Forum Rule: Always post complete source code & details to reproduce any issue!
Page 5 of 5 FirstFirst ... 3 4 5
Results 101 to 104 of 104

Thread: Different-Range FFT Algorithm

  1. #101
    Senior Member duff's Avatar
    Join Date
    Jan 2013
    Las Vegas
    Quote Originally Posted by Natefp View Post
    @bmillier, what was your easy way to disable the notefreq function when not needed? I'm doing the same thing, integrating the guitar tuner into a bigger project.
    I have a newer version on GitHub that has disable function implemented and also it can decimate the incoming signal and low pass filter it (for aliasing) to help with processor usage. You can decimate 2, 4, 8 which corresponds to ~22050, ~11025 and ~5512.5Hz sample rates. You have to add in the coefficients for the filter though, the examples show this though they are probably not great filter coefficients but worked well for my tests.

  2. #102
    Junior Member
    Join Date
    Mar 2017
    Thank you both. The updated version seems to be much better at recognizing singing as well.

  3. #103
    Everyone, thanks for this discussion! I followed a link from the example code to this forum thread. I had no idea there'd be such great sample code. I bought a Teensy 3.2 after seeing the Adafruit FFT: Fun with Fourier Transforms project.

    My goal is to create a conference-badge style instrument tuner in time for Defcon 28 (2020). I've decided to keep it simple and detect just one tone, like Concert A4 (440 Hz). (hackaday page) So this code is right on point!

    I'm definitely going to try to use it to tune a double bass, and so the conversation about the tuba was helpful. The fundamental frequency of its low E is also 41-ish Hz, but I can tune the instrument by using a tune on e.g. a harmonic played on the A string.

  4. #104
    Oh, I forgot to mention: I'm also going to try an entirely different method: using AI / machine learning. Now that TensorFlow Light enables some TensorFlow models to run on Arduino-compatible chips, this opens the possibility of "training" a model which can recognize an A4 440Hz being played on various instruments.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts