What is the pitch salience? I see it a lot when looking for papers about multiple pitch detection but haven't found a proper definition of it.

  • $\begingroup$ where do you see this term, "pitch salience", referred? $\endgroup$ – robert bristow-johnson May 31 '17 at 3:32

The perception of pitch, in particular pitch salience, is not a simple idea. As recommended by Robert, if you can cite the paper where you saw the term pitch salience it would give us a chance to understand the context of where it was used and perhaps allow us to give you more succinct answers.

If you look at most textbooks on the subject of psychoacoustics (in this post I am referencing "Psychoacoustics, Facts and Models" by Fastl and Zwicker, Third Edition, Springer Publishing, 2007), you'll find descriptions of pitch perception for simple pure tones, complex tones, and bands of noise. To give you an idea of the complex nature of pitch perception, the following paragraphs are brief summaries of the ideas found in Chapter 5 of the Fastl and Zwicker textbook.

As described in Section 5.1.1 of Fastl and Zwicker, the pitch of pure tones can be measured using different procedures. In particular, measurements for "half pitch" and "double pitch". In the "half pitch" case, a subject is presented with a pure tone of frequency f1 and has to adjust the frequency f1/2 of a second tone so that the second tone produces half the pitch of the first tone. For example, if a pure tone of 440 Hz is used as sound 1 and a pure tone of variable frequency as sound 2, then the subject's average setting for the second tone is 220 Hz (halving of the pitch sensation corresponds to a ratio of 2:1 in frequency). At higher frequencies, this relationship changes. At an f1 frequency of 8 kHz, the "half pitch" frequency, or f1/2 frequency is 1,300 Hz. Measurements at other frequencies above 1 kHz confirm this general observation: for the perception of "half pitch", a ratio of the corresponding frequencies larger than 2:1 is necessary. If the Fastl and Zwicker textbook is available to you, see Fig. 5.1.

The pitch of pure tones also depends on sound pressure level. This effect can be examined by comparing the pitches of pure tones at different presentation levels. The pitch of a pure tone at a presentation level L and frequency fL is measured by matching its pitch to that of a tone with level of 40 dB and a frequency of f40 dB. For example, if a 200 Hz tone is alternatively presented at levels of 80 dB and 40 dB, the louder tone produces a lower pitch that the softer one. On the other hand, if the same experiment is performed with pure tones at 6 kHz, the effect is reversed. See Figure 5.3.

The pitch of complex tones is, well, a bit more complicated.

As described in Fastl and Zwicker, complex tones can be thought of as the sum of several pure tones. If the frequencies of the pure tones are integer multiples of a fundamental frequency, then the complex tone is typically called an harmonic complex tone. Like pure tones, the pitch of complex tones can be assessed by pitch matches with pure tones. Although complex tones contain many pure tones, they do not produce many pitches, but only one single or perhaps one prominent pitch. Here, the pitch of the complex tone is close to the pitch of the fundamental frequency or the differences between adjacent pure tones that make up the complex tone. However, the relative frequency difference between the pitch of the complex tone and the pitch of the pure tone does depend on the fundamental frequency. For example, the difference at 60 Hz is almost -3% (i.e. a pure tone, with a frequency of 58.2 Hz produces the same pitch as the harmonic complex tone with a 60 Hz fundamental frequency. With a 400 Hz fundamental frequency, the relative frequency difference is about -1% (i.e. a pure tone of 396 Hz produces the same pitch as a complex tone with a 400 Hz fundamental frequency). Above 1 kHz, the frequency of the pure tone and the fundamental frequency of the complex tone of the same pitch are equal. As with pure tones, the pitch of harmonic complex tones is also influenced by presentation level.

Unlike the harmonic complex tones for which the component frequencies are integer multiples of the fundamental frequency, complex tones can also be produced by shifting the spectral components by some amount to produce an inharmonic complex tone. With inharmonic complex tones, from which the lower components have been removed, ambiguous virtual pitches can be sometimes produced. As HOTPAW2 alluded to, some complex tones have no discernible pitch and sound like noise.

If you are interested in the role that pitch salience has in music, read Trainor et. al (2014),"Explaining the high voice superiority effect in polyphonic music: Evidence from cortical evoked potentials and peripheral auditory models", Hearing Research, Vol. 308, pages 60-70. Further, see Degani et al. (2014), "A Pitch Salience Function Derived from Harmonic Frequency Deviations for Polyphonic Music Analysis", Proceedings of the 17th International Conference on Digital Audio Effects (DAFx-14), Erlangen, Germany, September 1-5.

I hope this helps.



One dictionary definition of "salience" is whether something is particularly noticeable. Pitch is a human psychoacoustic phenomena. Some sounds might have a clearly noticeable (salient) pitch. Some sounds not or not particularly. Some maybe seem pitched, depending.

You can synthesize a (pseudo-)periodic waveform, and gradually change it from something a human would clearly identify as having a certain pitch (high pitch salience), to something that they would identify as some sort of noise or other unpitched sound. For instance: If the duration of each repeating "period" started randomly varying from period to period to greater and greater variances. Or if the number of identical periods went from 100 down to 1. Or if the "overtones" varied from all being nice exact integer multiples, to some or all of them having highly irrational and unrelated ratios to each other and the "fundamental".

  • $\begingroup$ maybe "pseudo-" should be changed to "quasi-". $\endgroup$ – robert bristow-johnson May 31 '17 at 3:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.