Could anyone give me some technical specs or advice with the AGC disable? I know that I have to connect TH to MICBIAS but these pins are incredibly tiny. I know there are through holes but the datasheet does not mention these, to my knowledge. Does anyone know if the through holes are connected to the TH or MICBIAS? I am trying to get raw values from the mic and I do not want any clipping.
#AGC Disable - Adafruit Electret Mic Amp - MAX9814
cedar maple
sharp pine
Could anyone give me some technical specs or advice with the AGC disable? I know...
Hello Sir! Could you say more about what you are aiming to build?
I think there's an inconsistency in what you say: one of the aims of AGC is to prevent clipping, but you want to disable it?
You refer to https://www.adafruit.com/product/1713. Have you considered https://www.adafruit.com/product/1063, which I think (I'm not an expert in Adafruit products) is similar, but without the AGC?
cedar maple
I have a school project where we are measuring sound of water being poured at various temperatures. These were the boards that we were instructed to use and given. I am trying to work within the parameters given. Through research I've found that AGC is good for consistent volume, but when it comes to softer sound the sound becomes normalized toward the same output level. Hot and cold water (when poured) sound different but the difference is subtle. The AGC pulls the loudness of different subtle sounds towards the same target. Essentially right now, I can almost scream into the mic and have a very small fluctuation in an analog signal. The analog range that I am caping out at is from 250-350ish, my goal was to expand this range essentially by disabling the AGC.
https://www.analog.com/media/en/technical-documentation/data-sheets/max9814.pdf
Page 10 - Setting the AGC Threshold - states what to do but not how to do it. I was curious to find out about the through holes on the board to see if the were for that or not.
In other words, I am looking for raw data points to collect.
Where the AGC does not allow for this.
sharp pine
Yup, AGC is like having a robot that turns the mic gain (volume control) up and down automatically. I had one on a tape recorder years ago. It meant I could record someone speaking in a whisper, and then someone playing a trumpet, and get a clear recording of both without needing to adjust anything during recording. The trumpet is massively louder than the whisper. If the gain was fixed, I'd end up with a decent recording of the whisper, but the trumpet would be clipped and distorted... or a decent recording of the trumpet, and silence where the whisper should be.
So yes, AGC normalizes loud and soft sounds, so they both have the same output level.
I've just run my kitchen hot tap. It's cold weather here today, so the water that came out was cold to start with, then as the boiler started up, it was warm, then hot. Listening to the water as it ran, I think you are right that the sound changed as the temperature changed.
But I think the amplitude (loudness) of the sound didn't change much as I ran this experiment.
Have you been taught anything about sound frequencies, and FFT?
cedar maple
Sorry for the late response, I have two other projects going on at the same time.
Have you been taught anything about sound frequencies, and FFT?
Yes, very loosely due to the structure of the course, so my understanding is limited.
For background, the experiment set up is that we are using an adjustable temperature kettle. (6 different temperature settings) We don't plan to pour while boiling as to eliminate an extra noise variable. Once the water is still, we pour on top of a temperature sensor in a container, while recording. The theory is that the viscosity of the water changes with an increase/decrease in temperature. Then associate the voltage from the microphone to the temperature. From there take a guess as to what the temperature is based off the sound that the 'pour' makes. I only mention this to make it a bit clearer.
Are you saying that I should be more concerned with the frequency of the pour than the amplitude that is given?
Thanks again for speaking with me about this. Hope your Thanksgiving is a good one!
sharp pine
Interesting experiment!
Are you saying that I should be more concerned with the frequency of the pour than the amplitude that is given?
If you have a fairly steady & constant source of sound, like a running tap for example, then I think the amplitude of a recording (without AGC) would be mostly determined by the distance of the mic from the sound source. When the mic is nearer, it'll hear louder sound, and the amplitude will be higher.
If you could disable the AGC, you could test this.
But perhaps your teacher thinks the amplitude isn't very relevant, so they've provided a mic with AGC? (Can you ask them?)
the frequency of the pour
I did a quick web search for [audio fft demo] and found this analyser, which runs in a web page: https://www.scichart.com/demo/javascript/audio-analyzer-fft-example
If you are able, I'd suggest playing with that a bit - it might help you to think about frequencies. A good thing to try would be whistling to your computer, if you can. (Whistling is a fairly pure sound: it's easier to see on this display than singing for example.) Look at the display at the bottom left as you whistle. I just tried - my clearest results are below.
The whistling is (kinda) a single frequency - it looks like a single peak in those charts.
But most sounds combine many frequencies. If you think about an orchestra playing, if only the double basses are playing, the frequencies present are all low. If only the flutes are playing, there will be only (relatively) high frequencies. If the basses and the flutes are playing at the same time, you'll hear both at once.
I wonder if there is something similar going on with the sound of water pouring - a combination of lower and higher sounds.
cedar maple
But perhaps your teacher thinks the amplitude isn't very relevant, so they've pr...
This was just the microphone that they had on hand
I was ignoring values below 500 Hz per incremental temperature point
A lot of these inconsistencies came down to the experiment set up and some interesting rules per the classroom. Like we weren't allowed to leave to a quieter area to conduct the experiment. Honestly, after the instructor told us that I kind of gave up but I wanted to show you what I ended up getting. I think if I were to do it again. I would have performed this experiment with some soundproofing and a quiet room. Very loosely, there seems to be a trend but extremely loosely. If some mechanical components had changed as well it would have been nicer, Pour rates, pour distance, using a beaker instead of a cup for air column velocity, using a microphone without AGC, the combination of these things I think would get our R2 value up to at least 50-60% but currently 1st and 2nd orders are showing 17% with clear outliers
cedar maple
Also my apologies, I am not getting notifications for this thread for some reason