Re: Lower frequency limit for pitch perception

[Daniel Pressnitzer <Daniel.Pressnitzer@xxxxxxxx>]

Dear Paul,

 We just collected some data on the lower limit of pitch at the CNBH
in Cambridge, which you might be interested to hear about.

 In order to differentiate between pitch and other percepts
(e.g. roughness, whooshing, etc...), we asked listeners to perform a
simple melody task: if they could do it, we assumed they heard a
pitch. The task was as follows. First, a 4-note chromatic melody was
played with notes randomly chosen within a 4-semitone range.  The
lowest note was referred to as "the base note". Then, the melody was
repeated with a semitone change introduced on one of the notes. The
task was to indicate which note had changed. An adaptive procedure was
used to measure the lowest fundamental frequency of the base note for
which the task was possible. We called this frequency the lower limit
of melodic pitch (LLMP).

 We found that the LLMP was as low as 30Hz for broadband click trains.
The LLMP is approximately the same for bandpass-filtered harmonic
complexes with a lower cutoff less than 800Hz. When the lower cutoff
is above 800Hz and some low-pass pink noise is added to mask
distortion products, the LLMP increases. It is around 100Hz for a
lower cutoff of 1.6kHz, and higher than 250Hz for a cutoff of 3.2kHz.


Cheers,

Daniel


Pressnitzer, Patterson and Krumbholz (1999). The lower limit of
melodic pitch with filtered harmonic complexes. J. Acoust. Soc.
Am. 105, p. 1152 [This is an abstract. A paper has been submitted
since...]

A useful reference, in addition to the ones that have already been
mentioned:

Ritsma (1962). The existence region of the tonal residue I.
J. Acoust. Soc. Am. 34, p. 1224-1229