Signals processing and filtering

[daccordeon]

wants (if possible):

• filter out 50Hz mains noise
• playback pure tones more cleanly
• techniques to playback music
• techniques to playback voice

[daccordeon]

also determine if problem is weak speaker coupling

[daccordeon]

melaku (from the labs) mentions that having the circuit on a breadboard instead of a proper (soldered) board likely also introduces some noise, I don't have any estimates on the scale of this though (and so whether we should bother)

[daccordeon]

op amp to anti-alias at 8kHz sallen key low pass filter, 2nd order https://www.electronics-tutorials.ws/filter/second-order-filters.html

see Rob and Bill (or Sofia?) about comb filter for the beating n*50Hz

[daccordeon]

added in Sallen-Key filter, seems(?) to work

---

new problem: by not monitoring a single fringe we artificially pitch shift the signal, but to have a high enough SNR the speaker needs to be loud

[daccordeon]

sounding like a duck could be a compression artefact? https://books.google.com.au/books?id=ZfaXadTMeGEC&pg=PA65&lpg=PA65&dq=signal+processing+voice+sounding+like+duck&source=bl&ots=hbbkD6WHXZ&sig=ACfU3U0N0Wn5sMaIfFjK_Kfa1gpQgjFLiA&hl=en&sa=X&ved=2ahUKEwiFx8j1oZvnAhXtzDgGHbVmBvQQ6AEwEXoECAoQAQ#v=onepage&q=signal%20processing%20voice%20sounding%20like%20duck&f=false

[daccordeon]

https://github.com/wilsonchingg/logmmse https://www.researchgate.net/publication/224640875_Subjective_Comparison_of_Speech_Enhancement_Algorithms

---

https://www.researchgate.net/post/How_do_I_generate_time_series_data_from_given_PSD_of_random_vibration_input http://www2.oberlin.edu/physics/dstyer/Modern/Intensity.pdf

[daccordeon]

approximate rest of model away, just focus on the photodiode