Stuck on the first step… (first run)

[MueJosh]

See edit below
Ive never felt that stupid before

I cant get birdnet-go to run (first start)
Ive downloaded the tar file (for raspberry; v0.4.4), unpacked it in my home folder and now what?
Whats the next step to get it running?

Raspberry Pi 02W
Bullseye 64bit lite

Edit:
To clarify my issue a bit:
Is there a command to start the birdnet-go file?
I can read the txt file but this doesnt help me.

Edit 2:
I've downloaded the wrong .tar folder... arm and amd was too close

Now I can configure everything but that's something for tomorrow.
Still dont know how ro get it running (but i havent configured the settings/config yet, maybe the question will answer itself)

[JAnslinger]

Did you figure it out? I am stuck at the same point.

[JAnslinger]

Alright, I have indeed made some progress using the wiki entry (https://github.com/tphakala/birdnet-go/wiki/BirdNET%E2%80%90Go-Guide). In the following, I will document my journey so far, including the shell commands, for myself and other newcomers. Maybe it will help someone. I will expand the documentation as I make further progress.

I am using a Raspberry Pi 3+ equipped with the latest 64-bit Lite Version (Bullseye) of Raspberry OS.

What I did

1. I downloaded the current TensorFlow Lite C precomiled library from https://github.com/tphakala/tflite_c/releases for my raspberry pi which is the arm-version
   wget https://github.com/tphakala/tflite_c/releases/download/v2.14.0/tflite_c_v2.14.0_linux_arm64.tar.gz

2. I extracted this file using:
   tar -xzf tflite_c_v2.14.0_linux_arm64.tar.gz

3. The extracted file was then copied to the library directory:
   sudo cp libtensorflowlite_c.so /usr/local/lib/

4. Afterwards, I updated the library cache:
   sudo ldconfig

5. I created a directory for BirdNET-Go:
   mkdir birdnet-go

6. Then, I navigated into it and downloaded the latest release of the compiled version:
   cd birdnet-go wget https://github.com/tphakala/birdnet-go/releases/download/20240215/birdnet-go_v20240215_arm64.tar.gz

7. This file was also extracted:
   tar -xzf birdnet-go_v20240215_arm64.tar.gz

Upon checking the contents with ls, I noticed that there's also a libtensorflowlite_c.so file included. Therefore, when using ARM, it might be possible to skip steps 1 and 2.

8. Now, the BirdNET-Go executable can be launched with the command:
   ./birdnet-go

9. then I opened the then created config file using
   nano /home/username/.config/birdnet-go/config.yaml
   I edited
   -the locale according to the guide
   -as well as longitude and latitude (which I obtained from https://www.latlong.net/)
   and saved the file using strg+o and strg+x

10. I then tried to run the realtime analysis
    ./birdnet-go realtime

but got an error

username@raspberry3bplus:~/birdnet-go $ ./birdnet-go realtime
Read config file: /home/username/.config/birdnet-go/config.yaml
BirdNET GLOBAL 6K V2.4 FP32 model initialized, using 4 threads of available 4 CPUs
BirdNET locale not set, using English as default
Starting BirdNET-Go Analyzer in realtime mode
Threshold: 0.8, sensitivity: 1, interval: 15
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x40 pc=0x9b5698]

goroutine 1 [running]:
github.com/tphakala/birdnet-go/internal/analysis.RealtimeAnalysis(0x40001ec240)
/home/thakala/go/src/github.com/tphakala/birdnet-go/internal/analysis/realtime.go:47 +0x1f8
github.com/tphakala/birdnet-go/cmd/realtime.Command.func1(0x40001e6e00?, {0xcca26a?, 0x4?, 0xcca1a2?})
/home/thakala/go/src/github.com/tphakala/birdnet-go/cmd/realtime/realtime.go:20 +0x20
github.com/spf13/cobra.(*Command).execute(0x400011ef00, {0x54f6ba0, 0x0, 0x0})
/home/thakala/go/pkg/mod/github.com/spf13/cobra@v1.8.0/command.go:983 +0x840
github.com/spf13/cobra.(*Command).ExecuteC(0x400011ec00)
/home/thakala/go/pkg/mod/github.com/spf13/cobra@v1.8.0/command.go:1115 +0x344
github.com/spf13/cobra.(*Command).Execute(0x40000446e8?)
/home/thakala/go/pkg/mod/github.com/spf13/cobra@v1.8.0/command.go:1039 +0x1c
main.main()

I'll will open an issue and keep you posted.

Other aspects, that I was able to solve:

• How do I check whether my microphone works?

