Analyze Long Recordings (audio2csv)

  • Command: audio2csv
  • Config file: <any compatible config file>

This command applies an analysis to a long recording. It works well with recordings that are up to 1440 minutes (i.e. 24-hours) long.

This command takes a single long-duration audio recording as input, splits it into segments and calculates results and writes the output to several .csv files. This analysis process supports parallel processing.

Analyze long recordings can run many types of analyses. The provided config file determines what kind of analysis is run.

To see a list of all analyses that can be used by audio2csv, execute:

AnalysisPrograms.exe analysesavailable

As the command name implies, the output of this analysis is a set of CSV files that can contain:

  • recognized acoustic events
  • summary indices
  • spectral indices

Usage

This section describes the command line arguments required to, for example, calculate acoustic indices derived from a single audio file.

To run the command, type:

$ AnalysisPrograms.exe audio2csv [arguments] [options]

Here is an example of a command line with abbreviated path names:

$ AnalysisPrograms.exe audio2csv "audioPath\fileName.wav" "configPath\fileName.yml" "outputPath\directoryName"

The above three paths are required arguments in that order. The program will fail if they are not found on the command line.

Reminder: All path names should be double quoted, BUT make sure you use plain double quotes ("..") and NOT so called smart quotes (“..”).

Reminder: The paths to directories should never end in a forward-slash or back-slash.

Options

Typically, analyses will also include optional arguments that take default values if they are not found on the command line.

Here is more detail about the command line options:

  • Source: The path to the audio file to process
  • Output: The output will be placed the directory provided
  • Config:
  • IMPORTANT: The name of the config file chosen will determine the kind of analysis that is run.
  • The config file contains parameters that are unique for the chosen analysis. If the config file cannot be found, AP.exe will try and use a default.
  • Typically, if a required parameter is not found in the config file, a default value will be used or in certain cases, a fatal error may result.
  • --channels: The default value is nothing (not specified) which means that all channels are used. AP.exe can process files with up to 8 channels.
  • [TODO: CHECK THIS] Channels are numbered, 0, 1, 2, … In the usual case of stereo, left channel = 0 and right channel = 1. Note that, although the software should be able to accept recordings where channel number is greater than two, this use has not been debugged and results are undefined.
  • --mix-down-to-mono: The default value is true. Typically, indices are calculated on the mixed down waveform.
  • --parallel: If you have access to a multi-core CPU you can set this option to true. Otherwise, the segments will be cut and analysed in sequence.

Use the analysis-identifier option (-a or --analysis-identifier) followed by the <analysis type> to choose the analysis to run. If you do this AP will not have to guess the name of your config file and this your config file can be named anyway you like.