Is there a way to upload audio files to help identify either an animal or insect sound?
Yes, there is! Do you use an app or the website? I upload audio on the website sometimes, but I don’t know how the app works. I’m sure the process is somewhat different.
You will need to upload to inaturalist.org
The Android app has audio functionality directly parallel to its photo functionality: there are icons for “Record Sound” (create a recording) and “Choose Sound” (upload a recoding that you previously made).
Insects are animals though.
If you want to identify an animal sound with the INaturalist AI then just use some spectogram program and screenshot the spectogram and then upload it along with the audio file.
No, please don’t do this. The iNaturalist “AI” (we call it CV, or Computer Vision) is not intended to be used on spectrograms. It could theoretically be trained on them in the future, but currently if you upload one it will be lumped in with all the actual photos of the species, which makes for a confusing dataset. iNat staff do not encourage uploading spectrograms unless and until there’s a system in place to handle them separately. See the discussion here: https://forum.inaturalist.org/t/automatically-add-a-spectrogram-view-to-observations-with-sounds/
There’s a number of variables involved in generating a spectrogram. If one doesn’t know the specifics about the identifying characteristics of a particular audio, there’s a good chance that useful details will be missed. If needed, spectrograms can usually be easily generated by audio identifiers to the specifications that work best for them for that particular species or sound.
I don’t ID audio if there’s a spectrogram, app screenshot, or habitat only image included as evidence. I’m not sure how much of that is about guidelines, data integrity and consistency and how much is about what I find pleasant when I’m browsing the galleries. :)
Images can still be included as part of the notes or in a comment by hosting elsewhere and using the “<img src=…” tag
I post “worthy” (clean and interesting) spectrograms embedded into my ID and add the field “spectrogram: yes”.
When I use spectrograms to “proof my point” when in disagreement, I embed them into my comment.
Never as Specimen-photo.
When identifying sounds, I download the file either way, so I don’t rely on a user made spectrogram. I create my own (for you)
(Same for habitat-shots)
Does iNat have “CV “ identification for sound files yet? I record using Merlin and save the files through Merlin as .wav files. When I upload them on my apple laptop to iNat, make an observation, then click on the ID box, nothing happens - no ID suggestions appear.
No, it does not. You have to identify them all yourself, which is unfortunate for those of us who aren’t bird-call experts. Most of mine I just ID as “Birds” or “Insects” or even “Animals” if I really don’t know what it is. I suppose since you’re using Merlin you might get suggestions from that which you could use in place of CV.
Here’s an old discussion that you might like to read.
https://forum.inaturalist.org/t/recognize-sounds-automatically/
I think it’s unlikely to happen since iNat staff and developers are already a bit overloaded with ongoing projects, but it’s theoretically possible.