[Lab] Audio Spectrum display

[Justin Hornosty]

justin at slootsky.org writes:

> I've been thinking about doing something like this, but a completely 
> different display method (individually addressable pixels in a line,  
> colour to denote frequency,  or if that just works out to too close to all 
> white, then pixels for a frequency, and vary the intensity...  Or something)
>
>
> http://m.instructables.com/id/LoL-Shield-Audio-Spectrum-VU-Meter/?ALLSTEPS

I have been interested in making something like this myself (though just
a regular spectrum analyzer.

The FFT stuff for arduino seems interesting - never tried to run it
myself, but I get the feeling the arduino would be way too slow.

one of the common chips a lot of people use is the msgeq7 (
https://www.sparkfun.com/products/10468 - little inflated price, I got a
whole bunch of them for much cheaper off ebay)

I was reading about the teensy3 audio adapter the other day,
http://www.pjrc.com/store/teensy3_audio.html

I think I'm going to pick one up at some point, the pjrc guy is actively
writing code (I believe one of the demos at the moment is a spectrum
analyzer)

For dealing with the level stuff I wonder if you can get away with
simply tweaking the fft code to decrease the levels so you aren't maxing
it out the whole time. A lot of it might simply be the choice of
microphone (and it's position) along with the input stage of the
circuit. Someone with more knowledge in amplifier circuits might have
suggestions for the 'correct' way to deal with this.

Sounds like a interesting project :)

-jjrh.