[mp4 @ 00000229194b5d40] Application provided invalid, non monotonically increasing dts to muxer ...

[vasyl-zabolotniy]

This happens with specific file, but the thing is it worked earlier just fine on NVEncC 5.23 (I don't have NVidia GPU at the moment to check vs latest release). The video file and full console log are avail here:
https://drive.google.com/drive/folders/1K2dFIztS6Y1yFlJSt75BpeOjNxQijdvB?usp=sharing

[vasyl-zabolotniy]

fwiw, can confirm now that NVEncC (v5.23, v5.42) does not have this muxer error..

[unggulzmalqp]

Had similar problem. When the problematic file is opened using mp4box it shows something like "Missing REXT profile signaling, patching".

----------------------------- Demux audio -----------------------------

MP4Box 2.1-DEV-rev79-gdf29bc8a0-x64-gcc11.2.0 Patman

C:\apps\StaxRip-v2.11.0-x64\Apps\Support\MP4Box\MP4Box.exe -single 2 -out "C:\temp\ID1 SBR.m4a" "C:\video.mp4"

[isom/avcc] Missing REXT profile signaling, patching.

Tried add fps flag from rigaya/NVEnc#57, not working.

C:\apps\VCEEncC_6.17_x64\VCEEncC64.exe --avhw -i "C:\video.mp4" --video-streamid 1 --video-metadata clear --chapter-copy -c hevc --vbr 5000 --preset slow --tier high --colormatrix auto --transfer auto --colorprim auto --output-depth 8 --motion-est q-pel --chromaloc auto --colorrange auto --avsync cfr  --audio-copy 1    -m default_mode:infer_no_subs -o "C:\video-fastflix-01f0.mkv" --fps 30000/1001
--------------------------------------------------------------------------------------------------------------
C:\video-fastflix-01f0.mkv
--------------------------------------------------------------------------------------------------------------
storage->SetProperty(HevcColorBitDepth)=8 failed: invalid param..
VCEEnc (x64) 6.17 (r1002) by rigaya, Dec  6 2021 14:19:49 (VC 1929/Win)
OS:            Windows 11 x64 (22000) [UTF-8]
CPU:           AMD Ryzen 5 5500U with Radeon Graphics [4.05GHz] (6C/12T)
GPU:           AMD Radeon, AMF Runtime 1.4.18 / SDK 1.4.23
Input Info:    avvce: H.264/AVC, 1920x1080, 30/1 fps
Vpp Filters    copyDtoD
Output:        H.265/HEVC  main @ Level 4 (main tier)
               1920x1080p 1:1 30.000fps (30/1fps)
               avwriter: hevc, aac, chap => matroska
Quality:       slow
CQP:           I:28, P:28
VBV Bufsize:   12000 kbps
Bframes:       0 frames
Pre Analysis:  off
Ref frames:    2 frames
LTR frames:    0 frames
Motion Est:    Q-pel
Slices:        1
GOP Len:       300 frames
VUI:              chromaloc:left
Others:        deblock
avout: Error: Failed to write video frame: Invalid argument.
Error during output.
Error during output.
encoded 4 frames, 52.63 fps, 974.40 kbps, 0.02 MB
encode time 0:00:00, CPULoad: 15.2
frame type IDR 1
frame type I   1,  avgQP  28.00,  total size  0.02 MB
frame type P   3,  avgQP  28.00,  total size  0.00 MB
Finished with error in VCEEncC.
[matroska @ 00000246e5693ac0] Application provided invalid, non monotonically increasing dts to muxer in stream 0: 100 >= 33

[unggulzmalqp]

