Illusion Songs

If you listen through the cycle of scrambled and unscrambled versions a few times you might even find yourself starting to hear the real Beach Boys melody, even when the octaves are all wrong.

This Tuareg vocalist from Algeria is creating a dizzying cascade of yodels that give the impression of multiple overlapping melodic lines. It's another beautiful example of pseudo polyphony, where rapid switches between notes over a certain pitch range can cause the perception of several simultaneous voices. Recorded by Alain Joset BaM 115a-1948, track courtesy of David Murray

Here are two different melodies played one after the other, immediately followed by them both at once. What do you hear when they are played together? What's Happening? On its own, neither melody gives the impression of an ascending or descending pattern. When the two melodies are played together however, most people hear this combination as being two scales that either ascend or descend in unison together. Our brain groups sounds according to similar characteristics. When there are multiple ways to group something, a further hierarchy sets in. In this case, the close proximity of interleaved pitches between the parts takes precedence over individual melodies, creating the illusion of a third melodic trajectory. Interestingly, this illusion continues to occur over headphones when the two parts are simultaneously introduced, one to each ear. The Scale Illusion was discovered by Diana Deutsch. This video is the second part in a new series of auditory illusion demonstrations adapted for musical instruments and voice. Some of this work appeared originally in a performance at Davies Symphony Hall for Pop-Up Magazine.

A Shepard Tone appears to continually rise or fall in pitch, while actually staying within a set range that cycles around. How does this work? In the simplest demonstration of the phenomenon, each note is actually made up of many copies of itself in different octaves. As the pitch rises with the scale, the lower octaves of each note fade down in volume and the higher octaves fade up. At the top of the scale the sequence loops around and begins again at the bottom. Because we are busy tracking these changes in volume from note to note rather than hearing the sequence as a whole, the jump back to the bottom note is masked. Jean-Claude Risset subsequently made a continuous 'glissando' version of this illusion (known as the Shepard-Risset Glissando), as well as a rhythmic equivalent, called the Risset Rhythm. Diana Deutsch's Tritone Paradox also uses a similar concept. This demonstration was made by Roger Shepard and Edward Zajac at Bell Labs in 1967. It is accompanied by its visual equivalent, the Penrose Stairs

The Wessel Illusion demonstrates how timbre can determine the way in which we perceptually group notes in a melody. Three notes, rising in pitch but alternating in timbre, are played slowly. When this sequence is played faster, it's possible to hear the trajectory of the melody change. What's going on? When the notes are played slowly, most people hear a sequence that is rising in pitch. This is because there is enough available attention to track both the sound of the voice and the vibraphone as distinct timbres, and then group them together into one ascending melody. When the melody speeds up, it's no longer possible to track everything at once. At a certain point we are forced to perceptually cut our losses--we begin to group the notes solely based on their timbre. This creates the illusion of a descending sequence instead. The Wessel illusion was discovered by David Wessel of U.C. Berkeley's Center for New Music and Audio Technologies (CNMAT). This video is the first part in a new series of auditory illusion demonstrations adapted for musical instruments and voice. Some of this work appeared originally in a performance at Davies Symphony Hall for Pop-Up Magazine.

Silbo Gomero, a whistled language native to the Canary Islands is finding a resurgence and being taught in local schools again Silbo is based on a spoken dialect that is fundamentally non-tonal . This means that the pitches of spoken words or syllables are only used for inflection or to convey emotion--they don't change the inherent meaning of the word. The whistled version of Silbo is created by tracking the second formant of the vowel sounds found in the spoken version of the language. This means that the normal shapes made by the vocal tract when pronouncing the spoken language are preserved as much as possible during the whistled version. These whistled melodies retain and mimic the harmonic relationships that characterize the transitions between spoken vowel sounds. This process of approximation is strikingly similar to the one used by Yale's Haskins' Lab to create sinewave speech which I posted about previously. Consonants, physically hard to reproduce during a whistle, are replaced by a variety of techniques, including pauses, rising or falling tones, as well as a complex process of designating transitions between various formants as different types of consonants. As the Haskins' sinewave speech demonstrates, consonants are not always necessary in the comprehension of sentences if enough formant information is present. You can read a basic translation of this video here.

Recording: Dense Boogie from Sound Characters by Maryanne Amacher, Tzadik Records, 1999

Recording: Primitive Music Of The World, Selected and Edited by Henry Cowell, Ethnic Folkways Library FE 4581

Recording: Panpipe Ensemble, Olga Velitchkina, Ohio University.

