Content Description of One-minute Segments of Audio

[towsey]

Is your feature request related to a problem?

Content Description is intended to be a new functionality. It's purpose is to describe the ecologically relevant content in one minute segments of audio.

Describe the solution you'd like

Intention is to create a new folder within the AudioAnalysisTools project. In the first instance, it will contain classes to obtain content description using five previously calculated acoustic indices as a feature description of the minute.

Describe alternatives you've considered

Did consider deriving the content description at one hour resolution using features extracted from the ribbon plots already calculated.

Additional context

In the first instance the intention is to write code to obtain proof of concept.

[atruskie]

Sounds like a great idea 👍

As we discussed offline:

• I recommend a cleanroom implementation with unit tests for the file in AudioAnalysisTools
  • Use tests, avoid sandpit
  • do not pass in any files, paths, config files, or CSV files (you may pass the strongly typed/deserialized versions of these)
  • I'd expect to see roughly
  • one method for getting a signature for a minute
  • one method for getting an aggregate signature for a collection of minute signatures
• I'd like to see a strongly type Signature class
• As we move to developing this properly, it should have its own IAnalyzer2

More notes to come