Lower F0s Research Articles

Vocal pitch perception with cochlear implants

Cochlear implant (CI) users often struggle to accurately perceive vocal pitch. This suggests that they are either not able to adequately use vocal pitch cues, such as temporal modulations resulting from F0, or they are using cues that do not correspond to vocal pitch. A potentially unreliable cue they may be using is formant spacing, which is correlated with vocal tract length. There is a gross, probabilistic relationship between vocal pitch and formant spacing since males tend to have both lower F0s and longer vocal tracts, but the relationship is weak, especially within gender. To examine if CI users are using the acoustic cues resulting from vocal tract length to make pitch judgments, CI users were presented with stimuli where the F0 was held constant but the formant spacing was systematically manipulated. Participants were presented with pairs of stimuli and asked to indicate the similarity of the pitch of the two stimuli using a seven-point Likert scale. The preliminary results indicate that CI users systematically perceived the changes in formant spacing as a change in pitch.

Perception of melodies and triads at high frequencies

Accurate pitch perception is possible for harmonic complex tones with fundamental frequencies (F0s) in the musical range ( 6 kHz). However, it is unknown whether this basic pitch perception supports multiple simultaneous pitch perception. To address this, we measured (1) melody discrimination with and without a complex-tone masker and (2) major-minor discrimination for triads and arpeggios composed of complex tones with low (∼280 Hz) or high (∼1400 Hz) F0s. The tones were filtered to ensure that in high-F0 conditions only harmonics beyond the limits of phase locking were audible. Melody perception was poorer for isolated high-F0 tones than for isolated low-F0 tones, although performance was above chance in both cases. Adding a complex-tone masker in the same spectral region degraded performance for low- and high-F0 tones. Listeners could discriminate major and minor triads and arpeggios for low-F0 tones. For high-F0 tones, some listeners could discriminate major and minor arpeggios but none could discriminate major and minor triads. These results will help elucidate whether different mechanisms underlie the perception of combinations of complex tones at low and high frequencies. [Work supported by NIH R01DC005216 and NSF NRT-UtB1734815.]

Effect of lowest harmonic rank on fundamental-frequency difference limens varies with fundamental frequency.

This study investigated the relationship between fundamental frequency difference limens (F0DLs) and the lowest harmonic number present over a wide range of F0s (30-2000 Hz) for 12-component harmonic complex tones that were presented in either sine or random phase. For fundamental frequencies (F0s) between 100 and 400 Hz, a transition from low (∼1%) to high (∼5%) F0DLs occurred as the lowest harmonic number increased from about seven to ten, in line with earlier studies. At lower and higher F0s, the transition between low and high F0DLs occurred at lower harmonic numbers. The worsening performance at low F0s was reasonably well predicted by the expected decrease in spectral resolution below about 500 Hz. At higher F0s, the degradation in performance at lower harmonic numbers could not be predicted by changes in spectral resolution but remained relatively good (<2%-3%) in some conditions, even when all harmonics were above 8 kHz, confirming that F0 can be extracted from harmonics even when temporal envelope or fine-structure cues are weak or absent.

Glottal Source Contribution to Higher Order Modes in the Finite Element Synthesis of Vowels

Articulatory speech synthesis has long been based on one-dimensional (1D) approaches. They assume plane wave propagation within the vocal tract and disregard higher order modes that typically appear above 5 kHz. However, such modes may be relevant in obtaining a more natural voice, especially for phonation types with significant high frequency energy (HFE) content. This work studies the contribution of the glottal source at high frequencies in the 3D numerical synthesis of vowels. The spoken vocal range is explored using an LF (Liljencrants–Fant) model enhanced with aspiration noise and controlled by the R d glottal shape parameter. The vowels [ɑ], [i], and [u] are generated with a finite element method (FEM) using realistic 3D vocal tract geometries obtained from magnetic resonance imaging (MRI), as well as simplified straight vocal tracts of a circular cross-sectional area. The symmetry of the latter prevents the onset of higher order modes. Thus, the comparison between realistic and simplified geometries enables us to analyse the influence of such modes. The simulations indicate that higher order modes may be perceptually relevant, particularly for tense phonations (lower R d values) and/or high fundamental frequency values, F 0 s. Conversely, vowels with a lax phonation and/or low F0s may result in inaudible HFE levels, especially if aspiration noise is not considered in the glottal source model.

