ue4 Convert audio from 48 stereo to 16 mono - audio

Question

How to change sampling rate from 48000 (ue4 default) to 16000 samples and stereo to mono in a wav recording in ue4? I have searched in BPs but not lack. The image below shows what I have done with BPs. In order for this to work I had to change in WindowsEngine.ini the audio settings to XAudio (see this: https://www.youtube.com/watch?v=BpP1SxxwYIE)
Therefore I assume that this should be only possible with C++.

Answer 1

I did it with C++
File -> New C++ class -> VoiceCharacter -> Public
Change your character to have "VoiceCharacter" as parent. It can be found in Class settings of your character BP.
Add this method to VoiceCharacter C++ class and built and play.
void AVoiceCharacter::StereoToMono(TArray<uint8> stereoWavBytes, TArray<uint8>& monoWavBytes)
{
if(stereoWavBytes.Num() == 0)
{
GEngine->AddOnScreenDebugMessage(-1, 5.f, FColor::Red, "Stereo Bytes is empty");
return;
}
//Change wav headers
for (int i = 0; i < 44; i++)
{
//NumChannels starts from 22 to 24
if (i == 22)
{
short originalChannels = (*(short*)&stereoWavBytes[i]);
short NumChannels = originalChannels / 2;
FString message = FString::FromInt(originalChannels);
GEngine->AddOnScreenDebugMessage(-1, 25.f, FColor::Red, message);
monoWavBytes.Append((uint8*)&NumChannels, sizeof(NumChannels));
i++;
}//SamplingRate starts from 24 to 27
else if (i == 24)
{
int OriginalSamplingRate = (*(int*)&stereoWavBytes[i]);
int SamplingRate = OriginalSamplingRate / 3 ;
GEngine->AddOnScreenDebugMessage(-1, 25.f, FColor::Yellow, FString::FromInt(OriginalSamplingRate));
monoWavBytes.Append((uint8*)&SamplingRate, sizeof(SamplingRate));
i += 3;
} //ByteRate starts from 28 to 32
else if (i == 28)
{
int OriginalByteRate = (*(int*)&stereoWavBytes[i]);
int ByteRate = OriginalByteRate / 6 ;
GEngine->AddOnScreenDebugMessage(-1, 25.f, FColor::Yellow, FString::FromInt(OriginalByteRate));
monoWavBytes.Append((uint8*)&ByteRate, sizeof(ByteRate));
i += 3;
}
//BlockAlign starts from 32 to 34
else if (i == 32)
{
short BlockAlign = (*(short*)&stereoWavBytes[i]) / 2;
GEngine->AddOnScreenDebugMessage(-1, 25.f, FColor::White, FString::FromInt(BlockAlign));
monoWavBytes.Append((uint8*)&BlockAlign, sizeof(BlockAlign));
i++;
}
//SubChunkSize starts from 40 to 44
else if (i == 40)
{
int SubChunkSize = (*(int*)&stereoWavBytes[i]) / 2;
GEngine->AddOnScreenDebugMessage(-1, 25.f, FColor::Green, FString::FromInt(SubChunkSize));
monoWavBytes.Append((uint8*)&SubChunkSize, sizeof(SubChunkSize));
i += 3;
}
else
{
monoWavBytes.Add(stereoWavBytes[i]);
}
}
//Copies only the left channel and ignores the right channel
// for (int i = 44; i < stereoWavBytes.Num(); i += 4)
// {
// monoWavBytes.Add(stereoWavBytes[i]);
// monoWavBytes.Add(stereoWavBytes[i+1]);
// }
//Copies only the left channel and ignores the right channel. Also downsamples by 3,
// i.e. converts Windows 48000 sampling rate of ue4 to 16000
for (int i = 44; i < stereoWavBytes.Num(); i += 12)
{
monoWavBytes.Add(stereoWavBytes[i]);
monoWavBytes.Add(stereoWavBytes[i+1]);
}
}
Do not forget to add this in VoiceCharacter.h so that you can use it BPs.
UFUNCTION(BlueprintCallable, Category="Audio")
static void StereoToMono(TArray<uint8> stereoWavBytes, TArray<uint8>& monoWavBytes);
and here is how use it in BPs
Some useful plugins for (CUFile)
https://github.com/getnamo/nodejs-ue4
https://github.com/getnamo/socketio-client-ue4