General

[rdz-oss]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[rdz-oss]

First bat detected with the system. Next goal is to reduce the false classification of some types of noise.

[kjarkema]

Sweet!

[rdz-oss]

The noise is from the power converter. The fuzz around 22kHz. Trained the ML network to ignore it. But maybe there is a better way to shield from it?

[Pete-the-Geek]

Hi. I'm Pete-the-Geek. I enjoy observing and recording wildlife and this projected is amazing. I have BirdNET-Pi, BattyBirdNET-Pi, BirdNET-Analyzer and BattyBirdNET-Analyzer. Really enjoying it. Well done and keep up the amazing work.

Please feel free to browse my GitHub for scripts that I have used to further extend my use of the systems. You may find some of it useful too, particularly if you also post process the WAV files in additional analyzers like BTO. Please be kind :)

[rdz-oss]

Just tried the system with the wildlife acoustic touch 2 pro. Looks OK so far.

[BobSpong]

Did you do anything special to get it to work?

[rdz-oss]

No just plugged it in. And worked. I do remember others reporting issues, though. The noise level is significantly lower with the pro model.

I consider removing the noise profile of the wa echo meter 2 and audiomoth as a feature in the future, i.e. remove their typical noise profile from the data automatically.

[EZStreamer]

I was curious if anyone has tried BattyBirdNET-Pi with the new "AudioMoth USB Microphone"? Link: https://www.openacousticdevices.info/audiomoth. I provisioned a BattyBirdNET-Pi system on a Raspberry Pi 4 4GB using this AudioMoth mic and it's didn't recognize it. I assumed it would work after reading this: "AudioMoth USB Microphone is a variant of the standard AudioMoth 1.2.0 design which removes many of the components which aren't needed when an AudioMoth is being used as a USB microphone. It's designed to specifically run the AudioMoth USB Firmware and uses an identical audio front-end to the AudioMoth 1.2.0 and AudioMoth Dev variants."

I have placed an order for the original AudioMoth 1.2 to see if I can get a BattyBirdNET-Pi system operational in the meantime.

[rdz-oss]

Hi EZStreamer! Thank you for your interest and trying this out. I have two Audiomoth USB Microphones v 1.2 I believe .

• They were recognized as input sources for the audio signal.
• The system (classifier) is not trained on their particular noise profile well as yet, so there is a tendency for false detections (that bcan be remedied by retraining the classifier with such noise profile data)

May I ask for clarfication

• did you flash the USB Audiomoth with a certain firmware?

• What setting was used on the hardware itself (default, usb/off, custom)?

• which cable did you use for connection?

• -which usb port did you use?

• I use the hardware setting on 'Custom'

• set the settings on 256kHz I believe

• use a data transferring USB cable

• (have to check which USB port I use :)

and that actually works in the sense of being recognized as a sound input source.
After checking the above details, there might be one other source of error. In the Pi, you can set the audio settings to different values. I have notes on that somewhere I can check if the above leads to no working system.

BR
Richard

[rdz-oss]

If it turns out to be the audio setting on the Pi I will add a guide for that in the docs. But lets exclude other potential factors first :)

[EZStreamer]

Thanks rdz-oss, that worked putting the AudioMoth in Custom mode and setting it to 250Khz

[rdz-oss]

Glad to hear that. It actually also works at 386kHz setting. Currently, the software then downsamples that to 256kHz as that is what the classifier works with.

[EZStreamer]

Hi Richard, Thanks so much for the response. I have attached a photo of the AudioMoth MIC I'm talking about and a screen capture from BattyBirdNET-Pi showing no audio signal being processed. I'm using the standard USB with the USB micro cable to connect the AudioMoth mic. I'm using a Raspberry Pi 4 4Gb with a provisioned 64Gb SD card. I tried the AudioMoth in both the USB 2.0 and USB 3.0 connectors in the Raspberry Pi. Neither show output. I have the switch set on USB/Off and have the solid green light on the AudioMoth Mic. Thank you for the tips on the custom setting. I will try that next. I have not tried to play with the USB driver but this will give me something to try. I really appreciate you responding!! Thank you! Bill Powers

…

On 2024-02-02 13:08, Dr. Richard Zinck wrote: Hi EZStreamer! Thank you for your interest and trying this out. I have two Audiomoth USB Microphones v 1.2 I believe . * They were recognized as input sources for the audio signal. * The system (classifier) is not trained on their particular noise profile well as yet, so there is a tendency for false detections (that bcan be remedied by retraining the classifier with such noise profile data) May I ask for clarfication * did you flash the USB Audiomoth with a certain firmware? * What setting was used on the hardware itself (default, usb/off, custom)? * which cable did you use for connection? * -which usb port did you use? * I use the hardware setting on 'Custom' * set the settings on 256kHz I believe * use a data transferring USB cable * (have to check which USB port I use :) and that actually works in the sense of being recognized as a sound input source. After checking the above details, there might be one other source of error. In the Pi, you can set the audio settings to different values. I have notes on that somewhere I can check if the above leads to no working system. BR Richard -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #1 (reply in thread) [2] https://github.com/notifications/unsubscribe-auth/A3NA7XKSG5CME5CETA44HVTYRUTSXAVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DGNBZGU2DC

[rdz-oss]

Hi Bill sure. If memory serves me right, the switch was important - I think it needs to be at 'Custom' for use on the Pi. Perhaps you could give it a shot?

[BobSpong]

You have to setup the audiomoth using the pc app (I used 250khz) and set to custom on the switch. Usb is to connect to pc app.

[EZStreamer]

Thanks Richard, Trying that now. I know the Mic and USB Data cable work. Here's output from the Live App with me talking. Thanks for the help! I'll let you know when I have success. Thank you, Bill

…

On 2024-02-02 14:26, BobSpong wrote: You have to setup the audiomoth using the pc app (I used 250khz) and set to custom on the switch. Usb is to connect to pc app. -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #1 (reply in thread) [2] https://github.com/notifications/unsubscribe-auth/A3NA7XJUXSNWG6QJ3FCCISLYRU4W3AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DGNJQGEZTC

[EZStreamer]

Hi Richard, That worked!! Thank you!! I really appreciate you help. Bill

…

On 2024-02-02 14:26, BobSpong wrote: You have to setup the audiomoth using the pc app (I used 250khz) and set to custom on the switch. Usb is to connect to pc app. -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #1 (reply in thread) [2] https://github.com/notifications/unsubscribe-auth/A3NA7XJUXSNWG6QJ3FCCISLYRU4W3AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DGNJQGEZTC

[EZStreamer]

Hi Richard, I did build up my BattyBirdNET-Pi this weekend and put it out in the field. I run a site called PixCams, which is a wildlife live streaming educational site, https://pixcams.com/. I have a large amount of land and live stream 29 wildlife cams from my property. I have an extensive network build up on the property and installed a PoE version of the BattyBirdNET-Pi system in a remote setting away from houses and any electronic interference. However, I am getting tons and tons of false triggers. See attached photos of the set up and screech shots from the BattyBirdNET-Pi. I made the link public if you can look at it? Link here: http://73.236.105.104:8192/. I didn't put a password on the tools if you need to change anything. I assume I have something set up wrong? I can't imagine the PoE power supply is causing this much RF noise, but maybe I'm wrong? Any help would be greatly appreciated! Thank you, Bill

…

On 2024-02-02 14:26, BobSpong wrote: You have to setup the audiomoth using the pc app (I used 250khz) and set to custom on the switch. Usb is to connect to pc app. -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #1 (reply in thread) [2] https://github.com/notifications/unsubscribe-auth/A3NA7XJUXSNWG6QJ3FCCISLYRU4W3AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DGNJQGEZTC

[rdz-oss]

Down to a few false noise trigger per hour during the day. Will try to get that down even more.

[Pete-the-Geek]

This is great work Richard. Are you able to update the UK classifier with the data I provided? or is there something I can do to help further?

