How to Combine Buffers for Simultaneous Sound?

Tracktion Engine
1

Hello! I’m attempting to combine two buffers in within the context of the SamplerPlugin. What I want to do is combine the samples for SampledNotes with the same note value so I can process their buffers per-sound. The end result should be one buffer with the sound of the two clips playing simultaneously.

Here I am sorting through a list of notes to find SamplerNotes that match the currentNote that I, processing. I get the buffer with sn->addNectBlock and then add it in to sampleAccumulator.

AudioSampleBuffer sampleAccumulator(fc.destBuffer->getNumChannels(), fc.destBuffer->getNumSamples());
                sampleAccumulator.clear();

  // Test to see if other sampled notes share a note property
  for (int innerNote = 0; innerNote < playingNotes.size(); innerNote++)
  {
    if (playingNotes.getUnchecked(innerNote)->note == currentNote)
    {
      AudioSampleBuffer tempSampleBuffer(fc.destBuffer->getNumChannels(), fc.destBuffer->getNumSamples());
      tempSampleBuffer.clear();

      auto sn = playingNotes.getUnchecked(innerNote);

      // Build the audio buffer containing the sampler note's data
      sn->addNextBlock(tempSampleBuffer, fc.bufferStartSample, fc.bufferNumSamples);

      // Add it to the sampleAccumulator
      for (int channel = 0; channel < fc.destBuffer->getNumChannels(); channel++)
      {
        sampleAccumulator.addFrom(channel, fc.bufferStartSample, tempSampleBuffer, channel, fc.bufferStartSample, fc.bufferNumSamples);
      }

      for (int i = 0; i < sampleAccumulator.getNumSamples(); i++)
      {
        DBG("(" << i << ", " << sampleAccumulator.getSample(0, i) << ")");
      }
    }
  }

  for (int channel = 0; channel < fc.destBuffer->getNumChannels(); channel++)
  {
    fc.destBuffer->addFrom(channel, fc.bufferStartSample, sampleAccumulator, channel, fc.bufferStartSample, fc.bufferNumSamples);
  }

My problem seems to be when I call sampleAccumulator.addFrom() with more than one source. If it’s filled with only one buffer’s worth of data it sounds fine, but given 2 or more it does not.

In this example my input is a clip of a 440hz sine wave. Here are the samples in sampleAccumulator with one SamplerNote in the list:

image
image1332×470 103 KB

Here’s the two sets of samples I’m adding overlayed:

image
image1338×464 146 KB

Here are the samples in sampleAccumulator with two SamplerNotes in the list, the second sample having been triggered to play slightly after the first:

image
image1222×498 109 KB

As you might expect this sounds bad and is not what I want. What am I doing wrong here, or what intermediate step am I missing? How do I end up with a buffer that sounds like two audio clips playing simultaneously?

2

I found the solution!