Getting started with audio data

You can read, preprocess and analyze audio files in the same way as EEG or MEG files. This can for example be handy to determine the spectrogram (e.g., using a time-frequency analysis) or to determine the onset of speech.

The following reads the continuous audio data, in this case from a 4-channel WAV file. Most other audio file formats are also supported.

cfg = []; cfg.dataset = ‘4ch.wav’; data = ft_preprocessing(cfg)

The following information will be printed on screen:

processing channel { '4ch channel 1' '4ch channel 2' '4ch channel 3' '4ch channel 4' }
reading and preprocessing
reading and preprocessing trial 1 from 1

the call to "ft_preprocessing" took 1 seconds and required the additional allocation of an estimated 55 MB
data =
struct with fields:

         hdr: [1x1 struct]
       label: {4x1 cell}
        time: {[1x211355 double]}
       trial: {[4x211355 double]}
     fsample: 44100
  sampleinfo: [1 211355]
         cfg: [1x1 struct]

Also useful can be to use ft_databrowser to have a quick look at the time series.

cfg = [];
cfg.dataset = '4ch.wav';
ft_databrowser(cfg);