21. Two other ways to visualize sound signals

     As promised in the preceding page, this page will have only one external link. That is because it is a complete tutorial in itself, and it is more than enough material for one page.

     The page linked to below is a speech analysis tutorial created by the Department of Speech and Linguistics of Lund University in Sweden. It will help you review some of the methods of representing speech graphically that you are already familiar with: oscillograms, or waveforms; fundamental frequency (F0) analysis, or pitch tracks; spectrograms; and phonetic transcription. It will in addition introduce two new methods of visualizing a sound signal: the waterfall spectrogram and the spectrum (plural: spectra).

      The waterfall spectrogram contains the same information as a regular spectrogram; it is simply displayed in a different manner. You will see amplitude expressed as the peaks of what looks like a very knobby mountain range.

     The word "spectrum" is familiar to us, from our study of the color spectrum of light; and it also sounds a lot like "spectrogram". But don't confuse "spectrum" with "spectrogram". They represent two very different things. A spectrogram could be compared to a video movie, and a spectrum to a still photograph. A spectrogram gives a running display of a sound signal as it occurs in real time; a spectrum, on the other hand, gives us a snapshot of the sound at a specific point in time. A spectrum can enable you to see, for example, the energy distribution over the different frequencies of a single vowel, like [i]. You will see that the spectrum has no time scale; frequency (shown on the horizontal axis, or abscissa), measured in Hertz, and amplitude (on the vertical axis, or ordinate), measured in decibels, are its only parameters.

      Note that the kind of spectrum presented here is referred to more precisely as a short-term spectrum; this is contrasted with a long-term spectrum, in which the overall energy distribution of a sound signal of any length can be 'summarized' in one graphic image. And there are yet other ways to represent a sound spectrum which you may run into later in your study of phonetics.

     Link to the speech analysis tutorial here:

http://www.ling.lu.se/research/speechtutorial/tutorial.html


     Were you able to relate what you saw on this page with things we have covered in class?


     So far, WASP has been adequate for our speech analysis needs, but it cannot, for example, display a sound spectrum, so it's time to move on to more advanced speech analysis tools. The next few pages will introduce some of these tools and tell you where to get them and how to use them.

Next: Advanced speech analysis tools I: SFS

on to next page        back        index I        index II         home