[EZStreamer]

It was exciting to get our first real bat detection. Thanks Richard for making that possible with the recent changes you made.

[rdz-oss]

Here they come. A silver-haired bat?

[EZStreamer]

Yes, a silver-haired bat. Like you said we got 21 detections in like 15 minutes. One was false but the other were real detections. I didn't know of this species so I looked it up and it's a species that migrates in Pennsylvania, USA into wooded areas. They feed early, which makes sense all around. Exciting to see this. Nice work on your system and making it much better Richard!

[EZStreamer]

PS I live in Pennsylvania, USA. We don't have any bats out yet. Still too cold here. I'm just setting this up to see when they return.

…

On 2024-02-02 14:26, BobSpong wrote: You have to setup the audiomoth using the pc app (I used 250khz) and set to custom on the switch. Usb is to connect to pc app. -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #1 (reply in thread) [2] https://github.com/notifications/unsubscribe-auth/A3NA7XJUXSNWG6QJ3FCCISLYRU4W3AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DGNJQGEZTC

[EZStreamer]

Here's the images of the build and the AudioMoth Mic location.

[EZStreamer]

This is a screen shot of the crazy amount of false detections I am getting

[BobSpong]

How far is your POE running?

[EZStreamer]

That cable length is only 12' from the PoE switch. You can see it in the background.

[BobSpong]

I see your setup and raise my setup.
I'm also getting loads of a false positives ATM from birds etc. Am introducing night time usb power sensor switch to only have pi on when dark.

[EZStreamer]

I'm getting false positives all day long, day and night. Maybe this software isn't ready for US bat detection yet?

[rdz-oss]

Hi @BobSpong , very nice setup!
I updated the US classifier with noise data from @EZStreamer . In his case the daytime noise triggers are going down by a factor of at least 10. How is that for you? Tools - System Controls - Update if you want to try the new classifier.

You can also send some more noise data if there is an issue !
Looking forward to upcoming bat season. Will be great to see a few setups running including one here in Munich!

This will get better step by step. Thanks for being part of that!

BR
Richard

[rdz-oss]

Hi Bill,

first of all I like your setup, nice work on that! As for the false detections a few things can be the root cause. Some info to understand the issue and why it takes some work to fix:

• The system scans the input audio and tests 2 times per second. This means that it has roughly 120 tests per minute where it can have a false trigger. For reference, this is 6 times more often than BirdNet-Pi.
• The audio stream covers a 256kHz range, that is more than 5 times larger than the 48kHz used e.g. by BirdNet-Pi

Now, the classifier needs to be trained to ignore noise in the normal sound range as well as in ultrasound. This includes form

• the environement (cars, insects, etc)
• electrical equipment (chargers, plugs, the Raspberry Pi itself)
• The though issue is that this noise is also microphone dependent.

How can this be solved? It takes training sets of a lot of noise samples for each microphone.
These can then be used as non-events for making the classifier robust.
Without such sample data of false triggers, no robust classifier for that microphone.

The Bavaria classifier is fairly stable if used with the Wildlife acoustics echo meter 2. Which is what I use and that is why I have enough samples to make it work.

It is a frustrating experience for a hopeful user as well as for me if other microphones are plugged in, since this leads to the experience you describe above :)

Now, how to fix?

Can you send me the recordings of false detections? I can use them to make a training set of non-incidents. After that, it will be a lot more robust. Ideally more than 200 recordings.

I also plan to add a timer in the system that shuts it off during the day. Did not get to that yet as I am busy with job and family. Do have it on my list to do before next bat season starts.

I have an sftp server that is currently disbled. I can open it up again for your data upload if you like.

Best regards,

Richard

[rdz-oss]

Good evening! I have added the suntimer setting in the advanced settings. The goal was to make it configurable from there, too.

@hystrix1 the sun_info.py should be there if you are on the main branch. Maybe you are on adifferent branch? Such as the pi5 ?
check with
git status
it should be on the main branch. If not, and e.g. on the pi5 branch, then you can change that with
git fetch
git pull origin main
I believe. You would need to change to the main branch. I think I had a bug for a brief time in which the installer script would load on the pi5 branch. Perhaps that is what your system was installed with.

Let me know if that works or you need further help,
Richard

[rdz-oss]

Settings on the main branch look like this currently:

[Pete-the-Geek]

This works perfectly. Within a few seconds of the "official" sunset/rise time.

Incidentally, what happens if it is all processing a backlog of recordings? I assume, based on the services it is stopping, it will halt processing until sunset when it will pick up where it left off.

[hystrix1]

@rdz-oss I am on branch Pi5.

Following your suggestion...
git fetch
git pull origin main
...seems to have worked. I now see the 'Set dusk/dawn automatically from lat/lon' option.

A couple of questions:

1. Do the dusk/dawn times in the boxes update, or does the automatic option work in the background?
2. Do I need to permanently switch to the main branch? How would I do this?

Many thanks for your help so far.

[rdz-oss]

Great!.
The numbers change in the background, i.e. not visibly in the dusk dawn fileds.
You are now on the main branch 'permanently'.

[Pete-the-Geek]

Hi Richard
I see you have closed off your SFTP server. I'm guessing that means you do not require any more of my data streamed to your server. If you would like me to stop (trying to) sending you data, I can terminate my feed to you. How have you managed with it? Has it been worthwhile? I've not seen an update to the classifier yet, but am awaiting eagerly for whatever and whenever this may happen.
Many thanks once again

[rdz-oss]

I am moving the data to an S3 object store to make new space. Perhaps next step could be to directly upload to S3 storage. Need to think a bit on requirements for such an upload system. Would move towards big data potentially :)

[Pete-the-Geek]

I could do that. My sync routing is compatible with S3.
Just let me know.

Incidentally, I've not seen a classifier update for the UK. Is there anything I need to do to pull down a later one or is it just to apply an update through the GUI once I've become available?

[rdz-oss]

Good evening Pete! The PI system will only indicate changes to the BattyBirdNET-PI system at the momenet, not to the BattyBirdNET-Analyzer. By using Tools - System Controls - Update it should also update the classifier. You can verify in the output it displays and check if a .tflite file was changed (pulled by git).

After last years intense development work I needed a bit of a break. Am getting back to it now. I will look into the S3 issue. A quick option would be to setup a S3 bucket for your system and a corresponding user & pwd. I would still need to put some thought into a more genralizable solution, though sometime this year.

Any bats up around your area yet?

[Pete-the-Geek]

They haven't stopped... These are confirmed detections. I have taken the output from the Pi and run it through my application and confirmed each call as being a real call.

I suspected the update. I can see the USA classifier .../bats/v1.0/BattyBirdNET-USA-256kHz.tflite has updated but not the UK one, so I must be on the latest.

I'm glad you are back and in the swing of it. I have done a lot of work with this, BatDetect2 and Sonobat, as well as undertaking a Bat Acoustic course. And I've only scratched the surface!

Just let me know how I can help and I will gladly do so.

[Pete-the-Geek]

If I create my own classifier training tflite file, how would I then be able to use it in BattyBirdNET-Pi? Note that it wouldn't have all of the bats but does have many of them in varying quantities...

[Pete-the-Geek]

Sounds like a great plan.
I think that single species and no call noise files are the way to go. I'm hoping with my non bat calls, the classifier may even be able to interpret them too.

I'm more then happy to upload them to a cloud platform with their metadata. Mine is all stored in a mariadb so I can export in all sorts of ways. I run an SQL Query to categorise, locate and move the files to a "to classifier" folder so I can modify this to fit most possibilities.

I have to say, my bat work, identification, data categorisation etc has come on well, largely due to your battyBirdNET-Pi software and ideas. For that, I am most grateful to you!

[Pete-the-Geek]

All tweaks done and I am only using single species calls in training my classifier. This is not active on my Pi as I'm not sure how best to approach this. I have however forked your BattyBirdnet-Analyser and BattyBirdnet-Pi repositories and then uploaded my classifier called BattyBirdNET-SouthWales-256kHz.tflite.

