Don't leave the negative frequencies as 0 for inverse FFT

1 files changed, 7 insertions, 2 deletions

diff --git a/alc/effects/convolution.cpp b/alc/effects/convolution.cpp
index 102bed59..bbf8cb3f 100644
--- a/alc/effects/convolution.cpp
+++ b/alc/effects/convolution.cpp
@@ -266,8 +266,7 @@ void ConvolutionState::update(const ALCcontext *context, const ALeffectslot *slo
     /* The iFFT'd response is scaled up by the number of bins, so apply the
      * inverse to the output mixing gain.
      */
-    constexpr size_t m{ConvolveUpdateSize/2 + 1};
-    const float gain{slot->Params.Gain * (1.0f/m)};
+    const float gain{slot->Params.Gain * (1.0f/float{ConvolveUpdateSize})};
     auto &chans = *mChans;
     if(mChannels == FmtBFormat3D || mChannels == FmtBFormat2D)
     {
@@ -386,6 +385,12 @@ void ConvolutionState::process(const size_t samplesToDo,
                     mFftBuffer[i] += *input * *filter;
             }
 
+            /* Reconstruct the mirrored/negative frequencies to do a proper
+             * inverse FFT.
+             */
+            for(size_t i{m};i < ConvolveUpdateSize;++i)
+                mFftBuffer[i] = std::conj(mFftBuffer[ConvolveUpdateSize-i]);
+
             /* Apply iFFT to get the 1024 (really 1023) samples for output. The
              * 512 output samples are combined with the last output's 511
              * second-half samples (and this output's second half is