1. I checked if it is listed via arecord -l
2. Then I recorded a file arecord -D hw:1,0 -f cd -c1 test-mic.wav [the number behind hw must correspond with the number in the listed devices from before]----> STRG+C to stop recording
3. Then I played it using my headphones and aplay test-mic.wav

• How to enable I opening the birdnet-go executable from everywhere?

1. Going to the birdnet-go folder and using pwd to find out the path
2. Edit your .bashrc file: nano ~/.bashrc
3. And add the path from above to the end: export PATH="$PATH:/home/username/birdnet-go"
4. Now I can just type birdnet-go

My open questions to be adressed in future edits of this post:

• How do I remove the bug?

• How do I analyze a single file?
  In the readme.txt it is mentioned to use a command like birdnet-go file birdsound.wav --threshold 0.1 but that won't work. It does not know the command.

• How do I autostart the program?

• How do I install birdnet audio output?
  I found a readme.txt in the doc-folder and opened it cd doc && nano readme.txt
  Here is some information about the audio configuration. Amongst other things a recommendation to install Pulse audio. This I did : sudo apt-get install pulseaudio && sudo apt-get install gstreamer1.0-pulseaudio
  I then startet pulseaudio using: pulseaudio --start and checked, whether it started using ps -C pulseaudio
  Then I edited the config file as recommended in the readme.txt using sudo nano /etc/pulse/daemon.conf.
  But I still don't know how this is applicable in birdnet.

• How do I access the web platform?

• How do I add mqtt support to collaborate with home assistant

[JAnslinger]

Thanks a lot for the recommendation of the Nachtzuster fork. I just switched and it's much faster! :)

[isZumpo]

May I suggest using the docker setup mention ed in the readme. With it you don't have to deal with installing libraries etc. A docker run command will be enough.

[MueJosh]

May I suggest using the docker setup mention ed in the readme. With it you don't have to deal with installing libraries etc. A docker run command will be enough.

Is there a way to get the container running after a reboot without the need of typing a command and have another PC always on?

[isZumpo]

Yes there are ways to do that. However, I must warn you that troubleshooting issues etc will require more understanding of docker after this. Though it is not too tricky.

You can modify the first part of the docker run command to be this:

docker run -d --restart unless-stopped

The -d tells docker to run the container in detatched mode, meaning that it runs outside the context of your shell. If you do "docker ps" after this then you will be able to see the container running. If you then want to stop it you will have to do "docker container stop "

The --restart unless-stopped makes the container restart on reboots of your computer and even crashes of the container itself.

In your case it would look like this:

docker run -d --restart unless-stopped -p 8080:8080 --env ALSA_CARD=1 --device /dev/snd -v $HOME/birdnet-go-volumes/config:/config -v $HOME/birdnet-go-volumes/data:/data -e TZ="Europe/Berlin" ghcr.io/tphakala/birdnet-go:latest

Docker-compose is also a popular tool used for managing containers that I might add a little bit of info in the updated docs about. Could be worth reading up a bit on if you find the topic interesting.

[MueJosh]

I will see it as soon as there is trouble ^^

Thank you so much for the explanation!
I already started it last night and it ran just fine, stopping it also worked flawlessly.

Docker-compose is also a popular tool used for managing containers that I might add a little bit of info in the updated docs about. Could be worth reading up a bit on if you find the topic interesting.

I'll check this out on the weekend, thanks 👍

[tomlut]

How to fix your spectrogram issue: #78 (comment)

[MueJosh]

Thanks!
That was the issue 😄

[MueJosh]

May I suggest using the docker setup mention ed in the readme. With it you don't have to deal with installing libraries etc. A docker run command will be enough.
@isZumpo

If i do- i get this:

vogel@birdnetoutdoor01:~/birdnet-go $ docker run -ti
-p 8080:8080
--env ALSA_CARD=1
--device /dev/snd
-v /path/to/config:/config
-v /path/to/data:/data
ghcr.io/tphakala/birdnet-go:latest
"docker run" requires at least 1 argument.
See 'docker run --help'.

Usage: docker run [OPTIONS] IMAGE [COMMAND] [ARG...]

Create and run a new container from an image
-bash: --device: command not found

vogel@birdnetoutdoor01:~ /birdnet-go $ arecord -l
**** List of CAPTURE Hardware Devices ****
card 1: Device [USB Audio Device], device 0: USB Audio [USB Audio]
Subdevices: 1/1
Subdevice #0: subdevice #0
vogel@birdnetoutdoor01:~/birdnet-go $

Or did I get something wrong of setting this up?
Im completly new to docker! I installed it in /home/Username/
same for birdnet-go git clone https://github.com/tphakala/birdnet-go

[evolsoulx]

#81 fyi, my discussion thread has a pretty easy 'quickstart' in it if you use docker. i don't know how to use docker, but check out that one, might get you guys where you need.

