When I understand this right each pixel color is assigned to a specific frequency. They you look at all pixels at once, combine the frequencies into one waveform and then want to generate this waveform. Probably if two pixels have the same color the amplitude for that particular frequency ist higher.
This sounds as if you want to do an inverse FFT. From the image you can the amplitudes for each of the involved frequencies. Inverse FFT than can create you the waveform. You store this waveform into memory, and then use DMA with WaveDAC8 to create the actual signal.