I think (with my limited Python skills, I have updated advanced.php line 591 to say
$formats = array("Bavaria", "South Wales", "USA", "USA-EAST","USA-WEST","UK");

I think that is all that is needed.

What do you think is best @rdz-oss, how would you like to proceed with this? Do you want to do anything with this more importantly, as it is your product not mine, I'm just a leacher...

[rdz-oss]

Congratulations Pete. I would integrate your classifier into the project.
First the Analyzer, then in the Pi.
For the analyzer, it should be integrated into bat_ident.py and the server.py if memory serves me right.
Currently this is set to be done with pull requests.
You can check if your Analyzer fork is synchronized. If so, you can make a pull request.
Then I review. Perhaps I find something and comment. Then integration or another round of this until it fits.

We would take this step by step and learn together.

[Pete-the-Geek]

I have Synced both repositories, edited Pi's advanced.php and Analyser's bat_ident.py and server.py as well as uploading my classifier files. I have then done a pull request for the 2 repositories and hopefully that is all done correctly. (Sorry that you are also having to teach me GitHub functionality)

[rdz-oss]

Hi Pete, congratulations on the merge! I found one minor concern in the code that can be taken care of later.
What you can do now: Let the South Wales classifier run. Collect 30-100 false triggers by noise. Add them to you training set in a folder called 'background' or one of the other noise categories. Retrain and resubmit to get a version that is more robust to environmental triggers. Running it in Munich finds a lot of Pip. pyg. in the bird sounds :). But that can be dealt with using above approach.

[rdz-oss]

NEWS: 384kHz Setting enabled! Recommended for Pi 5 as Pi 4 has trouble keeping up at that rate. Working on settings to enable this also on Pi 4. Looks promising.

If you have a microphone that works at 384kHz you can now test the function under advanced settings.
The classifier still works at 256kHz yet the recordings are then made at 384kHz. This will allow to collect data that can be used in the future to train the classifier on 384 data as well. Also you can now manually inspect the files e.g. for myotis species up to 190kHz making the system suitable for biodiversity assessments.

Noise profiles at 384kHz not yet available but will come, too.

Enjoy the weekend!
Richard

[rdz-oss]

Testers for new feature wanted: bird during day and bat at night!

Use your setup as is (mic and Pi) . One microphone. Allows to set the classifier to:

• BIRDS 24/7 on vanilla BirdNET
• Switch to Bird detection during day via timer

You can try the new feature on this branch 'bb'
curl -s https://raw.githubusercontent.com/rdz-oss/BattyBirdNET-Pi/bb/newinstaller.sh | bash

Let me know if you find issues!

BR
Richard

[arne1921KF]

How much space does the installation need? I know you recommend a larger one, but I thought that would be due to recording size. I tried a 64 GB SD card, and the installation aborted here:

Collecting gradio==3.41.0
  Downloading https://www.piwheels.org/simple/gradio/gradio-3.41.0-py3-none-any.whl (20.1 MB)
ERROR: Could not install packages due to an OSError: [Errno 28] No space left on device

ETA: got it somewhat up and running on a 256 GB SD card.

Will report back ASAP, since my installation shows some weird behavior.

[rdz-oss]

Hi @arne1921KF , I run it on a system that has only 34GB , too. The large size is necessary for the recordings that are larger in 256kHz than at 48kHz as in BirdNET-Pi. So if you want an entire season or more you need a larger sd card. On my system that has 34GB, only 11 are currently used. So not likely the issue.
How have you prepared the sd card? Did you use a Pi Imager or another system? Raspberry PI OS 64 bit lite?
How much RAM does your Pi have? 4GB or more?

A quick internet search indicates there might be an issue with limited tmp space. https://askubuntu.com/questions/1326304/cannot-install-pip-module-because-there-is-no-space-left-on-device

[arne1921KF]

Thanks for your quick response.

I used a RasPi OS image (current Bookworm lite arm64 image, as of 15.03.2024). I do not use the Pi Imager software, since I'm on a Chromebook. Chromebook users can use the Chromebook recovery tool after zipping the content of the .xy archive to flash an SD with the RaspiOS. I assume this should not have any ill effect on the flashed OS, though.

The install went through with a larger SD, so there's that. Still, weird.

Just FYI: I use a Raspi 4B with 4 GB RAM. I had no problems installing and running vanilla BirdNet-Pi or the currently developing fork from @Nachtzuster, so far (who has, BTW, an experimental RAM disk feature implemented - that could be interesting for your fork as well).
Further setup: WA Echo Touch 2 Pro, and connected it via a 1.5 m USB-A extension cord and a USB-C adapter. USB-C power adapter with a 2 m cord. Grounded the Pi to a radiator to get rid of potential interference (works great with my standard mic and my standard UGREEN USB sound card - the WA mic seems not to care either way, but I got two false Nyctalus noctula detections directly after startup. I deleted those via the Tools - System Controls - Clear ALL Data option after configuring Batty BirdNET-Pi.

I note some weird behavior so far:
First of all, I had some common Pippistrellus about five meters away which didn't trigger a detection and also did not show up in the spectrogram at all. I'll have to look into that more deeply during the next days.

Second, the View Log tab in the GUI seems not to work any longer. I get no output whatsoever, in any level of logging set in Settings. This is weird and inconvenient - I had hoped to see some debug output there.

Third, in Settings, several entries irregularly seem not to be persistent. This concerns Time & Date Timezone settings in the basic settings, but also all combo boxes in the Bat Classifier Settings in the Advanced Settings section (but only the combo boxes, not the dropdown menus and not the entry fields. This doesn't happen all the time - right now, it seems like the system accepted my settings, and I have ticked combo boxes even after rebooting.

I'll look into that on Friday night, if I can. Right now, the system can try and gather some data for the 24 hours. Let's see what happens then.

One last thing: I'm really exited about your work, Richard. I hope the system evolves, and I can gather some support to use it for monitoring purposes.

[rdz-oss]

Hi @arne1921KF , a quick reply as it is getting late. The View Log should work if the classifier is running. You can check in the Services Tab to see if some are disabled. If they are, that might be an error or because of the timer (some services are shut off during the day to stop recording and detecting bats). You can manually restart Core services in the Services tab if you like any time.

If the classifier is not running, that might explain part of what you have seen with the pipistrelles. But lets wait and see.

[arne1921KF]

I checked on the services. All were running. Restarting services and rebooting so far did nothing, the View Log tab delivers a blank black screen.

Should I open an issue and deliver some debugging info there? (Disclaimer: I would need your advice what output you need and how to find it.)

[Svardsten53]

I now have my modified installation up and running and the bats are out flying and I have seen them too. It would be interesting if someone who knows this better than I can look at my spectrograms.

Dvärgfladdermus (Pipistrellus pygmaeus) and nordfladdermus (Eptesicus nilssonii) seem to match the spectrograms I compared them with.

Större brunfladdermus (Nyctalus noctula) has a more diffuse spectrogram but should be present in the area according to an inventory made in 2018 and is possibly correct. However, a lot of detections during daylight.

Trollpipistrell (Pipistrellus nathusii) appears to have a good spectrogram. That species was not included in the inventory but is one that migrates, so could be such a finding.

