Naive simple pitch detection

I whipped this app up to start experimenting with pitch detection, and also to test live audio input. It performs a naive (zero-crossing based) pitch detection algorithm in realtime. It works best today with whistling (which has a clear, simple waveform); it needs some severe help with more complex waveforms, even like guitar. I'll get back to this eventually. :) There are much better tuners out there right now - like Craig Spence's work.

Check it out, feel free to fork, submit pull requests, etc.

-Chris

Tutorial added by Silvia Pfeiffer

Note

• Chrome Web Audio API does not fully conform to standard - this uses Chris Wilson's monkey patch.

Step 1: Getting audio input

• open tutorial/step1.html
• what does it do?
  • access the microphone
  • set up a audio filter network ("audio context")
  • connect the microphone to the context as input
  • connect the the input to the output (i.e. speakers)

Take-aways:

• how to capture audio from the microphone
• what is an audio context
• how to route audio through nodes in the context
• what is the destination interface

Step 2: Getting access to the frequency bands

• open tutorial/step2.html
• what does it do?
  • get audio from microphone into filter network
  • mixes stereo input down to mono
  • performs a frequecy analysis of the mono data on a 2048 window
  • draws the frequency data every 10ms

Take-aways:

• understanding filter networks (mono filter, fft filter)
• how to access frequency data
• use setInterval for drawing

Step 3: Calculating pitch

• open tutorial/step3.html
• what does it do?
  • replace fft display with pitch analysis
  • finds zero crossings
  • calculates pitch & maps to note

Take-aways:

• pitch analysis can be done in real-time in the browser
• use requesteAnimationFrame for more immediate drawing