Heidy Suter

Formant pattern and spectral shape ambiguity in vowel synthesis: The role of fundamental frequency and formant amplitude

When investigating formant pattern and spectral shape ambiguity in Klatt synthesis, an earlier study showed that the perceived vowel quality of Standard German vowel sounds can be changed by varying fundamental frequency only [Maurer et al. (2017). J. Acoust. Soc. Am. 141(5):3469-3470]. In this follow-up study, the previous original synthesis experiment was repeated twice, firstly, with fundamental frequencies (f o) of the corresponding sounds lowered by one octave, and secondly, with different ratios of the first and second formant amplitudes. Here, the role of the f o range and the formant amplitudes for the investigation of formant pattern and spectral shape ambiguity in vowel synthesis was further examined. The same five phonetic expert listeners that participated in the previous experiment also identified all of the newly synthesised sounds in a multiple-choice identification task. Results revealed that the perceived vowel quality only changes for f os above 200 Hz and that, for back vowels, the ratio of the formant amplitudes used in the Klatt synthesis also affects vowel recognition. Thus, the results of the experiments confirm earlier indications of a non-systematic relation between f o or pitch and formant patterns or spectral envelopes for vowel recognition.When investigating formant pattern and spectral shape ambiguity in Klatt synthesis, an earlier study showed that the perceived vowel quality of Standard German vowel sounds can be changed by varying fundamental frequency only [Maurer et al. (2017). J. Acoust. Soc. Am. 141(5):3469-3470]. In this follow-up study, the previous original synthesis experiment was repeated twice, firstly, with fundamental frequencies (f o) of the corresponding sounds lowered by one octave, and secondly, with different ratios of the first and second formant amplitudes. Here, the role of the f o range and the formant amplitudes for the investigation of formant pattern and spectral shape ambiguity in vowel synthesis was further examined. The same five phonetic expert listeners that participated in the previous experiment also identified all of the newly synthesised sounds in a multiple-choice identification task. Results revealed that the perceived vowel quality only changes for f os above 200 Hz and that, for back vowels, the rati...

Sinewave vowel sounds: The role of vowel qualities, frequencies and harmonicity of sinusoids, and perceived pitch for vowel recognition

In the literature, the recognition of sinewave vowels replicating statistical formant patterns is reported as impaired when compared to natural sounds. However, the corresponding formant simulating sinusoids were harmonically unrelated, with synthesised signals only accidentally being quasi-periodic, and vowel confusion was indicated to relate to vowel height. Involving five phonetic expert listeners, the present study tested vowel and pitch recognition of three sinusoid replicas based on statistical F 1-F 2-F 3 patterns of the Standard German closed and mid-open vowels /i-y-e-o-o-u/ for women, “corrected” approximations of these patterns with harmonically related sinusoids, and harmonical patterns with fixed first and third sinusoids, yet varying only the second sinusoid so as to effect a change in harmonic relation. The results showed strong vowel confusions for mid-open but only limited confusions for closed vowels. Additional effects on vowel recognition were indicated to concern harmonicity and speci...

Formant pattern ambiguity of vowel sounds revisited in synthesis: Changing perceptual vowel quality by only changing fundamental frequency

The influence of varying fundamental frequency on the perception of vowel quality in synthesized vowels was tested in two experiments. In experiment 1, based on investigations of natural Standard German vowel sounds, various model formant patterns F1’ to F3’ were created and, for each single pattern, sounds were synthesised on two or three fundamental frequencies (range 200-600 Hz). In experiment 2, corresponding to open-tube resonance characteristics for men, women and children respectively, sounds were synthesised with formant patterns F1’ to F5’, formant frequencies being odd multiples of 500, 600, or 700 Hz and fundamental frequencies being 1/3, 1/2 or 1/1 of the first formant frequency. Five phonetic expert listeners identified all synthesised sounds in a multiple-choice identification tasks. The results of both experiments revealed that the perceived vowel quality can be changed systematically by varying fundamental frequency only and that the changes can exceed the perceptual boundaries of two neig...

