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_i420.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

620 lines
23 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.h"
#include <limits.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include "libyuv/cpu_id.h"
#include "libyuv/planar_functions.h"
#include "libyuv/row.h"
#include "libyuv/video_common.h"
#ifdef __cplusplus
namespace libyuv {
extern "C" {
#endif
static void HalfRow_C(const uint8_t* src_uv,
ptrdiff_t src_uv_stride,
uint8_t* dst_uv,
int width) {
int x;
for (x = 0; x < width; ++x) {
dst_uv[x] = (src_uv[x] + src_uv[src_uv_stride + x] + 1) >> 1;
}
}
static int NV16ToI420(const uint8_t* src_y,
int src_stride_y,
const uint8_t* src_uv,
int src_stride_uv,
uint8_t* dst_y,
int dst_stride_y,
uint8_t* dst_u,
int dst_stride_u,
uint8_t* dst_v,
int dst_stride_v,
int width,
int height) {
int y;
int halfwidth = (width + 1) >> 1;
int halfheight = (height + 1) >> 1;
// UV rows are halfwidth * 2 (even) bytes.
int awidth = halfwidth * 2;
void (*SplitUVRow)(const uint8_t* src_uv, uint8_t* dst_u, uint8_t* dst_v,
int width) = SplitUVRow_C;
void (*InterpolateRow)(uint8_t* dst_ptr, const uint8_t* src_ptr,
ptrdiff_t src_stride, int dst_width,
int source_y_fraction) = InterpolateRow_C;
if ((!src_y && dst_y) || !src_uv || !dst_u || !dst_v || width <= 0 ||
height == 0 || height == INT_MIN) {
return -1;
}
if (height < 0) {
height = -height;
halfheight = (height + 1) >> 1;
src_y = src_y + (ptrdiff_t)(height - 1) * src_stride_y;
src_uv = src_uv + (ptrdiff_t)(height - 1) * src_stride_uv;
src_stride_y = -src_stride_y;
src_stride_uv = -src_stride_uv;
}
if (dst_y) {
CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height);
}
#if defined(HAS_SPLITUVROW_SSE2)
if (TestCpuFlag(kCpuHasSSE2)) {
SplitUVRow = SplitUVRow_Any_SSE2;
if (IS_ALIGNED(halfwidth, 16)) {
SplitUVRow = SplitUVRow_SSE2;
}
}
#endif
#if defined(HAS_SPLITUVROW_AVX2)
if (TestCpuFlag(kCpuHasAVX2)) {
SplitUVRow = SplitUVRow_Any_AVX2;
if (IS_ALIGNED(halfwidth, 32)) {
SplitUVRow = SplitUVRow_AVX2;
}
}
#endif
#if defined(HAS_SPLITUVROW_AVX512BW)
if (TestCpuFlag(kCpuHasAVX512BW)) {
SplitUVRow = SplitUVRow_Any_AVX512BW;
if (IS_ALIGNED(halfwidth, 64)) {
SplitUVRow = SplitUVRow_AVX512BW;
}
}
#endif
#if defined(HAS_SPLITUVROW_NEON)
if (TestCpuFlag(kCpuHasNEON)) {
SplitUVRow = SplitUVRow_Any_NEON;
if (IS_ALIGNED(halfwidth, 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(halfwidth, 16)) {
SplitUVRow = SplitUVRow_LSX;
}
}
#endif
#if defined(HAS_SPLITUVROW_LASX)
if (TestCpuFlag(kCpuHasLASX)) {
SplitUVRow = SplitUVRow_Any_LASX;
if (IS_ALIGNED(halfwidth, 32)) {
SplitUVRow = SplitUVRow_LASX;
}
}
#endif
#if defined(HAS_SPLITUVROW_RVV)
if (TestCpuFlag(kCpuHasRVV)) {
SplitUVRow = SplitUVRow_RVV;
}
#endif
#if defined(HAS_INTERPOLATEROW_AVX2)
if (TestCpuFlag(kCpuHasAVX2)) {
InterpolateRow = InterpolateRow_Any_AVX2;
if (IS_ALIGNED(awidth, 32)) {
InterpolateRow = InterpolateRow_AVX2;
}
}
#endif
#if defined(HAS_INTERPOLATEROW_AVX512BW)
if (TestCpuFlag(kCpuHasAVX512BW)) {
InterpolateRow = InterpolateRow_AVX512BW;
}
#endif
#if defined(HAS_INTERPOLATEROW_NEON)
if (TestCpuFlag(kCpuHasNEON)) {
InterpolateRow = InterpolateRow_Any_NEON;