[MueJosh]

Ah ok I see 😅 I suspect there is something wrong with the docker installation on your computer. Did you follow the official guide?

https://docs.docker.com/engine/install/debian/ I would suggest following the apt repository installation guide.

YAY! I dont know what went wrong the first time but after reinstalling docker and getting docker for non-root, everything works as.

the species_config.csv should also work now (moved it to config folder)

Now the only issue is the Time, it is 2hours off.
Probably an easy fix- i just dont know how.
(the "non-docker - BirdNET-Go" gets the right time)

2024/03/31 15:47:37 New detection: Black Redstart with confidence: 0.834612
2024/03/31 15:47:44 Updating detection: Black Redstart with confidence: 0.933437

as soon as everything works ill write a small summary for new user (in a new discussion so its not that confusing to read)

EDIT:
It doesnt recognizes some owls, on my BirdNET-Pi i can set the "Species Occurence Frequency Threshold"
Is there something like this on birdnet-go?
I dont mean the sigmoid sensitivity.

[isZumpo]

Glad to hear that it works! Great job!

Yeah, we probably need some metod of setting the timezone in the docker container. Since the container is not aware of the host's time settings, it probably defaults to GMT-0 right now. I can take a look at it this coming weekend when I have access to a computer again. However, if you are up to try something, this should probably work since we are using debian as the base for our container image:

docker run -ti -p 8080:8080 --env ALSA_CARD=1 --device /dev/snd -e TZ="Europe/Stockholm" ghcr.io/tphakala/birdnet-go:latest

See the:

-e TZ="Europe/Stockholm"

Here you should fill in with your own timezone code. You can find it in this Wikipedia list:
https://en.m.wikipedia.org/wiki/List_of_tz_database_time_zones

If you try it, please let me know if it works.

It sounds like a great idea for you to write a summary. I have a in progress commit where I hope to improve the documentation on how to get started etc. However, it is a bit hard to get it right since I have a lot of experience using docker, making it hard to see what parts are difficult for those less familiar with it. Then maybe we can combine the two somewhat, making the docs great for everyone!

For the detection of owls, maybe this discussion could be worth taking a look at?
#91

[MueJosh]

Thanks!

yep, thats is!

Thanks for the explenation
this worked fine:
docker run -ti -p 8080:8080 --env ALSA_CARD=1 --device /dev/snd -v $HOME/birdnet-go-volumes/config:/config -v $HOME/birdnet-go-volumes/data:/data -e TZ="Europe/Berlin" ghcr.io/tphakala/birdnet-go:latest

Owls:
It doesnt really work, i have set up like this:

#Example configuration for species-specific confidence thresholds
#Format: species, confidence
#- Species: Localized common name or scientific name (case insensitive)
#- Confidence value: Range from 0.0 to 1.0 for species reporting.
#Values exceeding 1.0 are used to explicitly exclude species

#Exclude non-bird labels (e.g., environmental sounds, human activities)
Dog,2.0
Engine,2.0
Environmental,2.0
Fireworks,2.0
Gun,2.0
Human non-vocal,2.0
Human vocal,2.0
Human whistle,2.0
Noise,2.0
Power tools,2.0
Siren,2.0

#Higher threshold for uncommon species (which are frequently misidentified)
common kingfisher,0.85
eurasian moorhen,1.0
common snipe,1.0

#Lower threshold for owls (due to rarity or specific interest)
western barn owl,0.25
boreal owl,0.25
eurasian eagle-owl,0.25
eurasian pygmy owl,0.25
eurasian scops owl,0.25
little owl,0.25
long-eared owl,0.25
northern hawk-owl,0.25
short-eared owl,0.25
snowy owl,0.25
tawny owl,0.25

and it doesnt do anything.
Threshold / minimum Confidence is set to 0.70
Sigmoid Sensitivity 1.25
GPS is set correctly
(if its comparing its database to ebird detections this will become problematic since where I live there are eurasian eagle-owl but ebird doesnt know.. so they dont even get detected)

it only detects owls above 0.70 (like every other bird)
should I comment under #91 for this?

See Pictures below, idk if its supposed to look like this
(EDIT: ignore the config.yaml and species_config.csv in the /home/Username[Vogel]/ directory)



EDIT_2_: Mhmm, on my fresh install with my own Guide it looks different again... The birdnet-go folder doesnt even exist... thats something for the next weekend.

I start writing now 👌

[MueJosh]

see

#92