EFB2RAM: Apply copy filter as a float coefficient after sampling

Using 8-bit integer math here lead to precision loss for depth copies,
which broke various effects in games, e.g. lens flare in MK:DD.

It's unlikely the console implements this as a floating-point multiply
(fixed-point perhaps), but since we have the float round trip in our
EFB2RAM shaders anyway, it's not going to make things any worse. If we
do rewrite our shaders to use integer math completely, then it might be
worth switching this conversion back to integers.

However, the range of the values (format) should be known, or we should
expand all values out to 24-bits first.

1 files changed, 2 insertions, 2 deletions

diff --git a/Source/Core/VideoBackends/D3D/TextureCache.cpp b/Source/Core/VideoBackends/D3D/TextureCache.cpp
index e34858d5a4..41ad435817 100644
--- a/Source/Core/VideoBackends/D3D/TextureCache.cpp
+++ b/Source/Core/VideoBackends/D3D/TextureCache.cpp
@@ -276,7 +276,7 @@ void TextureCache::CopyEFBToCacheEntry(TCacheEntry* entry, bool is_depth_copy,
   };
   PixelConstants constants;
   for (size_t i = 0; i < filter_coefficients.size(); i++)
-    constants.filter_coefficients[i] = filter_coefficients[i] / 64.0f;
+    constants.filter_coefficients[i] = filter_coefficients[i];
   constants.gamma_rcp = 1.0f / gamma;
   constants.clamp_top = clamp_top ? src_rect.top / float(EFB_HEIGHT) : 0.0f;
   constants.clamp_bottom = clamp_bottom ? src_rect.bottom / float(EFB_HEIGHT) : 1.0f;
@@ -315,4 +315,4 @@ TextureCache::GetEFBToTexPixelShader(const TextureConversionShaderGen::TCShaderU
   m_efb_to_tex_pixel_shaders.emplace(uid, shader);
   return shader;
 }
-}
+}  // namespace DX11