if (IS_ALIGNED(awidth, 16)) {
InterpolateRow = InterpolateRow_NEON;
}
}
#endif
#if defined(HAS_INTERPOLATEROW_SVE2)
if (TestCpuFlag(kCpuHasSVE2)) {
InterpolateRow = InterpolateRow_SVE2;
}
#endif
#if defined(HAS_INTERPOLATEROW_SME)
if (TestCpuFlag(kCpuHasSME)) {
InterpolateRow = InterpolateRow_SME;
}
#endif
#if defined(HAS_INTERPOLATEROW_LSX)
if (TestCpuFlag(kCpuHasLSX)) {
InterpolateRow = InterpolateRow_Any_LSX;
if (IS_ALIGNED(awidth, 16)) {
InterpolateRow = InterpolateRow_LSX;
}
}
#endif
#if defined(HAS_INTERPOLATEROW_LASX)
if (TestCpuFlag(kCpuHasLASX)) {
InterpolateRow = InterpolateRow_Any_LASX;
if (IS_ALIGNED(awidth, 32)) {
InterpolateRow = InterpolateRow_LASX;
}
}
#endif
#if defined(HAS_INTERPOLATEROW_RVV)
if (TestCpuFlag(kCpuHasRVV)) {
InterpolateRow = InterpolateRow_RVV;
}
#endif
#if defined(HAS_INTERPOLATEROW_WASMSIMD)
if (TestCpuFlag(kCpuHasWASMSIMD)) {
InterpolateRow = InterpolateRow_Any_WASMSIMD;
if (IS_ALIGNED(awidth, 16)) {
InterpolateRow = InterpolateRow_WASMSIMD;
}
}
#endif
align_buffer_64(row_uv, awidth);
if (!row_uv) {
return 1;
}
for (y = 0; y < halfheight; ++y) {
if (y * 2 + 1 < height) {
// Average 2 rows of UV with rounding (source_y_fraction 128).
InterpolateRow(row_uv, src_uv, src_stride_uv, awidth, 128);
SplitUVRow(row_uv, dst_u, dst_v, halfwidth);
} else {
SplitUVRow(src_uv, dst_u, dst_v, halfwidth);
}
src_uv += (ptrdiff_t)src_stride_uv * 2;
dst_u += dst_stride_u;
dst_v += dst_stride_v;
}
free_aligned_buffer_64(row_uv);
return 0;
}
static int NV24ToI420(const uint8_t* src_y,
int src_stride_y,
const uint8_t* src_uv,
int src_stride_uv,
uint8_t* dst_y,
int dst_stride_y,
uint8_t* dst_u,
int dst_stride_u,
uint8_t* dst_v,
int dst_stride_v,
int width,
int height) {
int x, y;
int halfwidth = (width + 1) >> 1;
int halfheight = (height + 1) >> 1;
if ((!src_y && dst_y) || !src_uv || !dst_u || !dst_v || width <= 0 ||
height == 0 || height == INT_MIN) {
return -1;
}
if (height < 0) {
height = -height;
halfheight = (height + 1) >> 1;
src_y = src_y + (ptrdiff_t)(height - 1) * src_stride_y;
src_uv = src_uv + (ptrdiff_t)(height - 1) * src_stride_uv;
src_stride_y = -src_stride_y;
src_stride_uv = -src_stride_uv;
}
if (dst_y) {
CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height);
}
align_buffer_64(row_uv, width * 2);
if (!row_uv) {
return 1;
}
for (y = 0; y < halfheight; ++y) {
if (y * 2 + 1 < height) {
HalfRow_C(src_uv, src_stride_uv, row_uv, width * 2);
} else {
memcpy(row_uv, src_uv, (size_t)width * 2);
}
for (x = 0; x < halfwidth - 1; ++x) {
dst_u[x] = (row_uv[x * 4 + 0] + row_uv[x * 4 + 2] + 1) >> 1;
dst_v[x] = (row_uv[x * 4 + 1] + row_uv[x * 4 + 3] + 1) >> 1;
}
if (width & 1) {
dst_u[halfwidth - 1] = row_uv[(halfwidth - 1) * 4 + 0];
dst_v[halfwidth - 1] = row_uv[(halfwidth - 1) * 4 + 1];
} else {
dst_u[halfwidth - 1] = (row_uv[(halfwidth - 1) * 4 + 0] +
row_uv[(halfwidth - 1) * 4 + 2] + 1) >>
1;
dst_v[halfwidth - 1] = (row_uv[(halfwidth - 1) * 4 + 1] +
row_uv[(halfwidth - 1) * 4 + 3] + 1) >>
1;
}
src_uv += (ptrdiff_t)src_stride_uv * 2;
dst_u += dst_stride_u;
dst_v += dst_stride_v;
}
free_aligned_buffer_64(row_uv);
return 0;
}
// Convert camera sample to I420 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.