“Flat” vowel spectra revisited in vowel synthesis

Some studies of natural and of synthesized vowel sounds indicate “flat” vowel-related spectral envelopes or envelope parts in terms of vowel-related frequency ranges with harmonics equal in amplitude. The present investigation addresses this question in a vowel synthesis experiment in which sounds related to series of harmonics, multiples of 200 Hz in frequency and equal in amplitude, were created. Thereby, for various frequency ranges, the number of harmonics was increased stepwise from a single lower harmonic to an increasingly broader harmonic series, and, inversely, it was also decreased from a broad series of harmonics to a single higher harmonic. The entire frequency range of investigation was 0.2-4 kHz. Vowel recognition was investigated by means of a listening test in which five phonetic expert listeners were asked to assign the synthesized sounds to Standard German vowel qualities. The results of the experiment reveal that synthesized sounds with frequency bands of series of two or more equal-amp...

Vowel synthesis related to equal-amplitude harmonic series in frequency ranges > 1 kHz combined with single harmonics < 1 kHz, and including variation of fundamental frequency

Front vowels can be synthesized on the basis of series of harmonics equal in amplitude, with frequencies only above 1 kHz. In these cases, spectral energy usually attributed to the first formant frequency is lacking. The present paper reports results of an experiment in which sound synthesis was performed on the basis of harmonic series covering higher frequency ranges above 1 kHz, combined with a single lower harmonic < 1 kHz, all harmonics equal in amplitude. Thereby, two or three sounds were synthesized for which the higher frequency range and the frequency of the lower harmonic is identical, but the frequency distance of the higher harmonics differs resulting in different perceived pitches of the sounds. Vowel recognition of all sounds was investigated by means of a listening test in which five phonetic expert listeners were asked to assign the synthesized sounds to Standard German vowel qualities. The results of the experiment reveal that the perceived vowel quality of such types of sound pairs or so...

How listeners recognise vowel sounds under highpass or lowpass filtering of vowel-specific frequency ranges

The present paper reports findings of two experiments on filtered sounds of the Standard German vowels /i-y-e-o-ɛ-a-o-u/ produced by a female speaker at two fundamental frequencies fo  = 220 Hz and 659 Hz and a male speaker at fo  = 131 Hz and 523 Hz. High-pitched sounds were included in order to account for a possible impact of the fo level on the perception of filtered vowel sounds. In the first experiment, the frequency region of the first formant of the sounds was highpass filtered, and in the second experiment, the frequency region of the second formant of the sounds was lowpass filtered. Vowel recognition of all sounds was investigated in a listening test. Results revealed shifts in the perceived vowel quality which varied across (i) vowel categories, (ii) fo , and (iii) filter cutoff frequencies. Details of the filter parameters and of the perceived vowel quality shifts are given in the paper and implications for the relationship between acoustic cues and vowel recognition are discussed.

Gender and age differences in vowel-related formant patterns: What happens if men, women, and children produce vowels on different and on similar F0?

There is a broad consensus in the literature that vowel-specific formant patterns differ as a function of gender (men/women) or age (adults/children) due to different average vocal tract sizes. Although an additional influence of fundamental frequency F0 is discussed in corresponding normalization approaches, formant patterns relating to sounds of adults and children that exhibit the same F0, to sounds of adults with higher F0 than sounds of children, and to sounds of men with higher F0 than sounds of women are barely compared. Investigating vowels of men, women, and children producing sounds with varying F0, we observed (1) a possible decrease or even a disappearance of the expected speaker-group differences in the formant frequencies 1.5 kHz was found. These observations call for a further examination of the role of F0 when interpreting speaker-group related differences in formant patterns.

Vowel discrimination at high fundamental frequencies in real speech

Previous research showed that, in singing, vowel qualities of isolated vowel sounds can be discriminated up to a fundamental frequency (F0) of about 500 Hz. However, indications are reported in literature for vowel discrimination on F0 > 500 Hz for singing (raised larynx condition, CVC context) as well as for speech-like sounds. In this study, we tested vowel discrimination at a high F0 in speech using minimal pairs build from eight long German vowels. Words were produced in speech mode at F0 of about 650 Hz by two female speakers. For all samples except the words including /a/ and /ɛ/, F0 exceeded F1 values as given in vowel statistics for Standard German. In a listening test, stimuli were played back in random order to 14 listeners (7f, 7m) for identification. The results showed that vowel discrimination can be preserved at such high fundamental frequencies. This could mean that, for our speakers and the high fundamental frequency examined, (1) source-filter-characteristics were effective up to 650 Hz, or (2) transitions played a crucial role, or (3) other spectral characteristics than formants have to be taken into account in order to explain these results.