fzero~ anomaly

chrisbob12's icon

I'm interested in using fzero~ but found something odd, which I hope someone can help with.

I did a quick test on the first tab help patch with a tuning fork and the audio input. To my surprise the patch gave a frequency of about 179 Hz for the 440Hz fork.

To compound my puzzlement, I selected the oscillator option on the input, which is a cycle at 440Hz and this too gave 179 Hz for the sine and triangle waves, with only the square wave giving 440Hz.

Clearly, I'm missing something obvious; can anyone advise?

Roman Thilenius's icon

the problem is simply that (most) methods of finding the fundamental/loudest frequency in a signal wont work good with a static cosine tone (or tri, or simple harmonic sound) as input material.

try the same with a 440 Hz cello sample and you will see that it will work better.

-110

mzed's icon

I can verify that fzero~ isn't very good with sine waves. It was optimized for more natural spectra.

chrisbob12's icon

Roman and MZED - thank you for clarifying that. Since I want to use it for pitch tracking my voice, there will be plenty of other spectral content.