Bat detector

I am using Gmail and I am experiences the same. I got one or two replies after I had reacted and the last from Paul and Defragster never reached me ... Unfortunate.
 
The notifications from the forum somehow do not always seems to reach us. Maybe the forum doen not always realize we are offline or something or does think a notification was already sent out. It will not sent a message every time. If it sent out a message and new reply's come no new notifications are sent out.
I did get a notification last Wednesday.

On the echometer touch part of the noise you get could come come form the attached device so if the SNR on that device does not seen all that great try another phone or tablet. I have seen some nice recordings form EMT as wel as from the Teensy based bat detector I don't think it makes a huge difference but I never compared signal to noise ratio's on any other device.

Why is signal to noise important? As long as it is possible to determine the species a recording is good enough. I also use Batlogger M for transects here as the organisation for those transects use these. In Batexplorer the recordings look great, opening them in Audacity these look hideous they have a very high noise level dut still the species can be determinerd quite well.

Just for fun I attached a contrast measurement of one of my daily garden recordings. I also measured the background noise of one of the batlogger recordings which is already -28dB, I do not like the look of those recordings but they can still be used quite well.

1760888562060.png
 
AFAIK, the figures cannot be compared. Mic SNR is the difference between the mic noise and the max sound level over the total audio bandwidth and finally A-weighted. SNR in the Echometer touch datasheet is given in db/sqrt(Hz) for a given frequency band or the difference between recorder self noise and sensitivity for a given frequency band. For example, you can see in the 2nd lower plot, that SNR at 120kHz is quite low, about 30dB.

The special feature of your bat logger M is, that the bat loggers have far too high gain in the preamps, so that recordings contain lot of noise and nearby passes of bats commonly overdrive the recorder.

So, it is probably quite complex to answer the seemingly simple question of a comparison of SNR of different bat recorders.
 
I have some problems with the teensybat. One of my the encoders lately has an erratic behaviour.

What kind of encoder is used?
How many Impulses and how many stops?

Just want to source the right one to replace.

Steffen
 
It is indeed one of the most common encoders with switch. Please check the soldering first. The GND pin (middle) needs some extra heat. There are thermal reliefs but stil the heat is conducted away in the surrounding copper. Sometimes this results in a bad solder joint. These encoders somehow do not always seem to turn in the same direction (seems to depend on brand). The software allows to swap direction but only fot the pair. Maybe you better have one extra or be able te swat the leads. The encoder is the 20mm version with "plum" rod. There are different lengths and also D-style axles. (20mm probably is measured from the housing, measured form the board these are 26mm long).
 
Hi everyone,
just a quick question about gps on the teensybat:
What is the best recommended gps module I should buy to build into my teensybat?

Saw the first bats flapping around yesterday here in south-west germany. So I believe it is about time to prepare for the next summer season.
Happy listening to everyone!

Rainer
 
RadioSferica said:
Hi everyone,
just a quick question about gps on the teensybat:
What is the best recommended gps module I should buy to build into my teensybat?

Saw the first bats flapping around yesterday here in south-west germany. So I believe it is about time to prepare for the next summer season.
Happy listening to everyone!

Rainer
Click to expand...
I uses Beitian BN-220, and also tried a newer version of the same manufacturer which schould use less power but I did like the BN better. I did noit investigate further, any that talk ublox should work I guess.
 
Hi, I have another question to the developers of the wonderful teensybat.

Do I understand correctly that the 23kHz peak in the recieved signal originates from the microphone itself, as it is a little nonlinear in the region around this frequency?
I have found somewhere an idea to linearize the mic with a small network around it so that the filtering in a later stage of the signal-flow would be unnecessary. Unfortunately the original website is gone, but one can find an archived copy: https://web.archive.org/web/2024020...microphone-and-amplifier-1khz-80khz-bandwidth

Please forgive me if my question is naive or in any way inappropriate but would this not be an interesting modification to the teensybat?

all the best from southwest Germany
Rainer
 
Hi Rainer, that looks like some interesting information, I tried to apply some filtering in the microphone amplifier circuit but ended up with unwanted artifacts. Maybe tis is a good idea and if someone actually tried this and has good results it would be nice to know. As far as I can see there are no images of a built circuitboard and actual tests.

There is also something else worth to explore, and that is an acoustic filter as described by John Mears. https://twilighttravels.org/2026/02/03/horns-and-quarter-wave-stubs-with-mems-microphones/
 
Hello @CorBee and everybody!

The bats have started arriving in Virginia and I have been recording a few over the past couple of weeks. I have been working with the AI Claude over the past few months and I thought it might be fun to see if that AI could identify the bats I was recording. I asked it if it knew about TeensyBat and if so, what kind of data could I feed it so that it could identify species. It responded that a picture of the signal would work. So I took a picture of the TeensyBat screen with a recording displayed. It was able to make a pretty good guess as to the type of bat being recorded - but it got very accurate when I sent it a picture of the screen after the recording ended and the peak frequency was marked on the top of the display.

I am wondering if it would be possible to modify the waterfall display to show an indication of the frequencies on the top of the page in the form of a few labels for the tic marks while playback is underway? Or even just showing the summay at the top that is shown when the file is ended would work as well. This would make taking a snapshot to send to the AI much better because it woud indicate the freqency of the displayed traces. Often the best traces of the recording happen before the file has ended so the summary is not shown for the snapshot. This could be a configuration item, since some people may not want it.
 
Last edited:
So you allready have enough data, for the "live" view I think its not a usefull addition as its a very niche thing to use AI during a live-recording ... In my view (but I might be wrong) its more logical to concentrate on a good recording and do all determination afterwards using any software you can use.

cheers
Cor
 
Cor, 100% agree during a live recording - I am taking photos the next morning while doing playback - I would only want it during playback.
 
I suggest using "normal" software for that purpose. Audacity has a great system to properly analyse and show frequency distributions.
1775996756517.png
 
I think we are in different worlds - I casually use TeensyBat to record bats for fun - it is a great tool for that. In the past, I have done as you suggest and moved the recordings over to my desktop, spun up Audacity and processed the recordings and used that to determine the bat species. I have also spun up the BattyBirdNetPi software - which is an AI to determine the species of bat - all cool but IMHO too complicated. Being able to take a picture of the playback screen and send it to an AI and have it respond with a resonable answer as to what bat was recorded is on a different level.

This suggestion is to make using the TeensyBat a little easier and fun for people who are less interested in analysis and more interested in knowing the likely kind of bat they recorded last night.
 
You are free to modify the code, the latest available code is V1.6. By modifying the code you can adjust things to fit your demands better.
We are catering mainly for users that use the device to record/survey bats and for people that like to just "know" there are bats around.

Alternatively ... cant you feed the audio itself into AI directly when replaying ?
 
So I took a look at the 1.6 code and I have the fix for the large SD Card playback

1776019450785.png

I can send the code changes, but they are for verson 1.6

Matt
 
I see batty birdnet was already mentioned, the android app Bat Ai also uses the batty birdnet. If you uses android you could try to connect a cardreader to the USB port and view and classify the sounds via Bat Ai
 
Bat AI? I do not think I know that one. I did setup an app on my computer but it was for European bats ..
 
Since it appears from the page there is no plan to update the code:

1776782551225.png


Would it be possible to post the source so we could start some forks of our own? We certainly could start with 1.6, but your changes for 1.7.1 are very valuable and I would hate to have to try to figure out how to re-implement them.

With kind regards,

Matt
 
You must log in or register to reply here.
Back
Top