European Goldfinch and "Chardonneret élégant" are the same I think

[blogdemoi]

Hi,

The European Goldfinch and the " Chardonneret élégant" both are "Carduelis carduelis". They both appeared in my "Recordings" today. In settings I selected the French localisation, 6k model.

[ehpersonal38]

Did you temporarily switch your database language to English?

[blogdemoi]

I don't think so. I reinstalled everything this morning from scratch, so at first the database was in English then I switched to French a few minutes later.

[blogdemoi]

Maybe it is better to wait for tomorrow, to see what happens, as I have a lot of "Chardonneret élégant". :-)

[blogdemoi]

I didn't have any "European Goldfinch" detection today. So I suppose you were right @ehpersonal38, maybe I temporarily switched to English.

However I have a lot of " Chardonneret élégant" / "Carduelis carduelis" (the same specie as "European Goldfinch"). When I say a lot you can see by yourself:

I suspect that a colony may live just outside my window :-) and that is the reason why my system is always a few hours behind. What would be the best solution? Put this specie in the excluded species list? But in that case they won't appear in the detections...?

Edit: sadly an Eurasian collared dove was killed today at my home by my neighbour's cat :-(

[ehpersonal38]

That's interesting that your analysis system can't keep up with the demand... didn't you have this problem a month or so ago too? What exactly is happening? On my Pi 4, analysis takes about 1 second, and with recordings being 15 seconds this leaves plenty of time so that a backlog never occurs. On your system your analysis time must be > than your recording length, is that right?

I'm sorry to hear about that dove though. 🙁

[blogdemoi]

Yes, I have always had this problem, I don't know why. I suspected the Pi 3B not being powerful enough and bought a LePotato. It solved this problem for almost 2 weeks but now it is back. Could it be related to the high number of "European Goldfinch" detections. They do sing all day long, since a few weeks.

analysis takes about 1 second, and with recordings being 15 seconds this leaves plenty of time so that a backlog never occurs. On your system your analysis time must be > than your recording length, is that right?

I will check this tomorrow during the day.

[blogdemoi]

On your system your analysis time must be > than your recording length, is that right?

When I (just) look at my "View log" I have this:

