The perception of a synthetic nasal consonant is affected by the formant structure of a preceding vowel.

Below you can find audio examples of stimuli used in three experiments that were performed to examine the importance of formant continuity between vowel and nasal. Click on the vowel F2 frequency table under each spectrogram to play the stimuli. (Table entries shaded red tend to be heard as /m/ by listeners; those shaded blue tend to be heard as /n/).

For more information, see publications below.

Experiment 1 - no explicit transitions

When no explicit transitions are present between vowel and nasal, the vowel F2 frequency affects the perception of the nasal consonant.

For the experiment, vowel F2 frequency varies between 600 and 2200 Hz in steps of 200 Hz; other vowel formants remain constant.
Nasal prototypes /m/ and /n/ differ in their formant frequencies.

Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; nasal prototype /m/
	Stimuli with nasal prototype /m/ (click to play)
Vowel F2 (Hz)	600	800	1000	1200	1400	1600	1800	2000	2200
Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; nasal prototype /n/
	Stimuli with nasal prototype /n/ (click to play)
Vowel F2 (Hz)	600	800	1000	1200	1400	1600	1800	2000	2200

Nasal percept changes as vowel F2 frequency changes: /m/ is heard for low vowel F2, /n/ for high vowel F2, whatever the nasal prototype.

Experiment 2 - transitions of varying duration

When transitions of sufficient duration are present between vowel and nasal, the nasal percept corresponds to the nasal prototype.

Vowel F2 frequency takes values 800 and 2000 Hz; transitions occur from vowel formants to nasal formants; transition lengths are 0, 10, 20 and 40 ms.
Nasal prototypes /m/ and /n/ are as for experiment 1.

Sample spectrogram for vowel F2=2000 Hz; transition lengths 10ms and 40 ms; nasal prototype /m/
	Stimuli with nasal prototype /m/ (click to play)
	Vowel F2 = 800 Hz					Vowel F2 = 2000 Hz
Transition length (ms)	0	10	20	40		0	10	20	40
Sample spectrogram for vowel F2=800 Hz; transition lengths 10ms and 40 ms; nasal prototype /n/
	Stimuli with nasal prototype /n/ (click to play)
	Vowel F2 = 800 Hz					Vowel F2 = 2000 Hz
Transition length (ms)	0	10	20	40		0	10	20	40

Nasal percept corresponds to that of the nasal prototype for longer transition lengths.

Experiment 3 - transitions to a target frequency

When transitions from the vowel F2 to a target frequency (not necessarily corresponding to a nasal formant frequency) are present between vowel and nasal, the target frequency determines the perception of the nasal consonant.

Vowel F2 frequency varies between 600 and 2200 Hz in steps of 400 Hz.
Transitions from the vowel F2 point to target frequencies 1000 and 2000 Hz.
Nasal prototypes /m/ and /n/ are as for experiment 1.

Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; target 1000 Hz; nasal prototype /m/
Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; target 2000 Hz; nasal prototype /m/
	Stimuli with nasal prototype /m/ (click to play)
	Transition target 1000 Hz						Transition target 2000 Hz
Vowel F2 frequency (Hz)	600	1000	1400	1800	2200		600	1000	1400	1800	2200
Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; target 1000 Hz; nasal prototype /n/
Sample spectrogram for vowel F2=600, 1400 and 2200 Hz; target 2000 Hz; nasal prototype /n/
	Stimuli with nasal prototype /n/ (click to play)
	Transition target 1000 Hz						Transition target 2000 Hz
Vowel F2 frequency (Hz)	600	1000	1400	1800	2200		600	1000	1400	1800	2200

Nasal percept corresponds to the target vowel F2 frequency: /m/ is heard for target 1000 Hz; /n/ for target 2000 Hz.

Conclusions

• The target frequency of the vowel F2 transition into the nasal consonant determines the nasal percept: a target of around 1000 Hz gives an /m/ percept, while a target of around 2000 Hz gives an /n/ percept.
• The formant structure of the nasal prototype has little effect on the nasal percept when it is preceded by a vowel.
• The nasal percept does not depend on continuity between the formants of the vowel and nasal consonant.

• Harding S. and Meyer G. (2003), Changes in the perception of synthetic nasal consonants as a result of vowel formant manipulations , Speech Communication 39(3-4):173-189.
• Harding S.M. and Meyer G.F. (2001), Formant continuity and auditory scene analysis: The effect of vowel formant manipulations on the perception of synthetic nasal consonants. Journal of the Acoustical Society of America 109(5):2312. (Abstract; corresponding poster)

• This work was carried out at Keele University; SMH was supported by EPSRC research studentship award no. 99304828

Last updated 2/6/04