I will have the installation stationary for a couple of days and then maybe try to move it to temporary locations with a powerbank connected.

https://batnetpi.svardsten.se

[rdz-oss]

Nice! I see you had some great recordings this night @Svardsten53 ! Several hundred recordings including many social calls!
If you like you can try setting the Sampling Frequency Up (Tools - Settings- Advanced Settings - Bat Classifier). The audiomoth can be configured to run on 384kHz and you can use the entire spectrum now. The USB Version of he audiomoth (no batteries) has medium gain and 384kHz set on the 'Default' setting. That would also show you the upper end of some of the calls you have registered that go above 125 kHz.
Did you move your system or did you just have good weather or so?
BR
Richard

[Svardsten53]

Yesterday we had summer temperatures and that probably explains the amount of recordings and the Audiomoth USB mic is on the same place as I had the EMT2.
I changed the sampling frequency, but none of the species in Sweden are using so high frequencies according to the internet, so changed back to be able to easier read the spectrogram.
Myotis emarginatus is 100% a false positive because that species is not found in Sweden. Maybe you would like to collect the registrations to adjust your model. I will remove these registrations in a couple of days. I have marked the species that are not found in Sweden with an asterisk at the end of the species name.
Would it be possible to be able to exclude bats in the exclusion list instead of birds as well as use Flickr id to have pictures of the bats?
Hälsningar L-Å

[Svardsten53]

I'm now back to max Sampling Frequency in settings. Is this value important for species determination?
The disadvantage of this value is that the spectrogram becomes difficult to read, the sound is just like small dots. Would it be possible to make the frequency range shown in the spectrogram selectable but still keep the maximum value in the settings?

[rdz-oss]

I have thought about that. Use different frequencies for spectrogram and recording. That is possible and not too difficult.
For higher freq:

• The Myotis species can only be differentiated at the 'upper' end of their calls (the starting point). That can go as high as 185kHz.
• The data can be used to make better classifiers if the sampling frequency is higher
  Against:
• The file size becomes larger
• Most calls do not go higher than 140kH anyway even for the upper end
• If visualized on full spectrum birds ans some bats are harder to id by eye

So I believe I will add an option for birds and some bats to do spectrogram at lower kHz than is actually recorded.
Perhaps 60KHz range for birds and 125kHz for bats even if recordings can be at 384kHz

BR
Richard

[Svardsten53]

It sounds good. It is not a problem to truncate the upper part of the spectrogram as long as everything is included in the analysis and the lower part shows the true frequency. I have learned that if you want to analyze manually, it is important to see min, max and peak frequency.

[Pete-the-Geek]

Not sure if it is a bug but I find that when there are several files being processed, when it hits midnight it stops recording. Unfortunately, as I an not awake after midnight, I've not been able to check what services are running. Can you point me at the logs so I can see what was happening at that time please?

birdnet.conf file includes

These are just for debugging

LAST_RUN=
THIS_RUN=
IDFILE=/home/peter/BirdNET-Pi/IdentifiedSoFar.txt
ACTIVATE_FREQSHIFT_IN_LIVESTREAM="true"
RTSP_STREAM_TO_LIVESTREAM="0"
LogLevel_BirdnetRecordingService="error"
LogLevel_LiveAudioStreamService="error"
LogLevel_SpectrogramViewerService="error"
SOX_SPEED=0.1
BAT_TIMER="true"
BAT_DUSK="18:00"
BAT_DAWN="06:30"
NOISERED=false
NOISE_PROF="BattyBirdNET-Analyzer/checkpoints/bats/mic-noise/waem2.prof"
NOISE_PROF_FACTOR="0.10"
INPUT_NOISERED=false
INPUT_SPECTROGRAM_COLOR=""
BAT_SUNTIMER=true

[rdz-oss]

Curious. Have not noticed this behavior as yet. You can try
journalctl -eu birdnet_analysis -u birdnet_server -u batnet_server | sudo tee -a /var/log/syslog
and add a pipe for selecting the time.
Perhaps
cat /var/log/syslog | grep error
shows something.
Do you run a cron service at that time?

[Pete-the-Geek]

This is the log file from seconds either side of the "crash" of the service. It is like the service simply stopped. I hope it isn't something in the sun calc routine, and I would guess this to be unlikely as otherwise it would be likely to affect everyone...