this is the trace log level, hope it helps

--------------------------------------------------------------------------------------------------------------
C:\video.mp4-fastflix-8082.mkv
--------------------------------------------------------------------------------------------------------------
VCEEnc    6.17 (x64)
OS        Windows 11 x64 (22000) [UTF-8]
CPU Info  AMD Ryzen 5 5500U with Radeon Graphics [4.06GHz] (6C/12T)
Locale    C
Loaded amfrt64.dll.
device #0: CreateContext() Success.
loaded OpenCL.
OpenCL platform count: 1
OpenCL platform #0[00007FFD74C2A490]: AMD Accelerated Parallel Processing Advanced Micro Devices, Inc. OpenCL 2.1 AMD-APP (3180.7)[FULL_PROFILE]
  extensions:cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
Add platform #0[00007FFD74C2A490] to list.
Created OpenCL platform list: 1
device #0: Checking platform AMD Accelerated Parallel Processing...
createDeviceListD3D11(d3d11dev = 000001C1F514B3D0)
D3D11 device count = 1
clGetDeviceIDsFromD3D11KHR: Success
create OpenCL Context for [000001C1FB565810]
Enable d3d11 interop for 000001C1F514B3D0
createQueue for device : 000001C1FB565810
timeBeginPeriod(1)
Set csp thread param: all (0x0xfff), priority=normal, throttling=auto.
avhw reader selected.
InitInput: input selected : 7.
avvce: select audio track #1, codec copy
avvce: opened file "C:\video.mp4".
avvce: got stream information.
avvce: found video stream, stream idx 0
avvce: found audio stream, stream idx 1, trackID 1.0, aac, frame_size 2048, timebase 1/44100, delay 0 ms
avvce: use video stream #0 for input, codec h264, stream time_base 1/15360, codec_timebase 1/60.
avvce: hdr10plusMetadataCopy: off
avvce: can be decoded by vce.
avvce: initialized h264_mp4toannexb filter.
avvce: start predecode.
avvce: GetHeader: changed 38 bytes -> 35 bytes by h264_mp4toannexb.
avvce: GetHeader: 35 bytes.
avvce: initialized h264 codec context for parser: time_base: 1/60, pkt_timebase: 1/15360.
avvce: fps decoder 30/1, invalid: false
avvce: maxCheckFrames 48, maxCheckSec: 1.000e+99
avvce: found first key frame: timestamp 127488 (8.3), offset 249
avvce: read 48 packets.
avvce: checking 31 frame samples.
avvce: stream timebase 1/60
avvce: decoder fps     30/1
avvce: duration histgram of 1 frames
avvce: 512 [ 31 frames]
avvce: using popular duration...
avvce: fps mul:         1001
avvce: raw avgDuration: 512.000000
avvce: estimatedAvgFps: 30030/1001
avvce: use decoder fps...
avvce: final AvgFps (raw64): 30030/1001
avvce: final AvgFps (gcd): 30/1
avvce: final AvgFps (round): 30/1
avvce: checking for stream #1
avvce: adjust trim by offset 249.
avvce: avvce: H.264/AVC, 1920x1080, 30/1 fps
avvce: streamFirstKeyPts: 127488
avvce: chromaloc:left
avvce: sar 1:1, bitdepth 8
Input: trim options
0-2147483647  (offset: 249)
vfr mode automatically enabled with timebase 1/240
Checking GPU #0 (AMD Radeon) for codec H.265/HEVC.
  acceleration: Hardware-accelerated.
  Max Profile: main.
  Max Level: 6.2.
GPU #0 (AMD Radeon) available for encode.
InitDevice: device #0 (AMD Radeon) selected.
decoder: use codec "AMFVideoDecoderUVD_H264_AVC".
created decoder context.
set codec header to decoder: 35 bytes.
initialize decoder: 1920x1080, NV12.
Initialized decoder
Load resource type: EXE_DATA, name: RGY_FILTER_CL
VCEEnc    6.17 (x64)
OS        Windows 11 x64 (22000) (x64)
CPU Info  AMD Ryzen 5 5500U with Radeon Graphics [4.06GHz] (6C/12T)
GPU Info  AMD Radeon
Loaded resource type: EXE_DATA, name: RGY_FILTER_CL, size = 31486
--------------------------------------------------------------------------
building OpenCL source: size 31483.
options: -D MEM_TYPE_SRC=2 -D MEM_TYPE_DST=2 -D in_bit_depth=8 -D out_bit_depth=8
source
// MEM_TYPE_SRC
// MEM_TYPE_DST
// in_bit_depth
// out_bit_depth
const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE | CLK_ADDRESS_NONE | CLK_FILTER_NEAREST;
#if in_bit_depth <= 8
#define TypeIn  uchar
#define TypeIn4 uchar4
#define convert_TypeIn4 convert_uchar4
#elif in_bit_depth <= 16
#define TypeIn  ushort
#define TypeIn4 ushort4
#define convert_TypeIn4 convert_ushort4
#endif
#if out_bit_depth <= 8
#define TypeOut  uchar
#define TypeOut4 uchar4
#define convert_TypeOut4 convert_uchar4
#elif out_bit_depth <= 16
#define TypeOut  ushort
#define TypeOut4 ushort4
#define convert_TypeOut4 convert_ushort4
#endif
// samplerでnormalizeした場合、0 -> 0.0f, 255 -> 1.0f
#define RGY_MEM_TYPE_CPU                    (0)
#define RGY_MEM_TYPE_GPU                    (1)
#define RGY_MEM_TYPE_GPU_IMAGE              (2)
#define RGY_MEM_TYPE_GPU_IMAGE_NORMALIZED   (3)
#ifndef __OPENCL_VERSION__
#define __kernel
#define __global
#define __local
#define __read_only
#define __write_only
#define image2d_t void*
#define uchar unsigned char
#endif
#define IntegerIsSigned(intType)    ((intType)(-1) < 0)
inline int conv_bit_depth_lsft(const int bit_depth_in, const int bit_depth_out, const int shift_offset) {
    const int lsft = bit_depth_out - (bit_depth_in + shift_offset);
    return lsft < 0 ? 0 : lsft;
}
inline int conv_bit_depth_rsft(const int bit_depth_in, const int bit_depth_out, const int shift_offset) {
    const int rsft = bit_depth_in + shift_offset - bit_depth_out;
    return rsft < 0 ? 0 : rsft;
}
inline int conv_bit_depth_rsft_add(const int bit_depth_in, const int bit_depth_out, const int shift_offset) {
    const int rsft = conv_bit_depth_rsft(bit_depth_in, bit_depth_out, shift_offset);
    return (rsft - 1 >= 0) ? 1 << (rsft - 1) : 0;
}
inline int conv_bit_depth(const int c, const int bit_depth_in, const int bit_depth_out, const int shift_offset) {
    if (bit_depth_out > bit_depth_in + shift_offset) {
        return c << conv_bit_depth_lsft(bit_depth_in, bit_depth_out, shift_offset);
    } else if (bit_depth_out < bit_depth_in + shift_offset) {
        const int x = (c + conv_bit_depth_rsft_add(bit_depth_in, bit_depth_out, shift_offset)) >> conv_bit_depth_rsft(bit_depth_in, bit_depth_out, shift_offset);
        const int low = 0;
        const int high = (1 << bit_depth_out) - 1;
