Array methods for Audacity, SDRs and Javascript. Low cost array microphones and sensors

Jason, This is my first post on this forum.AudacityTeam.Org. I have used Audacity for a few years, and use it most every day to look at statistics in signals of many sorts. My priority is to do things in Javascript, since as Director of the Internet Foundation for 26 years, the only common language available to the 5.1 Billion Internet users is Javascript. Not to get diverted, I am writing because I was looking to see if Audacity can be adapted to support those 8 and 10 channel audio arrays. I have a 256 channel device and 16 channel devices. Since most humans on the planet are poor and cannot afford $1000 to play with stereo, data, statistics and invention, I always try to find low cost tools.

Today I took microphone elements from two $2.95 Karaoke microphones and wired them to a $1 3.5 mm stereo plug. Black is common, red and white the two channels. Tap the microphones to be sure. I used an “expensive” $12 USB sound card dongle that can do 96000 sps. It can be accessed through Audacity as a stereo microphone. Chrome Javascript can read it. Gradually all browsers are picking up sound and video. I have been doing that for about 20 years but audacity only in the last 5 years or so.

Sound is a good practice for things like superconducting gravimeters, seismometers, accelerometers, magnetometers, temperature, infrasound, vibration, sensors. Some of the experimental seismic arrays have thousands of three axis seismometers operating each at 1000 sps up to 5000 sps. There are also many software defined radio networks and data from many online experiments. I found about 2000 live videos on YouTube, and a few dozen to 100 with live sound. Some sites have multiple cameras with microphones. There are GPS/GNSS, and radio telescopes. The list is very long and I have been tracking such “emerging global networks” for about 20 years. I started with the superconducting gravimeter array to measure the speed of gravity, then found that the three axis seismometers can also track the sun and moon, so I went through all the seismometers to see which ones and which sites are useful. Then I found dozens of ways to measure gravity and hundreds of kinds of data streams where multiple sensors and arrays are used.

But getting that kind of ability to Internet users is hard. Every time someone says “interferometer” they jack up the prices by 10 times.  If they say “medical” 100 times. And it is often the same stuff underneath.

Nyquist language is NOT a global language. I know, I check all computer and human language and domain specific languages used on the Internet. I would recommend that Audacity adapt its algorithms to work in plain vanilla Javascript.  There are native compilers that compile to libraries that Python can use or that browsers can use. I am working on the Chromium source code to greatly simplify and organize it and separate DevTools and applications from the browsing.  I reviewed about 300 of the larger project on GitHub and am simplifying that as well. It should not take groups hundreds of volunteers to do things.  I am also working on “wikipedia” and many other site specific re-writes.

But to stay on track here.  I am just saying hello.  This equation is two traveling waves from one source to two detectors.  If you read the literature on radar and sonar this is pretty basic stuff. Even if the paper and pencil math is tedious and a bother to memorize and use.  I am trying to get all the LLM and other AI groups to handle mathematics in a formal and standard way on the Internet, so pretty much all humans using browsers can simply mention a method, get the full symbolic math for it, have it tied to calculators, simulators and visualization tools. I want to get humans out of “education by memorization” and “let the computers do it”. If someone want to do long division by hand, or symbolic manipulations in math, they can ask the computer to show them how. But that is like pushing a car.  Better to drive the car, or tell it where you want to go.

I just download Audacity-master again from Github. It has 606 folders and 7700 files. And for some reason is 143.871087 MegaBytes. There are lots of c cpp h files py nq – so the group is splitting its efforts and attention. There are many duplicate names, which usually indicates “which one should I use” issues.

When I try to run record while looking at spectrogram it ALWAYS crashes.

I am just saying hello.

Usually to minimize the difference between two things, you can minimize sum of squares of the differences. Auto correlation and pairwise correlation is often called regression. So simple linear regression works. The core of LLM methods is multiple linear regression and most machine learning has roots in probability and statistics. I am 75. I learned those things starting 60 years ago and have used some things for all those years. But I am just one person, so I mostly cannot do more than make videos, review plans and papers and websites, and give some examples and advice.  I am on Twitter(X), YouTube, ResearchGate, and I comment many places on the Internet as Richard Collins, The Internet Foundation.

I just wanted to share that microphones are not expensive. Those guitar pickups are fun and can be mounted to things to increase sensitivity. The SDR software also works with sound cards and oscilloscopes and that gives you nice FFT water falls and data tools. I am trying to build low cost, three axis, high sampling rate, time of flight, gravitational imaging arrays.

Filed as (Array methods for Audacity, SDRs and Javascript. Low cost array microphones and sensors)

I wrote quickly, so I apologize for grammar and tiredness.

Richard Collins, The Internet Foundation