Apr 30 23:59:52 battyBirdNET-Pi batnet_analysis.sh[3166997]: Finished /tmp/tmp6gsi6hin.wav in 4.64 seconds
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-78-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: GUANO|Version: 1.0
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: Samplerate: 256000
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:52 battyBirdNET-Pi env[3241085]: Note: BattyBirdNET-Pi
Apr 30 23:59:52 battyBirdNET-Pi env[3240658]: Line = 1.125;1.6875;Pipistrellus pipistrellus;Common pipistrelle;0.9381182
Apr 30 23:59:52 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-94-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-94-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: GUANO|Version: 1.0
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: Samplerate: 256000
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:55 battyBirdNET-Pi env[3241140]: Note: BattyBirdNET-Pi
Apr 30 23:59:55 battyBirdNET-Pi env[3240658]: Line = 1.6875;2.25;Pipistrellus pipistrellus;Common pipistrelle;0.7035523
Apr 30 23:59:55 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-70-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-70-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: GUANO|Version: 1.0
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: Samplerate: 256000
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:55 battyBirdNET-Pi env[3241186]: Note: BattyBirdNET-Pi
Apr 30 23:59:55 battyBirdNET-Pi env[3240658]: Line = 2.25;2.8125;Pipistrellus pipistrellus;Common pipistrelle;0.5352351
Apr 30 23:59:55 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-54-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-54-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: GUANO|Version: 1.0
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: Samplerate: 256000
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:55 battyBirdNET-Pi env[3241232]: Note: BattyBirdNET-Pi
Apr 30 23:59:56 battyBirdNET-Pi env[3240658]: Line = 2.8125;3.375;Pipistrellus pipistrellus;Common pipistrelle;0.5791876
Apr 30 23:59:56 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-58-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-58-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: GUANO|Version: 1.0
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: Samplerate: 256000
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:56 battyBirdNET-Pi env[3241283]: Note: BattyBirdNET-Pi
Apr 30 23:59:56 battyBirdNET-Pi env[3240658]: Line = 3.9375;4.5;Pipistrellus pipistrellus;Common pipistrelle;0.924536
Apr 30 23:59:56 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-92-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-92-2024-04-30-birdnet-21:01:54.wav
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: GUANO|Version: 1.0
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: Samplerate: 256000
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:56 battyBirdNET-Pi env[3241329]: Note: BattyBirdNET-Pi
Apr 30 23:59:57 battyBirdNET-Pi env[3240658]: Line = 2024-04-30-birdnet-21:01:59.wav.csv
Apr 30 23:59:57 battyBirdNET-Pi env[3240658]: Line = 0.5625;1.125;Pipistrellus pipistrellus;Common pipistrelle;0.8821471
Apr 30 23:59:57 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-88-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:57 battyBirdNET-Pi systemd[1]: chart_viewer.service: Scheduled restart job, restart counter is at 88.
Apr 30 23:59:57 battyBirdNET-Pi systemd[1]: Stopped BirdNET-Pi Chart Viewer Service.
Apr 30 23:59:57 battyBirdNET-Pi systemd[1]: chart_viewer.service: Consumed 4.088s CPU time.
Apr 30 23:59:57 battyBirdNET-Pi systemd[1]: Started BirdNET-Pi Chart Viewer Service.
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-88-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: GUANO|Version: 1.0
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: Samplerate: 256000
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:57 battyBirdNET-Pi env[3241383]: Note: BattyBirdNET-Pi
Apr 30 23:59:57 battyBirdNET-Pi env[3240658]: Line = 1.6875;2.25;Pipistrellus pipistrellus;Common pipistrelle;0.7976784
Apr 30 23:59:57 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-80-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-80-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: GUANO|Version: 1.0
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: Samplerate: 256000
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:57 battyBirdNET-Pi env[3241436]: Note: BattyBirdNET-Pi
Apr 30 23:59:58 battyBirdNET-Pi env[3240658]: Line = 2.25;2.8125;Pipistrellus pipistrellus;Common pipistrelle;0.7423408
Apr 30 23:59:58 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-74-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-74-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: GUANO|Version: 1.0
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: Samplerate: 256000
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:58 battyBirdNET-Pi env[3241483]: Note: BattyBirdNET-Pi
Apr 30 23:59:58 battyBirdNET-Pi env[3240658]: Line = 2.8125;3.375;Pipistrellus pipistrellus;Common pipistrelle;0.8311276
Apr 30 23:59:58 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-83-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-83-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: GUANO|Version: 1.0
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: Samplerate: 256000
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:58 battyBirdNET-Pi env[3241529]: Note: BattyBirdNET-Pi
Apr 30 23:59:59 battyBirdNET-Pi env[3240658]: Line = 3.375;3.9375;Pipistrellus pipistrellus;Common pipistrelle;0.69100547
Apr 30 23:59:59 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-69-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: /home/peter/BirdSongs/Extracted/By_Date/2024-04-30/Common_pipistrelle/Common_pipistrelle-69-2024-04-30-birdnet-21:01:59.wav
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: GUANO|Version: 1.0
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: Samplerate: 256000
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: Loc Position: 51.xxxxxx -3.xxxxxx
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: Species Auto ID: Pipistrellus pipistrellus
Apr 30 23:59:59 battyBirdNET-Pi env[3241579]: Note: BattyBirdNET-Pi
Apr 30 23:59:59 battyBirdNET-Pi env[3240658]: Line = 3.9375;4.5;Pipistrellus pipistrellus;Common pipistrelle;0.62141454
Apr 30 23:59:59 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-62-2024-04-30-birdnet-21:01:59.wav
------------------------- Crash was here somewhere -----------------------
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[304135]: 2.25-2.8125-['Pipistrellus pipistrellus_Common pipistrelle', '0.0689033']
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[304135]: 2.8125-3.375-['Pipistrellus pipistrellus_Common pipistrelle', '0.48833397']
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[304135]: 3.375-3.9375-['Pipistrellus pygmaeus_Soprano pipistrelle', '0.058473818']
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[304135]: 3.9375-4.5-['Pipistrellus pipistrellus_Common pipistrelle', '0.46215746']
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[304135]: 4.5-5.0625-['Pipistrellus pygmaeus_Soprano pipistrelle', '0.20719701']
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[302482]: RECORDING_LENGTH set to 5
May 01 12:37:03 battyBirdNET-Pi birdnet_analysis.sh[302482]: /home/peter/BirdNET-Pi/birdnet/bin/python3 /home/peter/BirdNET-Pi/scripts/analyze.py --i /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:13.wav --o /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:13.wav.csv --lat 52.XX --lon -2.XX --week 17 --overlap 0.0 --sensitivity 0.5 --min_conf 0.5
May 01 12:37:04 battyBirdNET-Pi batnet_analysis.sh[299762]: #################### 2024-05-01 12:37:04.293930 ####################
May 01 12:37:04 battyBirdNET-Pi batnet_analysis.sh[299762]: {'lat': 51.xxxx, 'lon': -3.xxxx, 'week': 18, 'overlap': 0.0, 'sensitivity': 1.5, 'pmode': 'max'}
May 01 12:37:04 battyBirdNET-Pi batnet_analysis.sh[299762]: /tmp/tmp81sh_kzi.wav
May 01 12:37:04 battyBirdNET-Pi batnet_analysis.sh[299762]: Analyzing /tmp/tmp81sh_kzi.wav
May 01 12:37:13 battyBirdNET-Pi batnet_analysis.sh[299762]: Finished /tmp/tmp81sh_kzi.wav in 9.32 seconds
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 0-0.5625-['Nyctalus noctula_Noctule', '0.6400669']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 0.5625-1.125-['Pipistrellus pygmaeus_Soprano pipistrelle', '0.11795248']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 1.125-1.6875-['Pipistrellus pipistrellus_Common pipistrelle', '0.18260115']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 1.6875-2.25-['Pipistrellus pipistrellus_Common pipistrelle', '0.10903599']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 2.25-2.8125-['Pipistrellus pipistrellus_Common pipistrelle', '0.14298466']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 2.8125-3.375-['Pipistrellus pipistrellus_Common pipistrelle', '0.84865564']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 3.375-3.9375-['Pipistrellus pipistrellus_Common pipistrelle', '0.3306274']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 3.9375-4.5-['Pipistrellus pipistrellus_Common pipistrelle', '0.14575876']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[304182]: 4.5-5.0625-['Pipistrellus pygmaeus_Soprano pipistrelle', '0.3351433']
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[302482]: RECORDING_LENGTH set to 5
May 01 12:37:20 battyBirdNET-Pi birdnet_analysis.sh[302482]: /home/peter/BirdNET-Pi/birdnet/bin/python3 /home/peter/BirdNET-Pi/scripts/analyze.py --i /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:16.wav --o /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:16.wav.csv --lat 52.XX --lon -2.XX --week 17 --overlap 0.0 --sensitivity 0.5 --min_conf 0.5
May 01 12:37:20 battyBirdNET-Pi batnet_analysis.sh[299762]: #################### 2024-05-01 12:37:20.697749 ####################
May 01 12:37:20 battyBirdNET-Pi batnet_analysis.sh[299762]: {'lat': 51.xxxx, 'lon': -3.xxxx, 'week': 18, 'overlap': 0.0, 'sensitivity': 1.5, 'pmode': 'max'}
May 01 12:37:20 battyBirdNET-Pi batnet_analysis.sh[299762]: /tmp/tmpk6c3hlmg.wav
May 01 12:37:20 battyBirdNET-Pi batnet_analysis.sh[299762]: Analyzing /tmp/tmpk6c3hlmg.wav
May 01 12:37:23 battyBirdNET-Pi batnet_analysis.sh[299762]: Finished /tmp/tmpk6c3hlmg.wav in 2.85 seconds
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 0-0.5625-['Pipistrellus pipistrellus_Common pipistrelle', '0.67182755']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 0.5625-1.125-['Pipistrellus pipistrellus_Common pipistrelle', '0.57435536']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 1.125-1.6875-['Pipistrellus pipistrellus_Common pipistrelle', '0.34716922']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 1.6875-2.25-['Pipistrellus pipistrellus_Common pipistrelle', '0.05237173']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 2.25-2.8125-['Pipistrellus pipistrellus_Common pipistrelle', '0.34838238']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 2.8125-3.375-['Pipistrellus pipistrellus_Common pipistrelle', '0.2263447']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 3.375-3.9375-['Nyctalus noctula_Noctule', '0.30641353']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 3.9375-4.5-['Pipistrellus pipistrellus_Common pipistrelle', '0.025642814']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[304303]: 4.5-5.0625-['Pipistrellus pygmaeus_Soprano pipistrelle', '0.2952344']
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[302482]: RECORDING_LENGTH set to 5
May 01 12:37:36 battyBirdNET-Pi birdnet_analysis.sh[302482]: /home/peter/BirdNET-Pi/birdnet/bin/python3 /home/peter/BirdNET-Pi/scripts/analyze.py --i /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:21.wav --o /home/peter/BirdSongs/May-2024/01-Wednesday/2024-05-01-birdnet-12:33:21.wav.csv --lat 52.XX --lon -2.XX --week 17 --overlap 0.0 --sensitivity 0.5 --min_conf 0.5
May 01 12:37:36 battyBirdNET-Pi batnet_analysis.sh[299762]: #################### 2024-05-01 12:37:36.578189 ####################
May 01 12:37:36 battyBirdNET-Pi batnet_analysis.sh[299762]: {'lat': 51.xxxx, 'lon': -3.xxxx, 'week': 18, 'overlap': 0.0, 'sensitivity': 1.5, 'pmode': 'max'}
May 01 12:37:36 battyBirdNET-Pi batnet_analysis.sh[299762]: /tmp/tmpffdwz9b_.wav
May 01 12:37:36 battyBirdNET-Pi batnet_analysis.sh[299762]: Analyzing /tmp/tmpffdwz9b_.wav
May 01 12:37:39 battyBirdNET-Pi batnet_analysis.sh[299762]: Finished /tmp/tmpffdwz9b_.wav in 2.87 seconds
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 0-0.5625-['Pipistrellus pipistrellus_Common pipistrelle', '0.0012005812']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 0.5625-1.125-['Pipistrellus pipistrellus_Common pipistrelle', '0.14460988']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 1.125-1.6875-['Pipistrellus pipistrellus_Common pipistrelle', '0.03008575']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 1.6875-2.25-['Pipistrellus pipistrellus_Common pipistrelle', '0.05121732']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 2.25-2.8125-['Pipistrellus pipistrellus_Common pipistrelle', '0.0032607738']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 2.8125-3.375-['Pipistrellus pipistrellus_Common pipistrelle', '0.08558872']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 3.375-3.9375-['Pipistrellus pipistrellus_Common pipistrelle', '0.01760557']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 3.9375-4.5-['Pipistrellus pipistrellus_Common pipistrelle', '0.6204782']
May 01 12:37:41 battyBirdNET-Pi birdnet_analysis.sh[304411]: 4.5-5.0625-['Nyctalus noctula_Noctule', '0.14615847']
May 01 12:37:42 battyBirdNET-Pi systemd[1]: birdnet_analysis.service: Succeeded.
May 01 12:37:42 battyBirdNET-Pi systemd[1]: birdnet_analysis.service: Consumed 7.857s CPU time.
May 01 12:37:44 battyBirdNET-Pi systemd[1]: birdnet_analysis.service: Scheduled

