DSP Ideas

• Great capstone site with ideas and examples. here

• Tartini Pitch Detection Program (and papers!) Cross-platform pitch detection algorithm, plus his thesis
• Great StackOverflow post, particularly on cepstral but includes everything.
• Pitch Detection (Wikipedia)
• Can do with time-domain (autocorrelation) techniques very accurately but bad with noise, octaves
• Can also do in frequency domain with FFT, but not as accurate
• Third method is to combine their strengths together in a dynamic programming problem: http://pods.binghamton.edu/~hhu1/paper/Zahorian2008spectral.pdf
• Somehow Roger B. Dannenberg did it back in the 70s with a trumpet, so for simple stuff it can't be too bad…

How do we sense pitches anyways?

• If it's just a simple tube, then there are specific resonances based on the length and it won't hear all frequencies well!

Turns out it's … surprisingly complicated! Reading in a book now…

• MIT guy made a multi-spectrum antenna based on hardware resonators. 59982.pdf
  • How does it do it??? Linear time… can you make it with a formula? Reread that section

Thus, it appears that the cochlear spectral analysis algorithm delivers the most efficient trade-off between analysis time and hardware cost. It exploits the scale-invariant nature of an exponential to achieve scaling O(n) in both quantities

• Virtual channel Really way cool., ends up being just splitting up the current proportionally, not really cancelling out waves like I thought (although it feels like they tried that too)

• Whitepaper links to more stuff https://www.advancedbionics.com/content/dam/ab/Global/en_ce/documents/libraries/Professional%20Library/AB%20Technical%20Reports/Sound%20Processing/AB_HiRes_Fidelity_120_Sound_Processing_report.pdf#page=3

• Virtual surround sound is still cool though

• Might not even be that great of an idea / iPhone is too slow?
• How It Works and Ways to use it. Great article, talks about dynamically splitting up down-beats based on grains
• Some more articles
• Does it even sound good?