First version of audio tagging interface now available
Tagging audio recordings is an essential function you can do to help build a set of examples of bird song and other sounds. We intend to use these examples to train a system to identify these sounds automatically. Once this is available you will be able to automatically and continuously measure bird populations and the impact of your trapping.
The first version of our audio tagging interface is now available. When you listen to a recording from your Bird Monitor you will now see this interface:
This lets you add multiple tags to each of your recordings. You can tag time points on the recording as:
- Nothing of Interest,
- Human, or
- You can enter a custom tag, for example if you recognise the bird species then you can enter this here, or you may want to collect recordings of something else, like goats for example.
When you add a tag a line is added to the bottom, noting the tag, the time in the recording when the tag was added, who added the tag and when they added it. You can also see buttons allowing you to play the recording from when the tag was entered and to delete a tag.
If you are going to add a custom tag we recommend pausing the recording, so the start time is accurate.
This first version of the interface does have some quirks that we expect will be improved upon over time.
- When you add a tag the recording starts playing from the beginning of the recording. In order to quickly get to where you were without listening to the whole recording again, you can press the play button next to the tag you just added (which will be top of the list).
- When we released the tagging interface we removed the ability for you to delete a recording. We will add this back in. If there is a recording you would like to be deleted before this functionality is available then please let us know at firstname.lastname@example.org
- At the moment you can't search through your audio recordings to find examples of recordings with particular tags, like you can with thermal videos.
If there are any other changes you would like to see in this interface then please let us know in the comments below.
Many thanks to Mark Nikora (Waikato Institute of Technology) and Emmanuel Turner (University of Waikato) who have built this interface as part of the Cacophony open source project.