About Audio Analysers

Audio analysers are modules within PsySound3 that analyse the audio waveform. The input to an audio analyser is a sound file, and the output is data, which could be in several forms.

Calibration

Calibration is an important preparatory step for many analysers. For analysers that output sound pressure level, that value will only be accurate in absolute terms if calibration is correct. Nevertheless, the values are useable in relative terms without proper calibration. The need for calibration is more important for psychoacoustical analysers (e.g. loudness) because such algorithms do not respond in a simple way to gain adjustments.

How to Run an Analyser

Check the ‘process files using this analyser’ box. You can check many (or even all) analysers. It is advisable to obtain a time estimate of the analysis prior to running the analysis. Also, consider whether output data from analysers should be synchronised.

Windows

Many analysers analyse a succession of windows that step through the audio file from beginning to end. In some cases, the window length and overlap can be set by the user. In other cases, these are fixed.
Some analysers use windows, but this is not evident from the output (i.e., the output time series are sewn together as if the analysis process was continuous).
Some analysers may use just a single window, and do not step through the audio file.

Output Data

Single numbers and statistics
Time series
Spectra
Time-spectra

Synchronisation of Output

The output of analysers can be synchronised, meaning that output data from some analysers are resampled (usually downsampled), and window-based analysers (such as FFT Spectrum) have windows centred on each output point. The advantage of synchronising outputs is that the resulting time data from what could be a diverse set of analysers is easy to compare. The disadvantage is that some algorithms have specified analysis rates, which are resampled when synchronisation is chosen.

Post-processing

The output of audio analysers can be used in many ways within the post-processing section of PsySound3. This includes visualisation (generating visual graphs), auralisation (generating auditory graphs), data export, and further processing for analysis.

Writing New Audio Analysers

Information on how to write a new audio analyser is given somewhere…

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License