        return (((x) <= (high)) ? (((x) >= (low)) ? (x) : (low)) : (high));
    } else {
        return c;
    }
}
#define NORM_SCALE_IN  (float)((1<<(sizeof(TypeIn)*8))-1)
#define NORM_SCALE_OUT (1.0f/(float)((1<<(sizeof(TypeOut)*8))-1))
#define BIT_DEPTH_CONV(x) (TypeOut)conv_bit_depth((x), in_bit_depth, out_bit_depth, 0)
#define BIT_DEPTH_CONV_FLOAT(x) (TypeOut)((out_bit_depth == in_bit_depth) \
    ? (x) \
    : ((out_bit_depth > in_bit_depth) \
        ? ((x) * (float)(1 << (out_bit_depth - in_bit_depth))) \
        : ((x) * (float)(1.0f / (1 << (in_bit_depth - out_bit_depth))))))
#define BIT_DEPTH_CONV_AVG(a, b) (TypeOut)conv_bit_depth((a)+(b), in_bit_depth, out_bit_depth, 1)
#define BIT_DEPTH_CONV_3x1_AVG(a, b) (TypeOut)conv_bit_depth(((a)<<1)+(a)+(b), in_bit_depth, out_bit_depth, 2)
#define BIT_DEPTH_CONV_7x1_AVG(a, b) (TypeOut)conv_bit_depth(((a)<<3)-(a)+(b), in_bit_depth, out_bit_depth, 3)
#define LOAD_IMG(src_img, ix, iy) (TypeIn)(read_imageui((src_img), sampler, (int2)((ix), (iy))).x)
#define LOAD_IMG_AYUV(src_img, ix, iy) convert_TypeIn4(read_imageui((src_img), sampler, (int2)((ix), (iy))))
#define LOAD_IMG_NV12_UV(src_img, src_u, src_v, ix, iy, cropX, cropY) { \
    uint4 ret = read_imageui((src_img), sampler, (int2)((ix) + ((cropX)>>1), (iy) + ((cropY)>>1))); \
    (src_u) = (TypeIn)ret.x; \
    (src_v) = (TypeIn)ret.y; \
}
#define LOAD_BUF(src_buf, ix, iy) *(__global TypeIn *)(&(src_buf)[(iy) * srcPitch + (ix) * sizeof(TypeIn)])
#define LOAD_BUF_AYUV(src_buf, ix, iy) *(__global TypeIn4 *)(&(src_buf)[(iy) * srcPitch + (ix) * sizeof(TypeIn4)])
#define LOAD_BUF_NV12_UV(src_buf, src_u, src_v, ix, iy, cropX, cropY) { \
    (src_u) = LOAD((src_buf), ((ix)<<1) + 0 + (cropX), (iy) + ((cropY)>>1)); \
    (src_v) = LOAD((src_buf), ((ix)<<1) + 1 + (cropX), (iy) + ((cropY)>>1)); \
}
#define LOAD_IMG_NORM(src_img, ix, iy) (TypeIn)(read_imagef((src_img), sampler, (int2)((ix), (iy))).x * NORM_SCALE_IN + 0.5f)
#define LOAD_IMG_NORM_AYUV(src_img, ix, iy) convert_TypeIn4(read_imagef((src_img), sampler, (int2)((ix), (iy))) * (float4)NORM_SCALE_IN + (float4)0.5f)
#define LOAD_IMG_NORM_NV12_UV(src_img, src_u, src_v, ix, iy, cropX, cropY) { \
    float4 ret = read_imagef((src_img), sampler, (int2)((ix) + ((cropX)>>1), (iy) + ((cropY)>>1))); \
    (src_u) = (TypeIn)(ret.x * NORM_SCALE_IN + 0.5f); \
    (src_v) = (TypeIn)(ret.y * NORM_SCALE_IN + 0.5f); \
}
#if MEM_TYPE_SRC == RGY_MEM_TYPE_GPU_IMAGE
#define IMAGE_SRC    1
#define LOAD         LOAD_IMG
#define LOAD_AYUV    LOAD_IMG_AYUV
#define LOAD_NV12_UV LOAD_IMG_NV12_UV
#elif MEM_TYPE_SRC == RGY_MEM_TYPE_GPU_IMAGE_NORMALIZED
#define IMAGE_SRC    1
#define LOAD         LOAD_IMG_NORM
#define LOAD_AYUV    LOAD_IMG_NORM_AYUV
#define LOAD_NV12_UV LOAD_IMG_NORM_NV12_UV
#else
#define IMAGE_SRC    0
#define LOAD         LOAD_BUF
#define LOAD_AYUV    LOAD_BUF_AYUV
#define LOAD_NV12_UV LOAD_BUF_NV12_UV
#endif
#define STORE_IMG(dst_img, ix, iy, val) write_imageui((dst_img), (int2)((ix), (iy)), (val))
#define STORE_IMG_AYUV(dst_img, ix, iy, val) write_imageui((dst_img), (int2)((ix), (iy)), convert_uint4(val))
#define STORE_IMG_NV12_UV(dst_img, ix, iy, val_u, val_v) { \
    uint4 val = (uint4)(val_u, val_v, val_v, val_v); \
    write_imageui((dst_img), (int2)((ix), (iy)), (val)); \
}
#define STORE_IMG_NORM(dst_img, ix, iy, val) write_imagef(dst_img, (int2)((ix), (iy)), (val * NORM_SCALE_OUT))
#define STORE_IMG_NORM_AYUV(dst_img, ix, iy, val) write_imagef(dst_img, (int2)((ix), (iy)), (convert_float4(val) * (float4)NORM_SCALE_OUT))
#define STORE_IMG_NORM_NV12_UV(dst_img, ix, iy, val_u, val_v) { \
    float4 val = (float4)(val_u * NORM_SCALE_OUT, val_v * NORM_SCALE_OUT, val_v * NORM_SCALE_OUT, val_v * NORM_SCALE_OUT); \
    write_imagef(dst_img, (int2)((ix), (iy)), (val)); \
}
#define STORE_BUF(dst_buf, ix, iy, val)  { \
    __global TypeOut *ptr = (__global TypeOut *)(&(dst_buf)[(iy) * dstPitch + (ix) * sizeof(TypeOut)]); \
    ptr[0] = (TypeOut)(val); \
}
#define STORE_BUF_AYUV(dst_buf, ix, iy, val)  { \
    __global TypeOut4 *ptr = (__global TypeOut4 *)(&(dst_buf)[(iy) * dstPitch + (ix) * sizeof(TypeOut4)]); \
    ptr[0] = (TypeOut4)(val); \
}
#define STORE_BUF_NV12_UV(dst_buf, ix, iy, val_u, val_v) { \
    STORE(dst_buf, ((ix) << 1) + 0, (iy), val_u); \
    STORE(dst_buf, ((ix) << 1) + 1, (iy), val_v); \
}
#if MEM_TYPE_DST == RGY_MEM_TYPE_GPU_IMAGE
#define IMAGE_DST     1
#define STORE         STORE_IMG
#define STORE_AYUV    STORE_IMG_AYUV
#define STORE_NV12_UV STORE_IMG_NV12_UV
#elif MEM_TYPE_DST == RGY_MEM_TYPE_GPU_IMAGE_NORMALIZED
#define IMAGE_DST     1
#define STORE         STORE_IMG_NORM
#define STORE_AYUV    STORE_IMG_NORM_AYUV
#define STORE_NV12_UV STORE_IMG_NORM_NV12_UV
#else
#define IMAGE_DST     0
#define STORE         STORE_BUF
#define STORE_AYUV    STORE_BUF_AYUV
#define STORE_NV12_UV STORE_BUF_NV12_UV
#endif
void conv_c_yuv420_yuv444_internal(
    int *pixDst11, int *pixDst12,
    int *pixDst21, int *pixDst22,
    int pixSrc01, int pixSrc02,
    int pixSrc11, int pixSrc12,
    int pixSrc21, int pixSrc22
) {
    pixSrc02 = (pixSrc01 + pixSrc02 + 1) >> 1;
    pixSrc12 = (pixSrc11 + pixSrc12 + 1) >> 1;
    pixSrc22 = (pixSrc21 + pixSrc22 + 1) >> 1;
    *pixDst11 = BIT_DEPTH_CONV_3x1_AVG(pixSrc11, pixSrc01);
    *pixDst12 = BIT_DEPTH_CONV_3x1_AVG(pixSrc12, pixSrc02);
    *pixDst21 = BIT_DEPTH_CONV_7x1_AVG(pixSrc11, pixSrc21);
    *pixDst22 = BIT_DEPTH_CONV_7x1_AVG(pixSrc12, pixSrc22);
}
void conv_c_yuv420_yuv444(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    const int dstPitch,
    const int dst_x, const int dst_y,
    int pixSrc01, int pixSrc02,
    int pixSrc11, int pixSrc12,
    int pixSrc21, int pixSrc22
) {
    int pixDst11, pixDst12, pixDst21, pixDst22;
    conv_c_yuv420_yuv444_internal(
        &pixDst11, &pixDst12, &pixDst21, &pixDst22,
        pixSrc01, pixSrc02, pixSrc11, pixSrc12, pixSrc21, pixSrc22
    );
    STORE(dst, dst_x+0, dst_y+0, (TypeOut)pixDst11);
    STORE(dst, dst_x+1, dst_y+0, (TypeOut)pixDst12);
