A musical pitch is defined by two dimensions, pitch quality (which we describe with a note name) and pitch height (the octave of the note). In 1964, Roger Shepard created tones with clear pitch quality (note) but ambiguous pitch height (octave). Use to hear a scale made up of these tones. First, listen to the scale and to the continuous slide devised by Jean-Claude Risset. Unless you listen closely, they seem to continuously rise or fall. (This demonstration works best with headphones or speakers with good bass response.) Next, check the box for show spectrogram to see how these tones are made.
Diana Deutsch uses ambiguous tones to investigate the perception of absolute pitch height (Deutsch, 1987; Deutsch et al., 1987). You can do an abbreviated version of one of her experiments using .
What did this experiment test? Each trial presented a tritone and asked whether it seemed to rise or fall. A tritone is a pair of notes that are exactly half an octave apart, which is the most distant two pitches can be in quality. Since the experiment used ambiguous tones similar to those devised by Shepard, neither tone was objectively higher than the other. If our perception of pitch were simply based on the frequency content of the tones, we would not perceive the tritone pairs as either rising or falling. This experiment tested whether you subjectively rank some notes as higher than others based on pitch quality alone.
In fact, nearly everyone consistently hears some pitches as being higher than others, independent of pitch height. At the end of the experiment, you saw a circle of pitches based on your data; the ones at the top were judged to be higher than the others. The one at the top is your peak pitch height. One surprising finding from Deutsch’s work is that one’s peak pitch height seems to be related to the language and regional dialect that one grew up hearing. See for data from three subjects from different English-speaking countries.
If you do not have time to do the experiment, use to experience the tritone pairs. This makes a fun demonstration with a group of people. Click the tritone buttons at random and ask everyone whether each pair seems to rise or fall. They will be surprised to see disagreement on the matter. (However, if you listen carefully, you may hear a pair rising and falling at the same time.)
Although each of us has a bias to hear some pitches as higher than others, that perception can be modified by prior stimuli. Again using , check the box labeled loop. This will go through all the tritones in order. You will probably hear all of them rising or all of them falling. After going through all 12 pairs, click the button for a tone pair directly opposite the one that is currently playing (e.g., if B-F just played, click F-B). Most people will suddenly hear the pairs reverse; if they were rising before, they will now be falling, and will stay that way for the entire cycle of 12 tritones. Thus you can hear each of the 12 tone pairs as either ascending or descending, depending on context!
Tritone stimuli used here were generated as described by Deutsch et al. (1987). However, to adapt the experiment to computer and keep its duration reasonable, the manner of presentation has been altered. To replicate Deutsch’s experiments exactly, use her audio CD (Deutsch, 1995), available along with other clever musical illusion CDs at www.philomel.com.
To play Shepard scale and Risset slide without graphics, use , , , and (loops; click the link to start and stop the sound).