#include "UnityCG.cginc"

RWStructuredBuffer<uint> _Vectorscope;
Texture2D<float4> _Source;

CBUFFER_START (Params)
    uint _IsLinear;
    float4 _Res;
CBUFFER_END

#define GROUP_SIZE 32

float3 RgbToYUV(float3 c)
{
    float Y =  0.299 * c.r + 0.587 * c.g + 0.114 * c.b;
    float U = -0.169 * c.r - 0.331 * c.g + 0.500 * c.b;
    float V =  0.500 * c.r - 0.419 * c.g - 0.081 * c.b;
    return float3(Y, U, V);
}

#pragma kernel KVectorscope
[numthreads(GROUP_SIZE,GROUP_SIZE,1)]
void KVectorscope(uint2 dispatchThreadId : SV_DispatchThreadID)
{
    if (dispatchThreadId.x < (uint)_Res.x && dispatchThreadId.y < (uint)_Res.y)
    {
        float3 color = saturate(_Source[dispatchThreadId].xyz);
        if (_IsLinear > 0)
            color = LinearToGammaSpace(color);

        float3 yuv = RgbToYUV(color);

        if (length(yuv.yz) > 0.49)
            yuv.yz = normalize(yuv.yz) * 0.49;

        yuv.yz += (0.5).xx;
        uint u = (uint)floor(yuv.y * _Res.x);
        uint v = (uint)floor(yuv.z * _Res.y);
        InterlockedAdd(_Vectorscope[v * _Res.x + u], 1);
    }
}

#pragma kernel KVectorscopeClear
[numthreads(GROUP_SIZE,GROUP_SIZE,1)]
void KVectorscopeClear(uint2 dispatchThreadId : SV_DispatchThreadID)
{
    if (dispatchThreadId.x < (uint)_Res.x && dispatchThreadId.y < (uint)_Res.y)
        _Vectorscope[dispatchThreadId.y * _Res.x + dispatchThreadId.x] = 0u;
}