    STORE(dst, dst_x+0, dst_y+1, (TypeOut)pixDst21);
    STORE(dst, dst_x+1, dst_y+1, (TypeOut)pixDst22);
}
__kernel void kernel_copy_plane(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstOffsetX,
    int dstOffsetY,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcOffsetX,
    int srcOffsetY,
    int width,
    int height
) {
    const int x = get_global_id(0);
    const int y = get_global_id(1);
    if (x < width && y < height) {
        TypeIn pixSrc = LOAD(src, x + srcOffsetX, y + srcOffsetY);
        TypeOut out = BIT_DEPTH_CONV(pixSrc);
        STORE(dst, x + dstOffsetX, y + dstOffsetY, out);
    }
}
__kernel void kernel_copy_plane_nv12(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int uvWidth,
    int uvHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int uv_x = get_global_id(0);
    const int uv_y = get_global_id(1);
    if (uv_x < uvWidth && uv_y < uvHeight) {
        TypeIn pixSrcU, pixSrcV;
        LOAD_NV12_UV(src, pixSrcU, pixSrcV, uv_x, uv_y, cropX, cropY);
        TypeOut pixDstU = BIT_DEPTH_CONV(pixSrcU);
        TypeOut pixDstV = BIT_DEPTH_CONV(pixSrcV);
        STORE_NV12_UV(dst, uv_x, uv_y, pixDstU, pixDstV);
    }
}
__kernel void kernel_crop_nv12_yv12(
#if IMAGE_DST
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitch,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int uvWidth,
    int uvHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int uv_x = get_global_id(0);
    const int uv_y = get_global_id(1);
    if (uv_x < uvWidth && uv_y < uvHeight) {
        TypeIn pixSrcU, pixSrcV;
        LOAD_NV12_UV(src, pixSrcU, pixSrcV, uv_x, uv_y, cropX, cropY);
        TypeOut pixDstU = BIT_DEPTH_CONV(pixSrcU);
        TypeOut pixDstV = BIT_DEPTH_CONV(pixSrcV);
        STORE(dstU, uv_x, uv_y, pixDstU);
        STORE(dstV, uv_x, uv_y, pixDstV);
    }
}
__kernel void kernel_crop_nv12_yuv444(
#if IMAGE_DST
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int src_x = get_global_id(0);
    const int src_y = get_global_id(1);
    const int dst_x = src_x << 1;
    const int dst_y = src_y << 1;
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int loadx = src_x + (cropX>>1);
        const int loady = src_y + (cropY>>1);
        TypeIn pixSrcU01, pixSrcV01, pixSrcU02, pixSrcV02;
        TypeIn pixSrcU11, pixSrcV11, pixSrcU12, pixSrcV12;
        TypeIn pixSrcU21, pixSrcV21, pixSrcU22, pixSrcV22;
        LOAD_NV12_UV(src, pixSrcU01, pixSrcV01,     loadx,                max(loady-1, 0),           0, 0);
        LOAD_NV12_UV(src, pixSrcU02, pixSrcV02, min(loadx+1, srcWidth-1), max(loady-1, 0),           0, 0);
        LOAD_NV12_UV(src, pixSrcU11, pixSrcV11,     loadx,                    loady,                 0, 0);
        LOAD_NV12_UV(src, pixSrcU12, pixSrcV12, min(loadx+1, srcWidth-1),     loady,                 0, 0);
        LOAD_NV12_UV(src, pixSrcU21, pixSrcV21,     loadx,                min(loady+1, srcHeight-1), 0, 0);
        LOAD_NV12_UV(src, pixSrcU22, pixSrcV22, min(loadx+1, srcWidth-1), min(loady+1, srcHeight-1), 0, 0);
        conv_c_yuv420_yuv444(dstU, dstPitch, dst_x, dst_y, pixSrcU01, pixSrcU02, pixSrcU11, pixSrcU12, pixSrcU21, pixSrcU22);
        conv_c_yuv420_yuv444(dstV, dstPitch, dst_x, dst_y, pixSrcV01, pixSrcV02, pixSrcV11, pixSrcV12, pixSrcV21, pixSrcV22);
    }
}
__kernel void kernel_crop_c_yuv444_nv12(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t srcU,
    __read_only image2d_t srcV,
#else
    __global uchar *srcU,
    __global uchar *srcV,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int dst_x = get_global_id(0);
    const int dst_y = get_global_id(1);
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int src_x = dst_x << 1;
        const int src_y = dst_y << 1;
        const int loadx = src_x + cropX;
        const int loady = src_y + cropY;
        const int pixSrcU00 = LOAD(srcU, loadx+0, loady+0);
        const int pixSrcU10 = LOAD(srcU, loadx+0, loady+1);
        const int pixSrcV00 = LOAD(srcV, loadx+0, loady+0);
        const int pixSrcV10 = LOAD(srcV, loadx+0, loady+1);
        TypeOut pixDstU = BIT_DEPTH_CONV_AVG(pixSrcU00, pixSrcU10);
        TypeOut pixDstV = BIT_DEPTH_CONV_AVG(pixSrcV00, pixSrcV10);
        STORE_NV12_UV(dst, dst_x, dst_y, pixDstU, pixDstV);
    }
}
__kernel void kernel_crop_yv12_nv12(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
#if IMAGE_SRC
    __read_only image2d_t srcU,
    __read_only image2d_t srcV,
#else
    __global uchar *srcU,
    __global uchar *srcV,
#endif
    int srcPitch,
    int uvWidth,
    int uvHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int uv_x = get_global_id(0);
    const int uv_y = get_global_id(1);
    if (uv_x < uvWidth && uv_y < uvHeight) {
        const TypeIn pixSrcU = LOAD(srcU, uv_x + (cropX>>1), uv_y + (cropY>>1));
        const TypeIn pixSrcV = LOAD(srcV, uv_x + (cropX>>1), uv_y + (cropY>>1));
        const TypeOut pixDstU = BIT_DEPTH_CONV(pixSrcU);
        const TypeOut pixDstV = BIT_DEPTH_CONV(pixSrcV);
        STORE_NV12_UV(dst, uv_x, uv_y, pixDstU, pixDstV);
    }
}
__kernel void kernel_crop_c_yv12_yuv444(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int src_x = get_global_id(0);
    const int src_y = get_global_id(1);
    const int dst_x = src_x << 1;
    const int dst_y = src_y << 1;
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int loadx = src_x + (cropX>>1);
        const int loady = src_y + (cropY>>1);
        const int pixSrc01 = LOAD(src,     loadx,                max(loady-1, 0)          );
        const int pixSrc02 = LOAD(src, min(loadx+1, srcWidth-1), max(loady-1, 0)          );
        const int pixSrc11 = LOAD(src,     loadx,                    loady                );
        const int pixSrc12 = LOAD(src, min(loadx+1, srcWidth-1),     loady                );
        const int pixSrc21 = LOAD(src,     loadx,                min(loady+1, srcHeight-1));
        const int pixSrc22 = LOAD(src, min(loadx+1, srcWidth-1), min(loady+1, srcHeight-1));
        conv_c_yuv420_yuv444(dst, dstPitch, dst_x, dst_y, pixSrc01, pixSrc02, pixSrc11, pixSrc12, pixSrc21, pixSrc22);
