[x] Verify all previous features work in the system with the new updates and endpoints
[x] Clean up the spectrogram generation to be more explicitly and only done in Tasks. I want them in tasks because I'm bit worried about scaling and having large audio files processed in a request. I think it may be better to make it so generation happens only in a task and removing the property reference generation of a spectrogram
[x] I believe that CompressedSpecotrgrams should be their own model that have their length, and the start/stop times as well as the duration associated with them. This makes it easer to cache them and invalidate the cache as users add more annotations to the system.
[x] Work on converting the base view to utilizing the django large image tile server instead of the base64 version
[x] Switch the system of storing the image to the S3 endpoint instead of utilizing the base64 version
[x] Going to create an admin task to generate the spectrograms automatically so I can perform the migration that way on the system instead of doing it through a regular migration. This gives me a bit more control.
Bryon's TODO: