The Windows Imaging Component (WIC) has a function for creating an IWICBitmap from a GDI HBITMAP. But how do you go the other way?

One way is to lock the bitmap bits and then use SetDIBits to place them into a DIB section. We saw this trick when we were debugging code that converts a Direct2D bitmap to a GDI HBITMAP. The downside of this is that it may result in a double-copy if the IWICBitmapSource has to generate the pixels on the fly (due to being the end of a chain of transforms).

Instead, we can ask CopyPixels to generate the pixels directly into our DIB section.

HRESULT ConvertWICBitmapSourceToHBITMAP(
    IWICBitmapSource* source,
    HBITMAP* bitmap)
{
    *bitmap = nullptr;

    // Convert to 32bpp if necessary.
    wil::com_ptr<IWICBitmapSource> source32bpp;
    RETURN_IF_FAILED(
        WICConvertBitmapSource(GUID_WICPixelFormat32bppBGR, source,
            source32bpp.put()));

    // Create a 32bpp DIB section of matching size
    UINT width, height;
    RETURN_IF_FAILED(source32bpp->GetSize(&width, &height));

    // Make sure the size won't create integer overflow.
    RETURN_HR_IF(HRESULT_FROM_WIN32(ERROR_ARITHMETIC_OVERFLOW),
                 width > (MAXDWORD / 4 / height));

    BITMAPINFO bi{};
    bi.bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
    bi.bmiHeader.biWidth = width;
    bi.bmiHeader.biHeight = -static_cast<LONG>(height); // top-down
    bi.bmiHeader.biPlanes = 1;
    bi.bmiHeader.biBitCount = 32;
    bi.bmiHeader.biCompression = BI_RGB;
    bi.bmiHeader.biSizeImage = width * height * 4;

    void* bits;
    wil::unique_hbitmap result(CreateDIBSection(nullptr, &bi,
            DIB_RGB_COLORS, &bits, nullptr, 0));
    RETURN_LAST_ERROR_IF_NULL(result);

    // Copy the pixels into the bitmap                            
    RETURN_IF_FAILED(source32bpp->CopyPixels(nullptr, width * 4,  
        bi.bmiHeader.biSizeImage, reinterpret_cast<BYTE*>(bits)));

    // All done
    *bitmap = result.release();
    return S_OK;
}

We convert the WIC bitmap to 32-bit BGR format, since that’s the format that GDI uses for 32-bit DIBs. Then we create a 32-bit DIB section with the same dimensions as the WIC bitmap and use CopyPixels to tell the WIC bitmap to generate pixels into the DIB section buffer.

This code always creates a 32-bit DIB section. You could extend the function to query the WIC bitmap for its pixel format and create a matching GDI bitmap, but in practice almost nobody uses any GDI color formats besides 32-bit DIB and 32-bit DIB with premultiplied alpha. (32-bit DIB with straight alpha is also popular, but that’s not a GDI color format.)