[rdz-oss]

There is a dependency between the recording_length, the extraction run that sometimes can hang up on itself. That would be independednt of the time of the day if it happends. I made a change with the restart mechanism in one of the newer commits.
What you can try: set the recording_lenth to 15 and see if it still happends.

it appears that the system is working up quite a backlog in processing , maybe 3 hours
Apr 30 23:59:59 battyBirdNET-Pi env[3240658]: NEWFILE=Common_pipistrelle-62-2024-04-30-birdnet-21:01:59.wav
That can be reduced by making the recording_time larger. Also, I made some mods (took out sleep times) that were unnecessary in some recent commits. On what version is your system running?

[Pete-the-Geek]

Recording length is currently 15 sec
Extraction length is currently 3 sec

[Pete-the-Geek]

This is all working well now I have upgraded to a RPi5 as it simply wasn't able to keep up and died overnight. Not sure why it was always midnight but it has run smoothly for 2 nights now... All good :)

[Pete-the-Geek]

Birdweather integration.

I've been looking at this and trying to work out how to add the suitable integration into Birdweather as it works seamlessly on BirdNET-Pi and I think I understand why. When it automatically attempts an upload, it gets two messages in the log stream stating "Soundscape POST Response Status - 201" and "Cannot POST right now". I tried to manually upload a soundscape to Birdweather and it neither accepted a WAV file, nor a "compressed" WAV file but does accept an mp3 compressed. My guess is that WAV files are not permitted.

I wonder if there is a way to modify server.py so that is does something like this if there is a suitable function, or if a call to ffmpeg can be done to achieve this
with open(args.i, 'rb') as f:
wav_data = f.read()
gzip_wav_data = gzip.compress(convert_to_mp3(wav_data))
response = requests.post(url=soundscape_url, data=gzip_wav_data, headers={
'Content-Type': 'application/octet-stream',
'Content-Encoding': 'gzip'})
------------------

[Pete-the-Geek]

This is now working in the latest update

[hystrix1]

I have had BattyBirdNET-Pi running for a few weeks now on a RPi4 but I seem to have a backlog of unprocessed files in the birdsong directory, starting from yesterday. The last file currently processing is 2024-05-11-birdnet-21:10:25.wav. I have 3627 remaining files on 2024-05-11 to process and another 2050 from last night (2024-05-12).
I'm using suntimer so I restarted all services except for recording to see if the backlog started clearing but nothing is happening. Any ideas how I can get the backlog processed and things back to normal?

[Pete-the-Geek]

Check the log. Is it stuck on one file? If it is, it may be a corrupt wav file. It may be that the file size is small too, that gives you a hint it is a corrupted wav file. If it is, delete the file it is stuck on and it will resume if that is your issue.

[hystrix1]

I removed the file it was stuck on, rebooted, started all Services except for recording. Still stuck and the log has errors with the next file in the queue...

May 12 16:16:22 batty systemd[1]: Started BirdNET Analysis Server.
May 12 16:16:47 batty systemd[1]: Started BirdNET Analysis.
May 12 16:16:47 batty birdnet_analysis.sh[3827]: Files loaded
May 12 16:16:48 batty birdnet_analysis.sh[3827]: RECORDING_LENGTH set to 3
May 12 16:16:48 batty birdnet_analysis.sh[3827]: /home/pi/BirdNET-Pi/birdnet/bin/python3 /home/pi/BirdNET-Pi/scripts/analyze.py --i /home/pi/BirdSongs/May-2024/11-Saturday/2024-05-11-birdnet-21:10:28.wav --o /home/pi/BirdSongs/May-2024/11-Saturday/2024-05-11-birdnet-21:10:28.wav.csv --lat 51.XX --lon -2.XX --week 18 --overlap 0.0 --sensitivity 1 --min_conf 0.8
May 12 16:16:48 batty python3[3524]: Exception in thread Thread-1:
May 12 16:16:48 batty python3[3524]: Traceback (most recent call last):
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connection.py", line 198, in _new_conn
May 12 16:16:48 batty python3[3524]: sock = connection.create_connection(
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/util/connection.py", line 85, in create_connection
May 12 16:16:48 batty python3[3524]: raise err
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/util/connection.py", line 73, in create_connection
May 12 16:16:48 batty python3[3524]: sock.connect(sa)
May 12 16:16:48 batty python3[3524]: ConnectionRefusedError: [Errno 111] Connection refused
May 12 16:16:48 batty python3[3524]: The above exception was the direct cause of the following exception:
May 12 16:16:48 batty python3[3524]: Traceback (most recent call last):
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connectionpool.py", line 793, in urlopen
May 12 16:16:48 batty python3[3524]: response = self._make_request(
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connectionpool.py", line 496, in _make_request
May 12 16:16:48 batty python3[3524]: conn.request(
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connection.py", line 400, in request
May 12 16:16:48 batty python3[3524]: self.endheaders()
May 12 16:16:48 batty python3[3524]: File "/usr/lib/python3.9/http/client.py", line 1250, in endheaders
May 12 16:16:48 batty python3[3524]: self._send_output(message_body, encode_chunked=encode_chunked)
May 12 16:16:48 batty python3[3524]: File "/usr/lib/python3.9/http/client.py", line 1010, in _send_output
May 12 16:16:48 batty python3[3524]: self.send(msg)
May 12 16:16:48 batty python3[3524]: File "/usr/lib/python3.9/http/client.py", line 950, in send
May 12 16:16:48 batty python3[3524]: self.connect()
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connection.py", line 238, in connect
May 12 16:16:48 batty python3[3524]: self.sock = self._new_conn()
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connection.py", line 213, in _new_conn
May 12 16:16:48 batty python3[3524]: raise NewConnectionError(
May 12 16:16:48 batty python3[3524]: urllib3.exceptions.NewConnectionError: <urllib3.connection.HTTPConnection object at 0x7f81a27e50>: Failed to establish a new connection: [Errno 111] Connection refused
May 12 16:16:48 batty python3[3524]: The above exception was the direct cause of the following exception:
May 12 16:16:48 batty python3[3524]: Traceback (most recent call last):
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/adapters.py", line 486, in send
May 12 16:16:48 batty python3[3524]: resp = conn.urlopen(
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/connectionpool.py", line 847, in urlopen
May 12 16:16:48 batty python3[3524]: retries = retries.increment(
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/urllib3/util/retry.py", line 515, in increment
May 12 16:16:48 batty python3[3524]: raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type]
May 12 16:16:48 batty python3[3524]: urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='127.0.0.1', port=7667): Max retries exceeded with url: /analyze (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f81a27e50>: Failed to establish a new connection: [Errno 111] Connection refused'))
May 12 16:16:48 batty python3[3524]: During handling of the above exception, another exception occurred:
May 12 16:16:48 batty python3[3524]: Traceback (most recent call last):
May 12 16:16:48 batty python3[3524]: File "/usr/lib/python3.9/threading.py", line 954, in _bootstrap_inner
May 12 16:16:48 batty python3[3524]: self.run()
May 12 16:16:48 batty python3[3524]: File "/usr/lib/python3.9/threading.py", line 892, in run
May 12 16:16:48 batty python3[3524]: self._target(*self._args, **self._kwargs)
May 12 16:16:48 batty python3[3524]: File "/usr/local/bin/server.py", line 233, in handle_client
May 12 16:16:48 batty python3[3524]: detections = sendRequest(ANALYSIS_SERVER, ANALYSIS_PORT, args.i, json.dumps(mdata))
May 12 16:16:48 batty python3[3524]: File "/usr/local/bin/server.py", line 109, in sendRequest
May 12 16:16:48 batty python3[3524]: response = requests.post(url, files=multipart_form_data)
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/api.py", line 115, in post
May 12 16:16:48 batty python3[3524]: return request("post", url, data=data, json=json, **kwargs)
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/api.py", line 59, in request
May 12 16:16:48 batty python3[3524]: return session.request(method=method, url=url, **kwargs)
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/sessions.py", line 589, in request
May 12 16:16:48 batty python3[3524]: resp = self.send(prep, **send_kwargs)
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/sessions.py", line 703, in send
May 12 16:16:48 batty python3[3524]: r = adapter.send(request, **kwargs)
May 12 16:16:48 batty python3[3524]: File "/home/pi/BirdNET-Pi/birdnet/lib/python3.9/site-packages/requests/adapters.py", line 519, in send
May 12 16:16:48 batty python3[3524]: raise ConnectionError(e, request=request)
May 12 16:16:48 batty python3[3524]: requests.exceptions.ConnectionError: HTTPConnectionPool(host='127.0.0.1', port=7667): Max retries exceeded with url: /analyze (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f81a27e50>: Failed to establish a new connection: [Errno 111] Connection refused'))