This stunning demonstration of Tuvan throat singing is the result of tuning the resonances of the vocal tract. By creating a single low fundamental note with a rich harmonic spectrum, singers are able to enhance certain harmonics and minimize others. This gives the mistaken impression that more than one note is being sung at once. In this case, the two voices are also interfering, creating beats and difference tones. A particularly amazing demonstration of this starts at around 1:00. If you're having trouble hearing what I'm talking about in the context of the song here's a more clinical version of the same difference tone phenomenon. For more technical information about harmonic singing, the University of New South Wales has an introductory page here. Thanks to Thor Magnusson for sharing the video.

The concept of beats is one of the most fundamental in acoustics. First order beats occur when two notes are of slightly different frequencies. Because the frequencies are so close together their phases become just out of line with one another. They then interfere constructively and destructively, causing a modulation of the amplitude. The beat frequency is equal to the distance between the two notes. The "roughness" that beating causes can alter our ability to perceive pitch. Other types of beats, such as second-order and binaural beats are themselves an illusion. Italian composer Pietro Grossi's 1965 album Battimenti is solely devoted to different beat ratios as produced on analog synthesizers. According to his liner notes: "the album is a collection of "94 sets of beats with 2, 3, 4, 5 frequencies taken from a catalogue of sound events...meant to be used for various compositional purposes. Each set of beats lasts approximately 30 seconds; they are arranged as follows: 10 sets with two frequencies, 25 with three frequencies, 31 with four frequencies, 28 with five frequencies. Listening at low level is suggested". The beginning of the first track is excerpted above. Recording: Battimenti A Due Frequenze, Battimenti Ants Records 2003

Hollerin’ is a form of yodeling, one of the oldest forms of long-distance communication. Often using the natural acoustics or resonances of a space to reinforce and carry the message, yodelers are also adept at transitioning between vocal registers, focusing on and emphasizing the break that occurs at the boundary rather than smoothing it. This creates a unique timbre in the middle of a rapid leap from low to high that further disrupts our ability to track the melody and maintain auditory streams, often resulting in a kind of pseudo-polyphony. This 1975 performance is from the still yearly contest in Spivey’s Corner, North Carolina. Recording: Old-timey Holler/Ditty by Leonard Emmanuel, Hollerin', Rounder Records, 1995

Swiss composer Andreas Stahel creates solo hockets for voice and flute. Using circular breathing he is able to sing and blow simultaneously for abnormally long periods of time. Stahel's hockets are composed of at least three different streams on the basis of timbre alone: the sound of his voice, the sound of the flute, and the altered sound of his voice when he sings into the flute. In turn, each of these timbral streams is also broken apart by pitch range--each part makes huge leaps from low to high, further muddying our ability to track who is making what, and giving the impression of multiple disparate sound sources. As is especially demonstrated in the second piece he plays (see 1:35), rapid repetition can also play a role. When rhythmic or melodic sequences are established by repetition it is easier for our brain to begin to expect and fill in missing parts. For another example of this, see my recent post on beatboxing and continuity illusions. Thanks to my friend Neil Young for the tip-off.

This highlights a really interesting part of sound perception: the way that our ears perceive and process information individually in isolation (like when you're wearing headphones) can be different than how they perform together in the open air. In this case the headphones and hard panning to left and right creates the functional sensation of two sound sources with very different locations. How well can you hear the effect without headphones?

This solo Russian panpipe song from the Kursk province is produced by a combination of blowing techniques, head movements, and simultaneous vocal sounds. This creates an erie timbre which seems to be made up of more than the sum of its parts--an incredible hybrid of pipe and vocal resonances. You can hear more examples of uniquely altered vocal timbres here. Ethnomusicologist Olga Velitchkina has also done fascinating research on the role of movement and speech sounds in the production of this tradition of music. You can read about it and hear more examples here. Recording: "Solo close-up showing breathing movements" recorded by Olga Velitchkina, Ohio University.

Beat-boxing champion Rahzel creates an astonishing display of simultaneous singing and vocal percussion—everything you hear is coming out of one mouth!  In order to make this happen, he's actually dropping parts of words and masking them with percussive bursts of noise. This allows our brains to recreate the missing speech information based on our expectation of a continuous sentence.  A continuity illusion like this occurs when missing information from a progression we perceive as ongoing is masked by bursts of noise containing at least some of the missing frequencies. While one of the challenges in beatboxing is to create percussive timbres that don't sound like they originate in the vocal tract, this can actually work to the illusion's advantage.  When a percussive noise originates in the same part of the vocal tract as the missing speech sound, this can aid in our reconstruction of the word.

Recording: Musique du Burundi, Ocora Records, 1968.