How to start exploring possibilities of using smartphones as recording devices?

Question

I am thinking about starting to explore use of smartphones (Android, iOS) for in-field sound recording for amateur citizen-science. I was not able to find technical information about existing projects. Is there an existing collection of information to start with?

• Which smartphone apps are good to start exploring sound-recording capabilities of smartphones? E.g. which would allow setting gain, auto-gain control parameters, record various metadata (geolocation, gain, clipping, phone movement...)

• How to calibrate the recording parameters using just cheap equipment?

• Possible cheap accessories - like sound reflector made of paper etc.

Answer 1

Start with almost any good-quality sound recording app -- it does not need to be bioacoustics related, any good "sound recorder". There are many apps, and the best choice keeps changing, so I make no recommendations by name. Instead, pick an app for which:

1. you can record uncompressed audio (e.g. WAV or FLAC) -- check this, since most "speech diary" apps only record poor-quality low-filesize audio formats
2. you can access the recorded files
3. allows you to control the gain
4. ideally visualises the audio in real-time: a gain/amplitude level and/or a spectrogram.

Since your question is about exploring the possibilities, I do not recommend using a task-specific app such as BirdNet or Warblr: both are bird sound recognisers, but they're not designed for general-purpose recording nor for accessing the files afterwards.

You should then ask a small group of users to make recordings with their own phone, in their own chosen location, and to send the files to you. Pretty quickly, with a small group of users (less than 10), you will have a first idea of the diversity of recording quality, noise levels, user error, for your goal.

When you have collected the audio files, you can use analysis software to determine what's possible (e.g. automated recognition), before going further and trying to connect an app to some more involved system.

Answer 2

I just download Birdnet on an Android phone and it gives you

• running spectrogram of the sound
• the possibility to get birds be classified (over net?)
• store data

Could not figure out, how to store data snippets, etc. Maybe the creators (Cornell, TU-Dresden) could chime in and pick-up the suggestions for transforming Birdnet to a better Citizen Science tool

Answer 3

I can share this outreach poster, which offers several apps for Android and iOS that we found worked pretty well for simple recording, viewing spectrograms, and saving sound snippets.

Caveat: This is from 2013. I'm sure there are newer, better apps out there, but these may provide a start.

Acknowledgement: This poster was created in collaboration with my former lab group at Oregon State University for the Hatfield Marine Science Day outreach event, which opens up the Science Center to the public on a Saturday. We had several folks successfully download these apps and have fun with their kids looking at spectrograms that day!

Answer 4

I worked on a project for several years where we used smartphones as autonomous recording devices. This work doesn't address the citizen science component of your project, but there may be some useful ideas here regarding settings, apps, workflows, and potential opportunities and pitfalls around managing data from smartphones:

Donovan, T., Balantic, C., Katz, J., Massar, M., Knutson, R., Duh, K., ... & Dias, J. (2021). Remote Ecological Monitoring with Smartphones and Tasker. Journal of Fish and Wildlife Management, 12(1), 163-173.

Answer 5

This isn't really an answer to the main question, but as people mentioned in other answers the BirdNET app does appear to have aspirations of collecting acoustic "point counts" by having users stand in given location for a set period of time. That way we're getting a bit closer to data that gives us absence information as well.

See Connor Wood et al.'s paper in Plos https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001670 . Where they write

An optional “point count mode” with which users could submit 3- to 5-minute continuous soundscape recordings would allow communities to document an estimated 5% to 15% of the local bird species per week even with minimal participation [9]. Moreover, these acoustic point counts would yield analytically valuable nondetections of the remaining species.

Seems like it would be helpful to figure out the differences in phone mic performance across models and account for that in any estimates of occupancy.

Answer 6

Wildlife Acoustics has the Echo Meter Touch, which allows you to record and visualize bat clicks in real-time on your phone or tablet. I've seen it in action and it's pretty cool, if limited in its scope.

However, what you can do is calibrate the Echo Meter in the same way you would any other microphone from WA. I don't know how to do that so much with the mic on your smart phone. I think the availability of cheap devices that can easily added to smartphones may be the way forward.

Answer 7

Definitely not an existing app solution, but maybe halfway there. I played around with the Python-Android interface package 'Kivy' and can only recommend it.

Kivy allows you to write a custom app without actually needing to dive into the depths of Java/Android packaging.

Kivy provides Sensor APIs and allows you to control them and generature GUIs for your own use case. If your use case is rather specific, perhaps it may be of interest?

Answer 8

Your question could be improved by being more specific about the goals of your "amateur citizen-science" project. For beginner bioacousticians, I strongly recommend focusing first on signal detection and classification goals and relative noise metrics, saving absolute calibrated noise levels for later. The latter in-air could be done with a decibel meter, but in-water will be more challenging and expensive (as of 2022).

For iOS/OSX users interested in free/low-cost, open source, and/or open data solutions, I recommend this work flow:

1. Use the native Voice Memo for recording from the phone (via built in mic, or external with e.g. a lightning dongle). No live spectrogram, but you can set format to lossless and even do initial trimming and file re-naming in-app.
2. Sync through iCloud with your Mac for initial data management
3. On OSX, use free, open source Audacity for visualizing waveforms and spectrograms, for generating label "tracks," and for basic analyses (e.g. relative power of spectral peaks or SNR)