LIBYUV_API
int ConvertToI420(const uint8_t* sample,
size_t sample_size,
uint8_t* dst_y,
int dst_stride_y,
uint8_t* dst_u,
int dst_stride_u,
uint8_t* dst_v,
int dst_stride_v,
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;
}
const int abs_src_height = (src_height < 0) ? -src_height : src_height;
const int abs_crop_height = (crop_height < 0) ? -crop_height : crop_height;
if (!dst_y || !dst_u || !dst_v || !sample || src_width <= 0 ||
src_width > INT_MAX / 4 || crop_width <= 0 || 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);
const uint8_t* src;
const uint8_t* src_uv;
int r = 0;
int need_buf = (rotation && format != FOURCC_I420 && format != FOURCC_NV12 &&
format != FOURCC_NV21 && format != FOURCC_YV12) ||
dst_y == sample;
uint8_t* tmp_y = dst_y;
uint8_t* tmp_u = dst_u;
uint8_t* tmp_v = dst_v;
int tmp_y_stride = dst_stride_y;
int tmp_u_stride = dst_stride_u;
int tmp_v_stride = dst_stride_v;
uint8_t* rotate_buffer = NULL;
const int inv_crop_height =
(src_height < 0) ? -abs_crop_height : abs_crop_height;
int aligned_src_width = (src_width + 1) & ~1;
// One pass rotation is available for some formats. For the rest, convert
// to I420 (with optional vertical flipping) into a temporary I420 buffer,
// and then rotate the I420 to the final destination buffer.
// For in-place conversion, if destination dst_y is same as source sample,
// also enable temporary buffer.
if (need_buf) {
size_t y_size = (size_t)crop_width * abs_crop_height;
size_t uv_size =
(size_t)((crop_width + 1) / 2) * ((abs_crop_height + 1) / 2);
if (uv_size > SIZE_MAX / 2 || y_size > SIZE_MAX - uv_size * 2) {
return -1; // Invalid size.
}
const size_t rotate_buffer_size = y_size + uv_size * 2;
rotate_buffer = (uint8_t*)malloc(rotate_buffer_size);
if (!rotate_buffer) {
return 1; // Out of memory runtime error.
}
dst_y = rotate_buffer;
dst_u = dst_y + y_size;
dst_v = dst_u + uv_size;
dst_stride_y = crop_width;
dst_stride_u = dst_stride_v = ((crop_width + 1) / 2);
}
switch (format) {
// Single plane formats
case FOURCC_YUY2: { // TODO(fbarchard): Find better odd crop fix.