    }
}
__kernel void kernel_crop_c_yuv444_yv12(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,     // 輝度と同じcropを想定
    int cropY      // 輝度と同じcropを想定
) {
    const int dst_x = get_global_id(0);
    const int dst_y = get_global_id(1);
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int src_x = dst_x << 1;
        const int src_y = dst_y << 1;
        const int loadx = src_x + cropX;
        const int loady = src_y + cropY;
        const int pixSrc00 = LOAD(src, loadx+0, loady+0);
        const int pixSrc10 = LOAD(src, loadx+0, loady+1);
        const TypeOut pixDst = BIT_DEPTH_CONV_AVG(pixSrc00, pixSrc10);
        STORE(dst, dst_x, dst_y, pixDst);
    }
}
TypeOut4 conv_rgb_yuv(const TypeIn4 rgb) {
    const float fr = (float)rgb.z;
    const float fg = (float)rgb.y;
    const float fb = (float)rgb.x;
    const float fy =  0.257f * fr + 0.504f * fg + 0.098f * fb +  16.0f;
    const float fu = -0.148f * fr - 0.291f * fg + 0.439f * fb + 128.0f;
    const float fv =  0.439f * fr - 0.368f * fg - 0.071f * fb + 128.0f;
    TypeOut4 yuv;
    yuv.x = (TypeOut)(BIT_DEPTH_CONV_FLOAT(fy) + 0.5f);
    yuv.y = (TypeOut)(BIT_DEPTH_CONV_FLOAT(fu) + 0.5f);
    yuv.z = (TypeOut)(BIT_DEPTH_CONV_FLOAT(fv) + 0.5f);
    return yuv;
}
__kernel void kernel_crop_rgb32_yuv444(
#if IMAGE_DST
    __write_only image2d_t dstY,
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstY,
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dst_x = get_global_id(0);
    const int dst_y = get_global_id(1);
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int loadx = dst_x + cropX;
        const int loady = dst_y + cropY;
        const TypeIn4 pixSrcRGB = LOAD_AYUV(src, loadx, loady);
        TypeOut4 yuv = conv_rgb_yuv(pixSrcRGB);
        STORE(dstY, dst_x, dst_y, yuv.x);
        STORE(dstU, dst_x, dst_y, yuv.y);
        STORE(dstV, dst_x, dst_y, yuv.z);
    }
}
__kernel void kernel_crop_rgb32_yv12(
#if IMAGE_DST
    __write_only image2d_t dstY,
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstY,
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitchY,
    int dstPitchU,
    int dstPitchV,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dstC_x = get_global_id(0);
    const int dstC_y = get_global_id(1);
    const int dstY_x = dstC_x << 1;
    const int dstY_y = dstC_y << 1;
    if (dstY_x + 1 < dstWidth && dstY_y + 1 < dstHeight) {
        const int loadx = dstY_x + cropX;
        const int loady = dstY_y + cropY;
        const TypeIn4 pixSrcRGB00 = LOAD_AYUV(src, loadx+0, loady+0);
        const TypeIn4 pixSrcRGB01 = LOAD_AYUV(src, loadx+1, loady+0);
        const TypeIn4 pixSrcRGB10 = LOAD_AYUV(src, loadx+0, loady+1);
        const TypeIn4 pixSrcRGB11 = LOAD_AYUV(src, loadx+1, loady+1);
        const TypeOut4 yuv00 = conv_rgb_yuv(pixSrcRGB00);
        const TypeOut4 yuv01 = conv_rgb_yuv(pixSrcRGB01);
        const TypeOut4 yuv10 = conv_rgb_yuv(pixSrcRGB10);
        const TypeOut4 yuv11 = conv_rgb_yuv(pixSrcRGB11);
        int dstPitch = dstPitchY;
        STORE(dstY, dstY_x+0, dstY_y+0, yuv00.x);
        STORE(dstY, dstY_x+1, dstY_y+0, yuv01.x);
        STORE(dstY, dstY_x+0, dstY_y+1, yuv10.x);
        STORE(dstY, dstY_x+1, dstY_y+1, yuv11.x);
        const TypeOut pixU = (TypeOut)(((int)yuv00.y + (int)yuv10.y + 1) >> 1);
        const TypeOut pixV = (TypeOut)(((int)yuv00.z + (int)yuv10.z + 1) >> 1);
        dstPitch = dstPitchU;
        STORE(dstU, dstC_x, dstC_y, pixU);
        dstPitch = dstPitchV;
        STORE(dstV, dstC_x, dstC_y, pixV);
    }
}
__kernel void kernel_crop_rgb32_nv12(
#if IMAGE_DST
    __write_only image2d_t dstY,
    __write_only image2d_t dstC,
#else
    __global uchar *dstY,
    __global uchar *dstC,
#endif
    int dstPitchY,
    int dstPitchC,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dstC_x = get_global_id(0);
    const int dstC_y = get_global_id(1);
    const int dstY_x = dstC_x << 1;
    const int dstY_y = dstC_y << 1;
    if (dstY_x < dstWidth && dstY_y < dstHeight) {
        const int loadx = dstY_x + cropX;
        const int loady = dstY_y + cropY;
        const TypeIn4 pixSrcRGB00 = LOAD_AYUV(src, loadx+0, loady+0);
        const TypeIn4 pixSrcRGB01 = LOAD_AYUV(src, loadx+1, loady+0);
        const TypeIn4 pixSrcRGB10 = LOAD_AYUV(src, loadx+0, loady+1);
        const TypeIn4 pixSrcRGB11 = LOAD_AYUV(src, loadx+1, loady+1);
        const TypeOut4 yuv00 = conv_rgb_yuv(pixSrcRGB00);
        const TypeOut4 yuv01 = conv_rgb_yuv(pixSrcRGB01);
        const TypeOut4 yuv10 = conv_rgb_yuv(pixSrcRGB10);
        const TypeOut4 yuv11 = conv_rgb_yuv(pixSrcRGB11);
        int dstPitch = dstPitchY;
        STORE(dstY, dstY_x+0, dstY_y+0, yuv00.x);
        STORE(dstY, dstY_x+1, dstY_y+0, yuv01.x);
        STORE(dstY, dstY_x+0, dstY_y+1, yuv10.x);
        STORE(dstY, dstY_x+1, dstY_y+1, yuv11.x);
        const TypeOut pixU = (TypeOut)(((int)yuv00.y + (int)yuv10.y + 1) >> 1);
        const TypeOut pixV = (TypeOut)(((int)yuv00.z + (int)yuv10.z + 1) >> 1);
        dstPitch = dstPitchC;
        STORE_NV12_UV(dstC, dstC_x, dstC_y, pixU, pixV);
    }
}
__kernel void kernel_crop_ayuv_yuv444(
#if IMAGE_DST
    __write_only image2d_t dstY,
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstY,
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dst_x = get_global_id(0);
    const int dst_y = get_global_id(1);
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int loadx = dst_x + cropX;
        const int loady = dst_y + cropY;
        TypeIn4 pix = LOAD_AYUV(src, loadx, loady); //RGBA = VUYA
        TypeOut pixY = (TypeOut)BIT_DEPTH_CONV(pix.z);
        TypeOut pixU = (TypeOut)BIT_DEPTH_CONV(pix.y);
        TypeOut pixV = (TypeOut)BIT_DEPTH_CONV(pix.x);
        STORE(dstY, dst_x, dst_y, pixY);
        STORE(dstU, dst_x, dst_y, pixU);
        STORE(dstV, dst_x, dst_y, pixV);
    }
}
__kernel void kernel_crop_ayuv_yv12(
#if IMAGE_DST
    __write_only image2d_t dstY,
    __write_only image2d_t dstU,
    __write_only image2d_t dstV,
