
halfwidth * ((ptrdiff_t)abs_src_height + crop_y) +
crop_x / 2;
r = U422ToARGB(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_I444:
case FOURCC_YV24: {
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u;
const uint8_t* src_v;
if (format == FOURCC_YV24) {
src_v =
sample + src_width * ((ptrdiff_t)abs_src_height + crop_y) + crop_x;
src_u = sample + src_width * ((ptrdiff_t)abs_src_height * 2 + crop_y) +
crop_x;
} else {
src_u =
sample + src_width * ((ptrdiff_t)abs_src_height + crop_y) + crop_x;
src_v = sample + src_width * ((ptrdiff_t)abs_src_height * 2 + crop_y) +
crop_x;
}
r = I444ToARGB(src_y, src_width, src_u, src_width, src_v, src_width,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_J444: {
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u =
sample + src_width * ((ptrdiff_t)abs_src_height + crop_y) + crop_x;
const uint8_t* src_v =
sample + src_width * ((ptrdiff_t)abs_src_height * 2 + crop_y) +
crop_x;
r = J444ToARGB(src_y, src_width, src_u, src_width, src_v, src_width,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_H444: {
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u =
sample + src_width * ((ptrdiff_t)abs_src_height + crop_y) + crop_x;
const uint8_t* src_v =
sample + src_width * ((ptrdiff_t)abs_src_height * 2 + crop_y) +
crop_x;
r = H444ToARGB(src_y, src_width, src_u, src_width, src_v, src_width,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
case FOURCC_U444: {
const uint8_t* src_y = sample + (ptrdiff_t)src_width * crop_y + crop_x;
const uint8_t* src_u =
sample + src_width * ((ptrdiff_t)abs_src_height + crop_y) + crop_x;
const uint8_t* src_v =
sample + src_width * ((ptrdiff_t)abs_src_height * 2 + crop_y) +
crop_x;
r = U444ToARGB(src_y, src_width, src_u, src_width, src_v, src_width,
dst_argb, dst_stride_argb, crop_width, inv_crop_height);
break;
}
#ifdef HAVE_JPEG
case FOURCC_MJPG:
r = MJPGToARGB(sample, sample_size, dst_argb, dst_stride_argb, src_width,
abs_src_height, crop_width, inv_crop_height);
break;
#endif
default:
r = -1; // unknown fourcc - return failure code.
}
if (need_buf) {
if (!r) {
r = ARGBRotate(dst_argb, dst_stride_argb, dest_argb, dest_dst_stride_argb,
crop_width, abs_crop_height, rotation);
}
free(rotate_buffer);
} else if (rotation) {
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = ARGBRotate(src, src_width * 4, dst_argb, dst_stride_argb, crop_width,
inv_crop_height, rotation);
}
return r;
}
#ifdef __cplusplus
} // extern "C"
} // namespace libyuv
#endif
Reference in New Issue View Git Blame Copy Permalink