[hystrix1]

I think the above log is misleading since that was done from a remote Terminal. Here's the one from the Web Terminal on the BattyBirdNET-Pi itself. Looks like the restart counter is just incrementing...

May 12 16:31:47 batty systemd[1]: birdnet_analysis.service: Service reached runtime time limit. Stopping.
May 12 16:31:47 batty systemd[1]: birdnet_analysis.service: Failed with result 'timeout'.
May 12 16:31:49 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 1.
May 12 16:31:49 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:31:49 batty systemd[1]: Started BirdNET Analysis.
May 12 16:31:49 batty birdnet_analysis.sh[15449]: Files loaded
May 12 16:31:49 batty birdnet_analysis.sh[15449]: RECORDING_LENGTH set to 3
May 12 16:32:14 batty birdnet_analysis.sh[15449]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:32:14 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:32:16 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 2.
May 12 16:32:16 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:32:16 batty systemd[1]: Started BirdNET Analysis.
May 12 16:32:16 batty birdnet_analysis.sh[15688]: Files loaded
May 12 16:32:16 batty birdnet_analysis.sh[15688]: RECORDING_LENGTH set to 3
May 12 16:32:41 batty birdnet_analysis.sh[15688]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:32:41 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:32:43 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 3.
May 12 16:32:43 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:32:43 batty systemd[1]: Started BirdNET Analysis.
May 12 16:32:44 batty birdnet_analysis.sh[15882]: Files loaded
May 12 16:32:44 batty birdnet_analysis.sh[15882]: RECORDING_LENGTH set to 3
May 12 16:33:09 batty birdnet_analysis.sh[15882]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:33:09 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:33:09 batty systemd[1]: birdnet_analysis.service: Consumed 1.009s CPU time.
May 12 16:33:11 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 4.
May 12 16:33:11 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:33:11 batty systemd[1]: birdnet_analysis.service: Consumed 1.009s CPU time.
May 12 16:33:11 batty systemd[1]: Started BirdNET Analysis.
May 12 16:33:11 batty birdnet_analysis.sh[16116]: Files loaded
May 12 16:33:11 batty birdnet_analysis.sh[16116]: RECORDING_LENGTH set to 3
May 12 16:33:36 batty birdnet_analysis.sh[16116]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:33:36 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:33:38 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 5.
May 12 16:33:38 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:33:38 batty systemd[1]: Started BirdNET Analysis.
May 12 16:33:38 batty birdnet_analysis.sh[16344]: Files loaded
May 12 16:33:38 batty birdnet_analysis.sh[16344]: RECORDING_LENGTH set to 3
May 12 16:34:03 batty birdnet_analysis.sh[16344]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:34:03 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:34:05 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 6.
May 12 16:34:05 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:34:05 batty systemd[1]: Started BirdNET Analysis.
May 12 16:34:05 batty birdnet_analysis.sh[16559]: Files loaded
May 12 16:34:05 batty birdnet_analysis.sh[16559]: RECORDING_LENGTH set to 3
May 12 16:34:30 batty birdnet_analysis.sh[16559]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:34:30 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:34:32 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 7.
May 12 16:34:32 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:34:32 batty systemd[1]: Started BirdNET Analysis.
May 12 16:34:32 batty birdnet_analysis.sh[16871]: Files loaded
May 12 16:34:32 batty birdnet_analysis.sh[16871]: RECORDING_LENGTH set to 3
May 12 16:34:57 batty birdnet_analysis.sh[16871]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:34:57 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:34:59 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 8.
May 12 16:34:59 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:34:59 batty systemd[1]: Started BirdNET Analysis.
May 12 16:35:00 batty birdnet_analysis.sh[17102]: Files loaded
May 12 16:35:00 batty birdnet_analysis.sh[17102]: RECORDING_LENGTH set to 3
May 12 16:35:25 batty birdnet_analysis.sh[17102]: Maximum number of attempts exceeded. Exiting & restarting service.
May 12 16:35:25 batty systemd[1]: birdnet_analysis.service: Succeeded.
May 12 16:35:27 batty systemd[1]: birdnet_analysis.service: Scheduled restart job, restart counter is at 9.
May 12 16:35:27 batty systemd[1]: Stopped BirdNET Analysis.
May 12 16:35:27 batty systemd[1]: Started BirdNET Analysis.

[Pete-the-Geek]

Hop on the Pi with ssh and see the size of all of the wav file. I think lots are corrupt as it can't keep up. Those errors look like mine on the RPi4... I don't have the same with RPi5

[hystrix1]

@Pete-the-Geek Many thanks for your help with this. The size of the files seems to be ok...all are 1.6MB. I moved all of the unprocessed files to a temporary folder and started the recording service. Recordings are being analysed...Sparrows are Noctules ;-).

Let's see what happens tonight. I'll put a RPi5 on the shopping list!

[EvaldoOliveira]

- Are you using data in 48kHz (birds) of bats with higher sampling rate (e.g. 384kHz) - In fact downloaded mp3, not wav - Did you set the sampling rate accordingly in the config.py? - Maybe this is why it is not working. I will convert to 48k and set it to 110k, correct? - Are you using BattyBirdNET-Analyzer or BirdNET-Analyzer? - Batty - You have your data in folders of 'Scientific name_Common name' ? - yes - you can run the training 'train.py' script? - Yes - Does that lead to a finished .tflite and _labels.txt files ? - Yes, but not funcional Em seg., 27 de mai. de 2024 às 14:40, Dr. Richard Zinck < ***@***.***> escreveu:

…

Hi @EvaldoOliveira <https://github.com/EvaldoOliveira> , may I ask some more details for clarification? - Are you using data in 48kHz (birds) of bats with higher sampling rate (e.g. 384kHz)? - Did you set the sampling rate accordingly in the config.py? - Are you using BattyBirdNET-Analyzer or BirdNET-Analyzer? - You have your data in folders of 'Scientific name_Common name' ? - you can run the training 'train.py' script? - Does that lead to a finished .tflite and _labels.txt files ? If so, you can try with the analyze.py script with a custom classifier ( option --classifier ) on some test data. BR Richard — Reply to this email directly, view it on GitHub <#1 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADYE3ZVKGTFFANGTQM3GLATZENVX7AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSG4YTM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Abraços, Evaldo +55 (11) 98476-8866

[rdz-oss]

In that case, set the Sampling rate to 48000 and the SIG_LENGTH should the be corrected automatically.
I believe you can the just use the .mp3 files even if .wav would be better. They should be resampled automatically

[rdz-oss]

How many data files do you have per species? How many species are you fine tuning for?