#else
    __global uchar *dstY,
    __global uchar *dstU,
    __global uchar *dstV,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t src,
#else
    __global uchar *src,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dst_x_C = get_global_id(0);
    const int dst_y_C = get_global_id(1);
    const int dst_x_Y = dst_x_C << 1;
    const int dst_y_Y = dst_y_C << 1;
    if (dst_x_Y < dstWidth && dst_y_Y < dstHeight) {
        const int src_x = dst_x_Y;
        const int src_y = dst_y_Y;
        const int loadx = src_x + cropX;
        const int loady = src_y + cropY;
        const TypeIn4 pixSrc00 = LOAD_AYUV(src, loadx+0, loady+0);
        const TypeIn4 pixSrc01 = LOAD_AYUV(src, loadx+1, loady+0);
        const TypeIn4 pixSrc10 = LOAD_AYUV(src, loadx+0, loady+1);
        const TypeIn4 pixSrc11 = LOAD_AYUV(src, loadx+1, loady+1);
        TypeOut pixY00 = (TypeOut)BIT_DEPTH_CONV(pixSrc00.z);
        TypeOut pixY01 = (TypeOut)BIT_DEPTH_CONV(pixSrc01.z);
        TypeOut pixY10 = (TypeOut)BIT_DEPTH_CONV(pixSrc10.z);
        TypeOut pixY11 = (TypeOut)BIT_DEPTH_CONV(pixSrc11.z);
        TypeOut pixU   = (TypeOut)BIT_DEPTH_CONV_AVG(pixSrc00.y, pixSrc10.y);
        TypeOut pixV   = (TypeOut)BIT_DEPTH_CONV_AVG(pixSrc00.x, pixSrc10.x);
        STORE(dstY, dst_x_Y+0, dst_y_Y+0, pixY00);
        STORE(dstY, dst_x_Y+1, dst_y_Y+0, pixY01);
        STORE(dstY, dst_x_Y+0, dst_y_Y+1, pixY10);
        STORE(dstY, dst_x_Y+1, dst_y_Y+1, pixY11);
        STORE(dstU, dst_x_C,   dst_y_C,   pixU);
        STORE(dstV, dst_x_C,   dst_y_C,   pixV);
    }
}
__kernel void kernel_crop_yuv444_ayuv(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t srcY,
    __read_only image2d_t srcU,
    __read_only image2d_t srcV,
#else
    __global uchar *srcY,
    __global uchar *srcU,
    __global uchar *srcV,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int dst_x = get_global_id(0);
    const int dst_y = get_global_id(1);
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int loadx = dst_x + cropX;
        const int loady = dst_y + cropY;
        const int pixY = LOAD(srcY, loadx, loady);
        const int pixU = LOAD(srcU, loadx, loady);
        const int pixV = LOAD(srcV, loadx, loady);
        TypeOut4 pix;
        pix.w = 0;
        pix.z = BIT_DEPTH_CONV(pixY);
        pix.y = BIT_DEPTH_CONV(pixU);
        pix.x = BIT_DEPTH_CONV(pixV);
        STORE_AYUV(dst, dst_x, dst_y, pix);
    }
}
__kernel void kernel_crop_yv12_ayuv(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int dstWidth,
    int dstHeight,
#if IMAGE_SRC
    __read_only image2d_t srcY,
    __read_only image2d_t srcU,
    __read_only image2d_t srcV,
#else
    __global uchar *srcY,
    __global uchar *srcU,
    __global uchar *srcV,
#endif
    int srcPitch,
    int srcWidth,
    int srcHeight,
    int cropX,
    int cropY
) {
    const int src_C_x = get_global_id(0);
    const int src_C_y = get_global_id(1);
    const int dst_x = src_C_x << 1;
    const int dst_y = src_C_y << 1;
    if (dst_x < dstWidth && dst_y < dstHeight) {
        const int load_C_x = src_C_x + (cropX>>1);
        const int load_C_y = src_C_y + (cropY>>1);
        const int pixSrcU01 = LOAD(srcU,     load_C_x,                max(load_C_y-1, 0)          );
        const int pixSrcU02 = LOAD(srcU, min(load_C_x+1, srcWidth-1), max(load_C_y-1, 0)          );
        const int pixSrcU11 = LOAD(srcU,     load_C_x,                    load_C_y                );
        const int pixSrcU12 = LOAD(srcU, min(load_C_x+1, srcWidth-1),     load_C_y                );
        const int pixSrcU21 = LOAD(srcU,     load_C_x,                min(load_C_y+1, srcHeight-1));
        const int pixSrcU22 = LOAD(srcU, min(load_C_x+1, srcWidth-1), min(load_C_y+1, srcHeight-1));
        const int pixSrcV01 = LOAD(srcV,     load_C_x,                max(load_C_y-1, 0)          );
        const int pixSrcV02 = LOAD(srcV, min(load_C_x+1, srcWidth-1), max(load_C_y-1, 0)          );
        const int pixSrcV11 = LOAD(srcV,     load_C_x,                    load_C_y                );
        const int pixSrcV12 = LOAD(srcV, min(load_C_x+1, srcWidth-1),     load_C_y                );
        const int pixSrcV21 = LOAD(srcV,     load_C_x,                min(load_C_y+1, srcHeight-1));
        const int pixSrcV22 = LOAD(srcV, min(load_C_x+1, srcWidth-1), min(load_C_y+1, srcHeight-1));
        int pixDstU11, pixDstU12, pixDstU21, pixDstU22;
        conv_c_yuv420_yuv444_internal(&pixDstU11, &pixDstU12, &pixDstU21, &pixDstU22, pixSrcU01, pixSrcU02, pixSrcU11, pixSrcU12, pixSrcU21, pixSrcU22);
        int pixDstV11, pixDstV12, pixDstV21, pixDstV22;
        conv_c_yuv420_yuv444_internal(&pixDstU11, &pixDstU12, &pixDstU21, &pixDstU22, pixSrcU01, pixSrcU02, pixSrcU11, pixSrcU12, pixSrcU21, pixSrcU22);
        const int load_Y_x = load_C_x << 1;
        const int load_Y_y = load_C_y << 1;
        const int pixSrcY11 = LOAD(srcY, load_Y_x+0, load_Y_y+0);
        const int pixSrcY12 = LOAD(srcY, load_Y_x+1, load_Y_y+0);
        const int pixSrcY21 = LOAD(srcY, load_Y_x+0, load_Y_y+1);
        const int pixSrcY22 = LOAD(srcY, load_Y_x+1, load_Y_y+1);
        TypeOut4 pix11, pix12, pix21, pix22;
        pix11.w = 0;
        pix11.z = BIT_DEPTH_CONV(pixSrcY11);
        pix11.y = pixDstU11;
        pix11.x = pixDstV11;
        pix12.w = 0;
        pix12.z = BIT_DEPTH_CONV(pixSrcY12);
        pix12.y = pixDstU12;
        pix12.x = pixDstV12;
        pix21.w = 0;
        pix21.z = BIT_DEPTH_CONV(pixSrcY21);
        pix21.y = pixDstU21;
        pix21.x = pixDstV21;
        pix22.w = 0;
        pix22.z = BIT_DEPTH_CONV(pixSrcY22);
        pix22.y = pixDstU22;
        pix22.x = pixDstV22;
        STORE_AYUV(dst, dst_x+0, dst_y+0, pix11);
        STORE_AYUV(dst, dst_x+1, dst_y+0, pix12);
        STORE_AYUV(dst, dst_x+0, dst_y+1, pix21);
        STORE_AYUV(dst, dst_x+1, dst_y+1, pix22);
    }
}
__kernel void kernel_separate_fields(
    __global uchar *dst0,
    __global uchar *dst1,
    int dstPitch,
    __global uchar *src,
