I am trying to minimize the delay on Android.
I checked Android console and saw that it is trying to resample 44100 source to 48000. So, I am going to convert my wav files into 48000.
BUT, is that enough? For example, it is mono as I do not need stereo and stereo takes more disk space. Shall I convert it to stereo to prevent it from entering to the resampling phase?
Resampling sound data to match the output rate does not reduce latency at all, so please don’t do that.
If you do determine that the output rate of the device is different to the mixer’s rate, then that is a step you can take to reduce latency. By calling System::getDriverInfo you can determine the device’s rate, then System::setSoftwareFormat can be called to make the 2 match. This removes a double buffered resampler from the final conversion step, which saves one block of latency.
If fmod and the hardware share the same rate then you don’t gain anything.