Hello from JP

Gemeinschaft / Community Introduce yourself
1

HI,

I am very new to this forum.
I am not a beekeeper, but I enjoy helping a friend to monitor his hive.
I have just designed a load cell scale + temperature + humidity system using ESP32 communicating with an android phone via firebase messenging.
All this work right now.

Looking forward adding sound acquisition + FFT into the hive !
And why not real time sound acquisition and streaming to phone ? (but I don’t know how)

NIce forum BTW !
JP (from France)

3 Likes
2

Dear JP,

welcome to our humble community. We are happy you enjoy reading within our forum. Good luck with all your endeavors and let us know if you want to put your feet on some of the open source technologies we are developing.

E.g. you might want to think about submitting telemetry data using MQTT instead of using Firebase and then get some beautiful visualization dashboards on https://swarm.hiveeyes.org/ or on a self-hosted instance.

With kind regards,
Andreas.

3

Hi Andreas,

I have looked at your grafana dashboard --> very nice. Impressive work.

Currently I have choosen firebase instead of MQTT for the following reasons :
-both act with the “same” behavior “publish/subscribe” and a broker in the middle

  • firebase has a proprietary free broker running 24/24 7/7… So I don’t have to care hosting it !
  • the client is directly my smartphone (or any other). It does receive the messages any time during the day (or night) even when the phone is locked (but not off). The app can be closed, the notification is fully transparent for the user, everything works in the background. It works connected to the Wifi or connected to 4G GSM network. (you can see your hive anywhere in the world in real time !)
    All the data are stored locally on your phone.

However MQTT is also a very clean solution and your grafana server is very very clean (much cleaner than my smartphone solution currently !)

I will first focus on sound acquisition, FFT and streaming to my phone before looking at another solution for data analysis !

Thanks for your welcome message
JP

4

Hallo JP, great to have you on board! The bad thing first: We have no ready to use audio recording or FFT analyzing code in our repos! But we did some research and you may find interesting informations via

Search results for 'esp32 fft' - Hiveeyes or
Search results for 'esp32 sound' - Hiveeyes or
Search results for 'esp32 i2s' - Hiveeyes

I think the best way would be to use an digital microphone for this purpose, they are tiny, cheap and you have not to deal with pre-amps or other stuff.

What ESP32 boards do you use at the moment and do you use Arduino C or MicroPython?

We have some working examples with Arduino M0 boards for FFT and also transmitting FFT values via LoRa but unfortunately not für ESP32.

5

Danke Clemens !

I bought yesterday evening a cheap I2s INMP441 microphone. Now waiting for reception from Aliexpress !
I am currently using a Lolin32 lite board under arduino IDE.
The FFT library is working on this platform. Transmiting FFT params will not be a problem, specially if I do not sample sound frequently. I have in mind that Bees sound should be quite constant during the same hour ?
And their frequency seems to be in the range 200-500Hz so not a lot of frequencies bins for the FFT.

Now my friend wants to listen to real Bees sound… so I have to find a way to stream audio from the hive. This could be done with ESP32 and I2s bus I found this link (i will try it) : How to stream audio from I2S Mic using Udp? - ESP32 Forum

Thanks for your help
JP

6

Ah, great to know, did you use the “normal” Arduino FFT lib ArduinoSound - Arduino Reference or an other library?

For real sound you may have a look at Sound Samples and Basic Analysis Hive with Queen vs. Queenless hives with vs. hives without queens are one of the most known examples for “different sound”. A lot of beekeeper say they hear the difference. But to be honest even with a long beekeeper experience I’m only in 80 % of the cases right. So I think it is more important to send aggregated data and do some (could also be quite stupid) AI. Live streaming could be boring. ;-)

But we are now in swarming time, so in case you are the lucky one you may hear some Queen piping: Visualization and analysis

7

HI,

I tried this one : GitHub - G6EJD/ESP32-8266-Audio-Spectrum-Display
explaination on youtube here :
https://www.youtube.com/watch?v=RnVeXkrrnPI

I have listened to your audio tracks. That’s really interesting.
I do believe that the Queen peeping tracking is more or less a matter of chance as it should be something rather seldom.
It seems that the difference between active hive and queenless one is obvious on your examples.
But is it always the same frequency signature ?

A lot to study !
JP