Using sound analysis software

Global notation can be produced using nothing but your ears, pens, and paper; but it can also take advantage of the capabilities of modern computer technology to measure certain aspects of sound precisely and present the results in a neat and professional-looking form. Sound analysis software can help identify the exact timing of onsets, even when there is no regular pulse, or the exact intervals between scale degrees, even when they are different from what your ears are accustomed to. It can reveal details of inflection and ornamentation that are difficult to detect with the unaided ear, either by generating a visual graph of the sound or simply by enabling you to play back the recording at a slower speed. Even better, software to do all these things is available to download for free—including all the programs discussed in this section of the website.

There is not (yet) a program specifically designed to produce global notation, but there are several programs that can help with various stages of the process. Each program provides its own manual or tutorials, and this website doesn’t attempt to teach you how to use them: it just discusses ways of using the software for the specific purpose of producing global notation.

It’s important to be realistic about what the software can and can’t do. No software program can produce acceptable notation by itself, or provide a substitute for listening and thinking. In many cases, sound analysis software can measure pitch accurately, but only if the music is limited to a single pitch sounding at a time. Even then, the software makes mistakes which need to be corrected manually. Conversely, some programs apply “pitch correction,” rounding each pitch to the nearest one on the keyboard, which we don’t want if we are trying to specify the pitches used in music that isn’t based on keyboard tuning.

A graph generated by a computer is not a piece of global notation in itself. It needs to be converted or annotated in ways that bring out the musical features we wish to specify. Various sound analysis programs do allow you to adapt and annotate the graph in certain ways, but you’ll have more control over the final appearance if you export the graph into a graphics program and do the annotating there.

Before running a recording through a sound analysis program, you’ll probably want to edit the sound file itself, if only to cut it down to the excerpt that you want to notate. A useful program for this is Audacity, which gives you a “picture” of the recording in the form of a waveform graph from which you can select and delete unwanted sections. (Audacity is also one of the many programs that allow you to play back the recording slowly without changing pitch, as an aid to aural transcription.)




For smoother starts and stops, select a second or so before or after the part you want to keep, open the “Effect” menu and choose “Fade In” or “Fade Out.” As you’ll see, you can also apply a wide range of other effects to the recording, some of which can make it work better with the sound analysis software in some cases. For instance, “Equalization” and “Noise Removal” can help highlight the sounds you want the software to recognize by downplaying any distracting background sounds. The only way to see about this is to experiment.

Once you have edited the sound file, it’s time to see how the software can help notate the music.


Next: Measuring onset timing with a waveform graph