Create d3d11Texture2D with initial data from camera - c++

Question

I am trying to read input from camera and use it to create background surface in D3D11. I receive memory errors all the time.
My render Target size is: 2364 * 1461
Image I get from the camera is an array of type unsigned char
unsigned char* p = g_pOvrvision->GetCamImage(OVR::OV_CAMEYE_LEFT, (OVR::OvPSQuality)processer_quality);
It returns 640 * 480 * 3 bytes. The code I am working with is below. CreateTexture2D gives a memory error. I tried filling the array with dummy data to fit all 2364 * 1461. This did not work either.
Could you please suggest me solution?
D3D11_TEXTURE2D_DESC desc;
ZeroMemory(&desc, sizeof(desc));
desc.Width = renderTargetSize.w;
desc.Height = renderTargetSize.h;
desc.MipLevels = desc.ArraySize = 1;
desc.Format = DXGI_FORMAT_R8G8B8A8_UNORM;
desc.SampleDesc.Count = 1;
desc.SampleDesc.Quality = 0;
desc.Usage = D3D11_USAGE_DYNAMIC;
desc.BindFlags = D3D11_BIND_SHADER_RESOURCE;
desc.CPUAccessFlags = D3D11_CPU_ACCESS_WRITE;
desc.MiscFlags = 0;
ID3D11Texture2D *pTexture = NULL;
D3D11_SUBRESOURCE_DATA TexInitData;
ZeroMemory(&TexInitData, sizeof(D3D11_SUBRESOURCE_DATA));
TexInitData.pSysMem = texArray;
TexInitData.SysMemPitch = static_cast<UINT>(2364 * 3);
TexInitData.SysMemSlicePitch = static_cast<UINT>(3 * 2364 * 1461 * sizeof(unsigned char));
d3dDevice->CreateTexture2D(&desc, &TexInitData, &d3dEyeTexture);
d3dDevice->CreateShaderResourceView(d3dEyeTexture, nullptr, &d3dEyeTextureShaderResourceView);
d3dDevice->CreateRenderTargetView(d3dEyeTexture, nullptr, &d3dEyeTextureRenderTargetView);

Answer 1

Chuck Walbourn's comment gets you 99% of the way there: there is no three-byte DXGI pixel format, so to properly pack the data you are receiving from the camera you must create a new buffer that will match the DXGI_FORMAT_R8G8B8A8_UNORM format, and copy your camera data to it adding a fourth value (an alpha value) to each pixel of 0xFF.
But as you mentioned your last line of code still fails. I believe this is because you haven't set d3dEyeTexture with the bind flag D3D11_BIND_RENDER_TARGET as well as D3D11_BIND_SHADER_RESOURCE. From what I understand, however, if you're trying to make the camera image of 640x480 the background of your 2364x1461 render target, you'll need to resample the data to fit, which is a pain to do (and slow) on the CPU. My recommendation would be to create a separate, D3D11_BIND_SHADER_RESOURCE flagged, 640x480 texture with your camera data, and draw it to fill the whole screen as the first step in your draw loop. This means creating a separate, dedicated 2364x1461 render target as your back buffer, but that's a common first step in setting up a Direct3D app.