    int srcPitch,
    int width,
    int height_field
) {
    const int x = get_global_id(0);
    const int y = get_global_id(1);
    if (x < width && y < height_field) {
        TypeIn pixSrc0 = LOAD_BUF(src, x, y*2+0);
        TypeIn pixSrc1 = LOAD_BUF(src, x, y*2+1);
        STORE_BUF(dst0, x, y, pixSrc0);
        STORE_BUF(dst1, x, y, pixSrc1);
    }
}
__kernel void kernel_merge_fields(
    __global uchar *dst,
    int dstPitch,
    __global uchar *src0,
    __global uchar *src1,
    int srcPitch,
    int width,
    int height_field
) {
    const int x = get_global_id(0);
    const int y = get_global_id(1);
    if (x < width && y < height_field) {
        TypeIn pixSrc0 = LOAD_BUF(src0, x, y);
        TypeIn pixSrc1 = LOAD_BUF(src1, x, y);
        STORE_BUF(dst, x, y*2+0, pixSrc0);
        STORE_BUF(dst, x, y*2+1, pixSrc1);
    }
}
__kernel void kernel_set_plane(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int width,
    int height,
    int cropX,
    int cropY,
    int value
) {
    const int x = get_global_id(0);
    const int y = get_global_id(1);
    if (x < width && y < height) {
        STORE(dst, x + cropX, y + cropY, value);
    }
}
__kernel void kernel_set_plane_nv12(
#if IMAGE_DST
    __write_only image2d_t dst,
#else
    __global uchar *dst,
#endif
    int dstPitch,
    int uvWidth,
    int uvHeight,
    int cropX,
    int cropY,
    int valueU,
    int valueV
) {
    const int uv_x = get_global_id(0);
    const int uv_y = get_global_id(1);
    if (uv_x < uvWidth && uv_y < uvHeight) {
        TypeOut pixDstU = valueU;
        TypeOut pixDstV = valueV;
        STORE_NV12_UV(dst, uv_x + cropX, uv_y + cropY, pixDstU, pixDstV);
    }
}
--------------------------------------------------------------------------
initialized Encoder component.
Max Profile: main.
Max Level: 1831027128.
pushed usage params.
storage->SetProperty(HevcColorBitDepth)=8 failed: invalid param..
pushed static params.
pushed dynamic params.
initalized encoder.
selected mode on : score 2: resolution 2, speed limit off.
perf monitor: Performace Monitor: none
perf monitor: Performace Plot   : none
Output: Using avformat writer.
Output: Audio/Subtitle muxing enabled.
Output: CopyAll=false
Output: Added audio track#4097 (stream idx 1) for mux, bitrate 0, codec: copy  , bsf: <none>, disposition: <copy>, metadata <copy>
avout: output filename: "C:\video.mp4-fastflix-8082.mkv"
avout: allocated internal buffer 8 MB.
perf monitor: Set perf monitor thread param all (0x0xfff), priority=background, throttling=auto.
avout: set external output buffer 8 MB.
avout: output video stream fps: 30/1
avout: opened video avcodec
avout: Set Video language: key language, value und
avout: Copy Video Metadata: key language, value und
avout: Copy Video Metadata: key handler_name, value VideoHandler
avout: start initialize hevc_metadata filter...
avout: set video_format 5 by hevc_metadata filter
avout: set colorprim 2 by hevc_metadata filter
avout: set transfer 2 by hevc_metadata filter
avout: set matrix 2 by hevc_metadata filter
avout: set chromaloc 0 by hevc_metadata filter
avout: initialized hevc_metadata filter
avout: output video stream timebase: 1/1000
avout: bDtsUnavailable: off
avout: Initialized video output.
avout: start initializing audio ouput...
avout: output stream index 1, trackId 1.0
avout: samplerate 44100, stream pkt_timebase 1/44100
avout: set extradata from original packet...
avout: Copy Disposition: default
avout: Set Audio #4097.0 language: key language, value und
avout: Copy Audio #4097.0 Metadata: key language, value und
avout: Copy Audio #4097.0 Metadata: key handler_name, value Sound Media Handler
avout: Initialized audio output - #0: track 1, substream 0.
avout: Copy Container Metadata: key major_brand, value isom
avout: Copy Container Metadata: key minor_version, value 512
avout: Copy Container Metadata: key compatible_brands, value isomiso2avc1mp41
avout: Copy Container Metadata: key encoder, value Lavf58.29.100
avout: set mux opt: default_mode = infer_no_subs.
avout: avwriter: hevc, aac => matroska
avout: starting output thread...
avout: Set output thread param: all (0x0xfff), priority=normal, throttling=on.
Output: Initialized avformat writer.
Set main thread param: all (0x0xfff), priority=normal, throttling=auto.
VCEEnc (x64) 6.17 (r1002) by rigaya, Dec  6 2021 14:19:49 (VC 1929/Win)
OS:            Windows 11 x64 (22000) [UTF-8]
CPU:           AMD Ryzen 5 5500U with Radeon Graphics [4.06GHz] (6C/12T)
GPU:           AMD Radeon, AMF Runtime 1.4.18 / SDK 1.4.23
Input Info:    avvce: H.264/AVC, 1920x1080, 30/1 fps
Vpp Filters    copyDtoD
Output:        H.265/HEVC  main @ Level 4 (high tier)
               1920x1080p 1:1 30.000fps (30/1fps)
               avwriter: hevc, aac => matroska
Quality:       slow
CQP:           I:28, P:28
VBV Bufsize:   30000 kbps
Bframes:       0 frames
Pre Analysis:  off
Ref frames:    2 frames
LTR frames:    0 frames
Motion Est:    Q-pel
Slices:        1
GOP Len:       300 frames
VUI:              chromaloc:left
Others:        deblock
Started Decode thread.
Started Output thread.
avout: calc dts, first dts 0 x (timebase).
avout: videoDts=       0: 0.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: audioDts=       0: 0, maxDst=   12000.
avout: videoDts=    6030: 0.067.
avout: audioDts=       0: 0, maxDst=   18030.
avout: videoDts=    9000: 0.1.
avout: Error: Failed to write video frame: Invalid argument.
avout: videoDts=    2970: 0.033.
avout: audioDts=       0: 0, maxDst=   14970.
Error during output.
Error during output.
Flushed Encoder
Waiting for ouput thread to be finieshed...
Closed output thread.
avout: Closing...
avout: Flushed audio buffer.
avout: audioDts=9223372036854775807: 1.02482e+14, maxDst=   14970.
avout: closed output thread...
avout: closed queues...
avout: Closed avformat context.
avout: Closed File Pointer.
avout: Closed format.
avout: Closed audio.
avout: Closed video context.
avout: Closed video.
avout: Closed.
avvce: Closing...