uint8_t* u = (crop_x & 1) ? dst_v : dst_u;
uint8_t* v = (crop_x & 1) ? dst_u : dst_v;
int stride_u = (crop_x & 1) ? dst_stride_v : dst_stride_u;
int stride_v = (crop_x & 1) ? dst_stride_u : dst_stride_v;
src = sample + ((ptrdiff_t)aligned_src_width * crop_y + crop_x) * 2;
r = YUY2ToI420(src, aligned_src_width * 2, dst_y, dst_stride_y, u,
stride_u, v, stride_v, crop_width, inv_crop_height);
break;
}
case FOURCC_UYVY: {
uint8_t* u = (crop_x & 1) ? dst_v : dst_u;
uint8_t* v = (crop_x & 1) ? dst_u : dst_v;
int stride_u = (crop_x & 1) ? dst_stride_v : dst_stride_u;
int stride_v = (crop_x & 1) ? dst_stride_u : dst_stride_v;
src = sample + ((ptrdiff_t)aligned_src_width * crop_y + crop_x) * 2;
r = UYVYToI420(src, aligned_src_width * 2, dst_y, dst_stride_y, u,
stride_u, v, stride_v, crop_width, inv_crop_height);
break;
}
case FOURCC_RGBP:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = RGB565ToI420(src, src_width * 2, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_RGBO:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = ARGB1555ToI420(src, src_width * 2, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_R444:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 2;
r = ARGB4444ToI420(src, src_width * 2, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_24BG:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 3;
r = RGB24ToI420(src, src_width * 3, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_RAW:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 3;
r = RAWToI420(src, src_width * 3, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_ARGB:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = ARGBToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_BGRA:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = BGRAToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_ABGR:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = ABGRToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_RGBA:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = RGBAToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_AR30:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = AR30ToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_AB30:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x) * 4;
r = AB30ToI420(src, src_width * 4, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, crop_width,
inv_crop_height);
break;
case FOURCC_I400:
src = sample + (ptrdiff_t)src_width * crop_y + crop_x;
r = I400ToI420(src, src_width, dst_y, dst_stride_y, dst_u, dst_stride_u,
dst_v, dst_stride_v, crop_width, inv_crop_height);
break;
// Biplanar formats
case FOURCC_NV12:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample + ((ptrdiff_t)src_width * abs_src_height) +
((ptrdiff_t)(crop_y / 2) * aligned_src_width) +
((crop_x / 2) * 2);
r = NV12ToI420Rotate(src, src_width, src_uv, aligned_src_width, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v,
dst_stride_v, crop_width, inv_crop_height, rotation);
break;
case FOURCC_NV21:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample + ((ptrdiff_t)src_width * abs_src_height) +
((ptrdiff_t)(crop_y / 2) * aligned_src_width) +
((crop_x / 2) * 2);
// Call NV12 but with dst_u and dst_v parameters swapped.
r = NV12ToI420Rotate(src, src_width, src_uv, aligned_src_width, dst_y,
dst_stride_y, dst_v, dst_stride_v, dst_u,
dst_stride_u, crop_width, inv_crop_height, rotation);
break;
case FOURCC_NV16:
src = sample + ((ptrdiff_t)src_width * crop_y + crop_x);
src_uv = sample + ((ptrdiff_t)aligned_src_width * abs_src_height) +
((ptrdiff_t)crop_y * aligned_src_width) + ((crop_x / 2) * 2);
r = NV16ToI420(src, src_width, src_uv, aligned_src_width, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v, dst_stride_v,
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)crop_y * aligned_src_width * 2) + (crop_x * 2);
r = NV24ToI420(src, src_width, src_uv, aligned_src_width * 2, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v, dst_stride_v,
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 = P010ToI420(src_y16, src_width, src_uv16, aligned_src_width, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v, dst_stride_v,
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 = P210ToI420(src_y16, src_width, src_uv16, aligned_src_width, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v, dst_stride_v,
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 = P410ToI420(src_y16, src_width, src_uv16, src_width * 2, dst_y,
dst_stride_y, dst_u, dst_stride_u, dst_v, dst_stride_v,
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 = I420Rotate(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_y, dst_stride_y, dst_u, dst_stride_u, dst_v,
dst_stride_v, crop_width, inv_crop_height, rotation);
break;
}
case FOURCC_I422:
case FOURCC_YV16: {
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;
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 = I422ToI420(src_y, src_width, src_u, halfwidth, src_v, halfwidth,
dst_y, dst_stride_y, dst_u, dst_stride_u, dst_v,
dst_stride_v, 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 = I444ToI420(src_y, src_width, src_u, src_width, src_v, src_width,
dst_y, dst_stride_y, dst_u, dst_stride_u, dst_v,
dst_stride_v, crop_width, inv_crop_height);
break;
}
#ifdef HAVE_JPEG
case FOURCC_MJPG:
r = MJPGToI420(sample, sample_size, dst_y, dst_stride_y, dst_u,
dst_stride_u, dst_v, dst_stride_v, 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 = I420Rotate(dst_y, dst_stride_y, dst_u, dst_stride_u, dst_v,
dst_stride_v, tmp_y, tmp_y_stride, tmp_u, tmp_u_stride,
tmp_v, tmp_v_stride, crop_width, abs_crop_height,
rotation);
}
free(rotate_buffer);
}
return r;
}
#ifdef __cplusplus
} // extern "C"
} // namespace libyuv
#endif
Reference in New Issue View Git Blame Copy Permalink