Welcome to our Frequently Asked Questions!

If you have a question that has not been answered here, please ask us on our Talk Page.

List of questions asked - Scroll to see the answers
Do I need prior knowledge on bird sounds in order to participate?
Where can I learn more about these birds and get extra identification help?
What if I am unsure if a sound is from the target species or not?
I can hear a bird, but it is too faint to make out. What should I do?
Why are some audio clips louder than others?
I am hearing a lot of background noise that doesn't sound like it came from a bird. Is this okay?
What if I feel like the AI is always right/wrong for a particular species? Should I be concerned I am not doing something right?
What is a spectrogram?
Are you using volunteer data to train the AI?
I am extremely knowledgeable when it comes to bird vocalizations. Is there a way I can further contribute to this project?

Do I need prior knowledge on bird sounds in order to participate?

• No, you don't need prior knowledge. All you need is a willingness to learn!
• While participating in this project does require some initial investment from volunteers to learn bird vocalizations, we provide resources to help you quickly and easily get up to speed.
• Check out our field guide and education page for educational resources that will help you be successful in this project!

Where can I learn more about these birds and get extra identification help?

• Check out the field guide on the right side of the screen, if you haven't already!
• The Cornell Lab of Ornithology has a great, free resource for learning about birds called All About Birds. This website provides a great introduction to each bird as well as their habitat, diet, and cool facts. Each bird also has a "Sounds" section, which teaches you about the different vocalization each bird makes. We highly recommend you utilize this resource in addition to the provided field guide.
• For more resources, see our "Education" page! There are a bunch of helpful links to websites and books provided there.

What if I am unsure if a sound is from the target species or not?

• We understand that some audio clips can be very challenging to make a determination on for a variety of reasons. All we ask is you try your best and trust your gut. You're likely better at this than you think you are. Remember that many other volunteers will classify the same audio clip as you, so we will be able to determine the correct answer via group consensus, even if you happen to be incorrect.
• If you are very stuck on an audio clip, you can share it on the Talk page to get feedback from researchers and other volunteers as well.

I think I can hear a bird, but it is too faint to make out. What should I do?

• We recommend volunteers wear headphones if the bird sound is too faint to hear out of computer speakers. Headphones make it much easier to hear bird sounds clearly.
• Changing your volume settings can also help to amplify quiet sounds. Just remember to lower it before moving onto the next clip!
• You can also look at the spectrogram to see if you can visually make out a distinguishable pattern. See the field guide for example spectrogram patterns.

Why are some audio clips louder than others?

• We collected these data from audio recorders that stayed in one place. Birds vocalized at different distances around the audio recorder, meaning a bird singing from 100 meters away will be much quieter than a bird singing directly in front of the recorder.
• It is important that we test the AI's ability to accurately identify bird vocalizations from all distances, which is why there is a range of volumes within the subject.
• We did our best to boost the volume of the quieter audio clips to make them easier to hear and to limit the number of times you have to adjust your volume.

I am hearing a lot of background noise that doesn't sound like it came from a bird. Is this okay?

• Yes, this is to be expected. These recordings were made in places where humans and birds coexist, which means you will likely hear human-made noise in addition to bird song.
• The AI sometime mixes up human-made sounds with birds. For example, sirens could be mistaken for an owl, hammering could be mistaken for a woodpecker, or a dog could be mistaken for a goose. So, if you are only hearing human noise, it is likely because the AI made a mistake. That is where your classification help comes in!
• There is also other natural noise that can be in the audio clips. Wind, rain, insects, and more can be heard in some audio clips. While these sounds might make identification more difficult, it is important to understand how the AI reacts to such sounds, which are un-avoidable in the real world.

What if I feel like the AI is almost always right/wrong for a particular species? Should I be concerned I am not doing something right?

• Trust your gut! The AI seems to be much better at detecting some species over others. There are some species in this project where the AI seems to do a good job, some where it seems to do a bad job, and some with a mix. It is extremely helpful to know which species the AI struggles and succeeds with, as those results can impact or bias our research if we are not aware of them.
• In summary, if you feel like the AI is mostly right or mostly wrong, that is not a cause for concern! Keep on keeping on.

What is a spectrogram?

• A spectrogram is a graphical representation of sound frequency over time. For this project, the y-axis represents frequency in Hertz, and the x-axis represents time in seconds. The greater the contrast between the sound and the background, the louder that sound is. When birds vocalize, they make recognizable spectrogram patterns. Therefore, spectrograms can be a helpful tool for identification.

Are you using volunteer data to train the AI?

• No, we are not. Our team is not associated with the development or training of the AI used in this project.
• We are using volunteer data to refine the AI detections. This means we are using the classifications to get an understanding of which species the AI knows well, which ones it doesn't, and how confident we can be about the AI's identifications.
• To learn more about the AI used in this project, visit their website.

I am extremely knowledgeable when it comes to bird vocalizations. Is there a way I can further contribute to this project?

• Yes! If you are a bird-song-extraordinaire and willing to spend time on this project, please get in touch (see Team page).
• We can promote volunteers to "Expert" status, which make authoritative classifications. This would be extremely beneficial!