Great Resource for Us Beginners at Audio Recordings

I’ve been really struggling with submitting decent audio recordings - not loud enough, extraneous noises, too much dead space, etc. Then I found these great guides from the Cornell Lab of Ornithology.

https://support.ebird.org/en/support/solutions/articles/48001064341-audio-preparation-and-upload-guidelines

https://ebird.freshdesk.com/en/support/solutions/articles/48001064298

https://support.ebird.org/en/support/solutions/articles/48001064305-smartphone-recording-tips

I spent an evening reading them, selecting one of their recommended apps for editing recordings and working through their tutorial on how to use it, and now you can actually hear the bird songs in my submittals. :-)

Example: while the airplane is still annoying, the ovenbird song comes through clearly enough to ID.
https://www.inaturalist.org/observations/120755513

I know these resources are old hat for a lot of people, but they might help out other folks like me.

9 Likes

Also: the metadata on recordings does not transfer when uploading the recordings. A work-around I found on another thread is to take a random photo at the same time as the recording. Make the observation with the photo, which will bring in the metadata. Then edit the observation, remove the photo, and add the recording. The metadata will be retained even after the photo is deleted.

6 Likes

You may want to read https://forum.inaturalist.org/t/sound-tips-and-tricks/28005

1 Like

I highly recommend recording through the Birdnet app. Not only do you see a spectrogram, you can select a range, save that clip as an observation, and export it to iNat. Unfortunately, none of the metadata exports, so you have to manually enter the time, date, and location fields, but it gets the audio file into an iNat observation.

Thank you for those resources for better sound files! I’ve been trying to make more sound files OBs this year, so these are helpful.

OTOH, I believe iNaturalist does not really like to have random photos uploaded to a record to provide meta data and stored on iNat because it could impact the CV training protocols.

So, I think what you can do to avoid that is some version of:

  • upload the random photo to pick up the metadata
  • upload your sound files
  • then delete the random photo after making your record.
2 Likes

Yes, I really like both Birdnet and Merlin for recording.

I chose the ocenaudio app for editing: I can increase the volume of the recording, as well as select a range. And Cornell has a tutorial for it that walked me through the initial set-up, and is super easy to follow.
https://ml-rds-wordpress-prod-s3.s3.amazonaws.com/uploads/2022/03/Editing-sound-recordings-in-ocenaudio.pdf

Yes, I should have emphasized deleting the photo after adding the sound file.

2 Likes

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.