A few questions #72
Replies: 1 comment
-
Hi @ElisendaPeris, thanks for your comments! About your questions:
It's the 3s fragments (windows) with no overlap, as with the default BirdNET settings. In my experience, having a call span two 3s windows hasn't been a particular issue for identifying calls. However, there is an issue where a call chopped off at the beginning or end of a window is more likely to generate a false positive detection. Since the overlap adds more window edges, you end up with more false positives. It also takes longer to analyse. That's been my experience, anyway.
No, it's just the 2.4 model, I'm afraid. It is generally better than the older versions, although as you say, there may be some species where the performance has deteriorated.
I'm sorry to hear that. It's not a issue I've heard from anyone before, and it works for me, Perhaps you could go into a little more detail - if you follow the guidance in the issue / bug report template, this will help me pin things down. |
Beta Was this translation helpful? Give feedback.
-
Hello,
I am new using Chirpity and I've been comparing it to the BirdNET-analyser. I really like that Chirpity is way more user friendly, and allows to directly listen to the audio fragments where a detection is made, in fact, in general I think it is way easier to use. Nevertheless I have a few questions that I haven't been able to find:
Thank you very much for your imputs and for developing such an interface!
Cheers,
Beta Was this translation helpful? Give feedback.
All reactions