

Forest vs. Grasslands Pi Box Placement ... 50 Yards Apart
I have two Birdnet-Pi's deployed within 50 yards of each other, but which get rather dramatic different bird ID's. However I do not believe either box is wrong in their birdsong identifications. Both Pi boxes are 100% identical in terms of both hardware and Pi configurations. From a setup vantage point both boxes require a confidence level of 75%, and listen for 18 seconds with zero overlap ... thus 6 snippets of sound per file.
There are some HUGE differences between the two systems aside from 50 yards between the microphones:
- Amity is near my prairie grass and has a much more "open" sound corridor to the stream across the road which is about 125 yards distant
- Greenwood is only a few yards from the forest edge, and within 5 yards of my nearest of three birdfeeders
- In general, my basic habitat is Boreal forest, and my home is about 800 yards uphill from Lake Superior. I live on the northern edge of Duluth right where forest meets town.
Anyhow, I have provided both images of yesterday's ID's as placement of your BirdNET-Pi microphone makes a big difference. For instance, my Amity Pi box picks up both my local Great Horned Owls and Belted Kingfishers much more often than my Greenwood Pi box. If I only used my Greenwood Pi Box I might think both my owls and kingfishers are rarely around, which is definitely not true. In summary, give significant thought to birding habitat in terms of Pi box placement.