avvce: Closed Stream Packet Buffer.
avvce: Closed caption handler.
avvce: Closing avformat context...
avvce: Closed avformat context.
avvce: Free formatOptions...
avvce: Freed formatOptions.
avvce: Closed format.
avvce: Close parser...
avvce: Closed parser.
avvce: Close codecCtx for parser...
avvce: Closed codecCtx for parser.
avvce: Free bsf...
avvce: Freed bsf.
avvce: Free extra data...
avvce: Freed extra data.
avvce: Closed video.
avvce: Closing Stream #0...
avvce: Free packet sample...
avvce: Freed packet sample.
avvce: Closed Stream #0.
avvce: Cleared frame pos list.
avvce: Closed.
                                                                               
encoded 4 frames, 41.67 fps, 974.40 kbps, 0.02 MB
encode time 0:00:00, CPULoad: 10.7
frame type IDR 1
frame type I   1,  avgQP  28.00,  total size  0.02 MB
frame type P   3,  avgQP  28.00,  total size  0.00 MB
timeEndPeriod(1)
Pipeline Stopped.
Closing Encoder...
Closed Encoder.
Closing Decoder...
Closed Decoder.
Closing CL Context...
--------------------------------------------------------------------------
build log of gfx902...

--------------------------------------------------------------------------
clBuildProgram success!
clReleaseProgram...
clReleaseProgram: fin.
Closed CL m_copy program.
Closed CL Queue.
Closed CL Context.
Closed CL Platform.
devDX11: Closing DX11 device...
devDX11: Closing DX11 device...
avout: Closing...
avout: closed queues...
avout: Closed format.
avout: Closed video.
avout: Closed.
avvce: Closing...
avvce: Closed Stream Packet Buffer.
avvce: Closed caption handler.
avvce: Closed format.
avvce: Closed video.
avvce: Cleared frame pos list.
avvce: Closed.
avvce: Closing...
avvce: Close...
Closing EncodeStatus...
perf monitor: Closing thread...
perf monitor: Closed thread.
perf monitor: Closing perf counter...
perf monitor: Closed perf counter.
Closing logger...
Finished with error in VCEEncC.
[matroska @ 000001c180e93080] Application provided invalid, non monotonically increasing dts to muxer in stream 0: 100 >= 33