search

orcasound / orcagsoc

Google Summer of Code projects and products related to Orcasound & orca sounds
http://orcasound.net
MIT License
19 stars 14 forks source link

Stream external hydrophones to Orcasound website #29

Open valentina-s opened 2 years ago

valentina-s commented 2 years ago

Orcasound hosts a collection of hydrophones in the Puget Sound, which allow monitoring whale behavior when they are visiting the region. However, when orcas are out in the Pacific Ocean, it is harder to know where they are. This can be improved by integrating data streams from hydrophones maintained off the Oregon coast by the Ocean Observatories Initiative into the Orcasound web app and machine learning pipeline via the orcanode code. The hydrophone data is archived here and is stored in an .mseed format (popular within the seismic community). Some initial efforts are here.

On an extended timeline, this project could also involve an extra step of processing the audio data to remove the annoying “clicks” of a current measurement device (ADCP). It could also extend by ingesting hydrophone data from the outer coast of Vancouver Island (British Columbia) via open access provided by Ocean Networks Canada.

Expected outcomes: Increase access to NSF-funded audio data from hydrophones in killer whale habitat on the outer coast, opening the door to wintertime acoustic detections of endangered orcas.

Required skills: Python, Docker

Bonus skills: Cloud Computing, Audio Processing, javascript, NextJS, Elixir and GraphQL

Mentors: Steve, Scott, Val, Valentina

Difficulty level: Medium

Project Size: 175 or 350 h

Resources: Obspy OOIPY miniSEED https://rawdata-west.oceanobservatories.org/files/CE02SHBP/LJ01D/11-HYDBBA106/2021/

Getting Started: Run the docker setup locally Check out how the other streams are played on orcasite

Points to consider in the proposal: How do you handle missing data, how do you handle lags in the data? How do you optimize the performance? What cloud tools you can use to continuously pull the data. What preprocessing on the audio stream can you do?

karan2704 commented 2 years ago

Hey, @mcshicks @valentina-s @scottveirs I'm interested in this project. I had a few questions regarding this: -Do I need some audio drivers / external sound cards to be able to work on this? -What is the role of Orcasound web app in this project? As far as I understand it is used to stream the audio data received from the archive. Should I run orcasite locally as well (if yes then what branch)? -I tried building the Dockerfile as suggested by steve on the slack channel but I face some authorization error. image -Lastly, is it okay to ask queries here on GitHub or should I stick to the slack channel

mcshicks commented 2 years ago

Hey @karan2704 probably better on the hydrophone-nodes channel in slack for questions or you can also just DM me there.. But specific to the error above you need to build the base image first which is in the orcanode/base directory. We use this image to isolate differences rpi, arm64 and amd64 docker images. You don't need the webapp to test. It's convenient to be able to listen locally, but not strictly necessary. But it seems like you are using docker on a PC so in that case you should be able to listen to the .ts files with something like vlc. Once you can listen to files locally we would need to get you an AWS key so you can use dev.orcasound.net (as documented in the orcanode readme) to test.