[EvaldoOliveira]

I am just using one species and 30 files. Just to check that the whole flow is working. I am playing one of the files with the cell phone to check. If this works, then I will make a good model properly Em seg., 27 de mai. de 2024 às 14:57, Dr. Richard Zinck < ***@***.***> escreveu:

…

How many data files do you have per species? How many species are you fine tuning for? — Reply to this email directly, view it on GitHub <#1 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADYE3ZX6QYXO6YF2CGMLU6DZENXXHAVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSHAZDM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Abraços, Evaldo +55 (11) 98476-8866

[rdz-oss]

Ok. You can test the classifier before integrting into the Pi using the analzyer.py script in BattyBirdNET-Pi with the --classifier option. That way you you know whether it works on test.mp3 files before moving on. Leave the config.py with the 48000 sampling rate for that test. Once that works, next steps can be taken.

[gutocarv]

Thanks for the advice @rdz-oss

[EvaldoOliveira]

Thanks, I will try it. Em seg., 27 de mai. de 2024 às 15:08, Dr. Richard Zinck < ***@***.***> escreveu:

…

Ok. You can test the classifier before integrting into the Pi using the analzyer.py script in BattyBirdNET-Pi with the --classifier option. That way you you know whether it works on test.mp3 files before moving on. Leave the config.py with the 48000 sampling rate for that test. Once that works, next steps can be taken. — Reply to this email directly, view it on GitHub <#1 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADYE3ZTVJSRPZQWQRUVAFQDZENZC7AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSHEYTC> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Abraços, Evaldo +55 (11) 98476-8866

[EvaldoOliveira]

It did recognize using just "python bat_ident.py" As it has worked fine, I will integrate. Thanks a lot so far! Em seg., 27 de mai. de 2024 às 15:08, Dr. Richard Zinck < ***@***.***> escreveu:

…

Ok. You can test the classifier before integrting into the Pi using the analzyer.py script in BattyBirdNET-Pi with the --classifier option. That way you you know whether it works on test.mp3 files before moving on. Leave the config.py with the 48000 sampling rate for that test. Once that works, next steps can be taken. — Reply to this email directly, view it on GitHub <#1 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADYE3ZTVJSRPZQWQRUVAFQDZENZC7AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSHEYTC> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Abraços, Evaldo +55 (11) 98476-8866

[EvaldoOliveira]

Ups! Looked like it did but it did not Will use with Classifier Servus aus Brasilien! Em seg., 27 de mai. de 2024 às 15:14, Evaldo Oliveira ***@***.***> escreveu:

…

It did recognize using just "python bat_ident.py" As it has worked fine, I will integrate. Thanks a lot so far! Em seg., 27 de mai. de 2024 às 15:08, Dr. Richard Zinck < ***@***.***> escreveu: > Ok. You can test the classifier before integrting into the Pi using the > analzyer.py script in BattyBirdNET-Pi with the --classifier option. That > way you you know whether it works on test.mp3 files before moving on. Leave > the config.py with the 48000 sampling rate for that test. Once that works, > next steps can be taken. > > — > Reply to this email directly, view it on GitHub > <#1 (reply in thread)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ADYE3ZTVJSRPZQWQRUVAFQDZENZC7AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSHEYTC> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> > -- Abraços, Evaldo +55 (11) 98476-8866

-- Abraços, Evaldo +55 (11) 98476-8866

[EvaldoOliveira]

Analyzer successfully run!! Em seg., 27 de mai. de 2024 às 15:17, Evaldo Oliveira ***@***.***> escreveu:

…

Ups! Looked like it did but it did not Will use with Classifier Servus aus Brasilien! Em seg., 27 de mai. de 2024 às 15:14, Evaldo Oliveira ***@***.***> escreveu: > It did recognize using just "python bat_ident.py" > As it has worked fine, I will integrate. > > Thanks a lot so far! > > Em seg., 27 de mai. de 2024 às 15:08, Dr. Richard Zinck < > ***@***.***> escreveu: > >> Ok. You can test the classifier before integrting into the Pi using the >> analzyer.py script in BattyBirdNET-Pi with the --classifier option. That >> way you you know whether it works on test.mp3 files before moving on. Leave >> the config.py with the 48000 sampling rate for that test. Once that works, >> next steps can be taken. >> >> — >> Reply to this email directly, view it on GitHub >> <#1 (reply in thread)>, >> or unsubscribe >> <https://github.com/notifications/unsubscribe-auth/ADYE3ZTVJSRPZQWQRUVAFQDZENZC7AVCNFSM6AAAAAA5SEXGWOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKNZSHEYTC> >> . >> You are receiving this because you were mentioned.Message ID: >> ***@***.***> >> > > > -- > Abraços, > Evaldo > +55 (11) 98476-8866 > -- Abraços, Evaldo +55 (11) 98476-8866

-- Abraços, Evaldo +55 (11) 98476-8866

[rdz-oss]

Congratulations!
We will get the integration into BattyBirdNET-Pi to run, too!

Next step - Install BattyBirdNET-Pi (if you have not already), get it to run smootly (audiomoth microphone recommended on 384kHz setting, median gain). The system then downsamples automatically.

Once that is running (Pi with audiomoth 384kHz ) you can use the file manager to replace the CUSTOM-BIRD-48kHz.tflite and labes file at /home/name_of_user/BattyBirdNET-Analyzer/checkpoints/bats/v1.0/CUSTOM-BIRD-48kHZ.tflite and _labests files with the ones you made and gave the exact same names!

Then go to 'Tools'-Settings-Advanced Settings'- Bat classifier and select classifier CUSTOM_BIRD it will automatically set the 48 kHz for you.

Cross fingers and let me know if something gets stuck.
We will get it to work eventually !

BR
Richard

[rdz-oss]

Will call it a day as it is late here. Will check in again tomorrow morning.

[BobSpong]

Another microphone type available to buy now http://shop.smithrobotics.co.uk

Look forward to testing with battybirdnet pi

[rdz-oss]

Hi @BobSpong
the exclusion list mechanism is working for bats and birds. The bats have not yet been added to the selection list. This is a future TODO. You could get it to work if you manually add the species to the exclusion file (instead of using the WebUI). It is in /home/batusername/BattyBirdNET-Pi/scrips/exclude_species_list.txt.

That should work as a hack. You can use the WebUI to modify the file via Tools-FileEditor

Best regards,
Richard

[MueJosh]

@rdz-oss @arne1921KF

For now I'm ordering 10 USB- Microphones V2:
Bat_Microphone V2

I'm definitely interested in the "full" bat detector but my wallet can't handle both:)
Some time in the future I'm ordering the handheld bat detector (10pcs).
Maybe there will be a less noisy V2 Version?

[MueJosh]

Just a little update:

I've received the PCBs, now I'm just waiting for some missing smaller parts. I just ordered them 😄

[rdz-oss]

Very cool!

[rdz-oss]

Thank you for making this possible @MueJosh !

[rdz-oss]

INFO: I updated the training set for the BAVARIA classifier to better detect pipistrellus pipistrellus. It was shadowed by nathusis pipistrelle. Is much better now in keeping those apart.
Also, I added mor myotis myotis calls to reduce confusion with myotis dasycmene.
FYI, let me know if you run into issues,
Richard