23:21:29---9.0;12.0-('Erithacus rubecula_Rougegorge familier', 0.21187726)
23:21:29---12.0;15.0-('Turdus viscivorus_Grive draine', 0.50283855)
23:21:29---15.0;18.0-('Turdus merula_Merle noir', 0.40342802)
23:21:29---18.0;21.0-('Ficedula hypoleuca_Gobemouche noir', 0.11939934)
23:21:29---21.0;24.0-('Turdus merula_Merle noir', 0.12543051)
23:21:29---24.0;27.0-('Turdus merula_Merle noir', 0.16160779)
23:21:29---27.0;30.0-('Turdus merula_Merle noir', 0.24549358)
23:21:29---RECORDING_LENGTH set to 30
23:21:30---BirdNET-Pi/birdnet/bin/python3 BirdNET-Pi/scripts/analyze.py --i BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:10:43.wav --o BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:10:43.wav.csv --lat 49.XX --lon 2.XX --week 12 --overlap 0.0 --sensitivity 1.25 --min_conf 0.75 --include_list BirdNET-Pi/include_species_list.txt --exclude_list BirdNET-Pi/exclude_species_list.txt --birdweather_id "IN_USE"
23:21:30---ANALYZING AUDIO... DONE! Time 5.1 SECONDS
23:21:30---WRITING RESULTS TO BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:10:13.wav.csv ... DONE! WROTE 0 RESULTS.
23:21:30---READING AUDIO DATA... DONE! READ 10 CHUNKS.
23:21:30---Date-time: 2023-03-30 16:10:43
23:21:35---0.0;3.0-('Serinus serinus_Serin cini', 0.061669007)
23:21:35---3.0;6.0-('Turdus merula_Merle noir', 0.32614532)
23:21:35---6.0;9.0-('Turdus merula_Merle noir', 0.26374075)
23:21:35---9.0;12.0-('Ficedula hypoleuca_Gobemouche noir', 0.09956491)
23:21:35---12.0;15.0-('Turdus merula_Merle noir', 0.36165792)
23:21:35---15.0;18.0-('Ficedula hypoleuca_Gobemouche noir', 0.19359885)
23:21:35---18.0;21.0-('Turdus merula_Merle noir', 0.2797583)
23:21:35---21.0;24.0-('Turdus merula_Merle noir', 0.3243402)
23:21:35---24.0;27.0-('Turdus merula_Merle noir', 0.18265428)
23:21:35---27.0;30.0-('Turdus merula_Merle noir', 0.44913417)
23:21:35---RECORDING_LENGTH set to 30
23:21:35---BirdNET-Pi/birdnet/bin/python3 BirdNET-Pi/scripts/analyze.py --i BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:11:13.wav --o BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:11:13.wav.csv --lat 49.XX --lon 2.XX --week 12 --overlap 0.0 --sensitivity 1.25 --min_conf 0.75 --include_list BirdNET-Pi/include_species_list.txt --exclude_list BirdNET-Pi/exclude_species_list.txt --birdweather_id "IN_USE"
23:21:36---ANALYZING AUDIO... DONE! Time 5.0 SECONDS
23:21:36---WRITING RESULTS TO BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:10:43.wav.csv ... DONE! WROTE 0 RESULTS.
23:21:36---READING AUDIO DATA... DONE! READ 10 CHUNKS.
23:21:36---Date-time: 2023-03-30 16:11:13
23:21:41---0.0;3.0-('Ficedula hypoleuca_Gobemouche noir', 0.084815785)
23:21:41---3.0;6.0-('Turdus merula_Merle noir', 0.3460208)
23:21:41---6.0;9.0-('Turdus merula_Merle noir', 0.09523814)
23:21:41---9.0;12.0-('Pica pica_Pie bavarde', 0.109659694)
23:21:41---12.0;15.0-('Buteo buteo_Buse variable', 0.1336259)
23:21:41---15.0;18.0-('Turdus viscivorus_Grive draine', 0.09129031)
23:21:41---18.0;21.0-('Turdus merula_Merle noir', 0.22668992)
23:21:41---21.0;24.0-('Turdus merula_Merle noir', 0.22320637)
23:21:41---24.0;27.0-('Turdus merula_Merle noir', 0.15958188)
23:21:41---27.0;30.0-('Turdus merula_Merle noir', 0.30877343)
23:21:41---RECORDING_LENGTH set to 30
23:21:42---BirdNET-Pi/birdnet/bin/python3 BirdNET-Pi/scripts/analyze.py --i BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:11:43.wav --o BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:11:43.wav.csv --lat 49.XX --lon 2.XX --week 12 --overlap 0.0 --sensitivity 1.25 --min_conf 0.75 --include_list BirdNET-Pi/include_species_list.txt --exclude_list BirdNET-Pi/exclude_species_list.txt --birdweather_id "IN_USE"
23:21:42---ANALYZING AUDIO... DONE! Time 5.1 SECONDS
23:21:42---WRITING RESULTS TO BirdSongs/March-2023/30-Thursday/2023-03-30-birdnet-16:11:13.wav.csv ... DONE! WROTE 0 RESULTS.
23:21:42---READING AUDIO DATA... DONE! READ 10 CHUNKS.

[ehpersonal38]

So from that log, your recording length is 30 seconds, and it takes on average 5 seconds to process a recording. Meaning you theoretically can process 6 files without starting a backlog, yet your system is struggling to even process 1 file in time.

I'm not sure exactly what this means, maybe at some point your processing time is shooting up from ~5s to ~30s+ randomly. Maybe you can watch to see if that happens, and then run htop to see what exactly your system is doing that's causing taking all that processing power.

[blogdemoi]

After a few days with the "European Goldfinch" in the exclusion list I can tell you that my system is back to normal. Too many detections of "European Goldfinch" I presume...