Pitch perception of concurrent high-frequency complex tones

Accurate pitch perception is possible for harmonic complex tones with fundamental frequencies (F0s) in the musical range (e.g., 1.4 kHz) but with all harmonics beyond the putative limits of phase locking. However, it is unknown whether pitch perception in more complex scenarios, such as with concurrent complex tones, is possible using these stimuli. To address this question, we measured (1) F0 difference limens (F0DLs) and (2) target-to-masker ratios (TMRs) required to detect a fixed F0 difference in a mixture of complex tones with low F0s (∼280 Hz) or high F0s (∼1400 Hz). The target tones were filtered to ensure that in the high-F0 case, only harmonics beyond the limits of phase locking were present. Pitch perception was poorer for isolated high-F0 tones than for isolated low-F0 tones and adding a masker complex tone with a geometrically centered F0 impaired performance for both high-F0 and low-F0 tones. The TMRs required to achieve good performance in the presence of two complex tone maskers were higher for high-F0 tones than for low-F0 tones. The results should help determine whether different mechanisms underlie the perception of combinations of complex tones at low and high frequencies. [Work supported by Grants NIH R01 DC005216 and NSF NRT-UtB1734815.]Accurate pitch perception is possible for harmonic complex tones with fundamental frequencies (F0s) in the musical range (e.g., 1.4 kHz) but with all harmonics beyond the putative limits of phase locking. However, it is unknown whether pitch perception in more complex scenarios, such as with concurrent complex tones, is possible using these stimuli. To address this question, we measured (1) F0 difference limens (F0DLs) and (2) target-to-masker ratios (TMRs) required to detect a fixed F0 difference in a mixture of complex tones with low F0s (∼280 Hz) or high F0s (∼1400 Hz). The target tones were filtered to ensure that in the high-F0 case, only harmonics beyond the limits of phase locking were present. Pitch perception was poorer for isolated high-F0 tones than for isolated low-F0 tones and adding a masker complex tone with a geometrically centered F0 impaired performance for both high-F0 and low-F0 tones. The TMRs required to achieve good performance in the presence of two complex tone maskers were higher...

Frequency difference limens as a function of fundamental frequency and harmonic number

Several studies have investigated the relation between the lowest harmonic present in a complex tone and fundamental frequency (F0) difference limens (F0DLs). It is generally assumed that F0DLs are smaller when lower harmonics are present and that the ability to discriminate small changes in F0 worsen as harmonic number increases. This worsening of performance has been attributed to a lack of peripherally resolved harmonics. This assumption was tested by measuring F0DLs for harmonic complexes where the lowest harmonic present in a twelve-harmonic complex tone varied from the 3rd to the 15th harmonic, with F0s varying from 30 Hz to 2000 Hz. The harmonics were presented in either sine or random phase and were embedded in threshold-equalizing noise. Aside from F0s between 100 and 400 Hz, performance did not follow the expected pattern of good performance with low-numbered (resolved) harmonics and poorer performance with high-numbered (unresolved) harmonics. At lower F0s, performance was relatively constant a...

The Perception of Formant Tuning in Soprano Voices

At the upper end of the soprano range, singers adjust their vocal tract to bring one or more of its resonances (Rn) toward a source harmonic, increasing the amplitude of the sound; this process is known as resonance tuning. This study investigated the perception of (R1) and (R2) tuning, key strategies observed in classically trained soprano voices, which were expected to be preferred by listeners. Furthermore, different vowels were compared, whereas previous investigations have usually focused on a single vowel. Listeners compared three synthetic vowel sounds, at four fundamental frequencies (f0), to which four tuning strategies were applied: (A) no tuning, (B) R1 tuned to f0, (C) R2 tuned to 2f0, and (D) both R1 and R2 tuned. Participants compared preference and naturalness for these strategies and were asked to identify each vowel. The preference and naturalness results were similar for /ɑ/, with no clear pattern observed for vowel identification. The results for /u/ showed no clear difference for preference, and only slight separation for naturalness, with poor vowel identification. The results for /i/ were striking, with strategies including R2 tuning both preferred and considered more natural than those without. However, strategies without R2 tuning were correctly identified more often. The results indicate that perception of different tuning strategies depends on the vowel and perceptual quality investigated, and the relationship between the formants and (f0). In some cases, formant tuning was beneficial at lower f0s than expected, based on previous resonance tuning studies.

Re-examining the upper limit of temporal pitch.

Five normally hearing listeners pitch-ranked harmonic complexes of different fundamental frequencies (F0s) filtered in three different frequency regions. Harmonics were summed either in sine, alternating sine-cosine (ALT), or pulse-spreading (PSHC) phase. The envelopes of ALT and PSHC complexes repeated at rates of 2F0 and 4F0. Pitch corresponded to those rates at low F0s, but, as F0 increased, there was a range of F0s over which pitch remained constant or dropped. Gammatone-filterbank simulations showed that, as F0 increased and the number of harmonics interacting in a filter dropped, the output of that filter switched from repeating at 2F0 or 4F0 to repeating at F0. A model incorporating this phenomenon accounted well for the data, except for complexes filtered into the highest frequency region (7800-10 800 Hz). To account for the data in that region it was necessary to assume either that auditory filters at very high frequencies are sharper than traditionally believed, and/or that the auditory system applies smaller weights to filters whose outputs repeat at high rates. The results also provide evidence on the highest pitch that can be derived from purely temporal cues, and corroborate recent reports that a complex pitch can be derived from very-high-frequency resolved harmonics.

Accuracy of six techniques for measuring formants in isolated words

In previous work comparing formant measurements of synthetic vowels, manual measurements and four of the six algorithms tested were shown to have large errors and be biased by fundamental frequency (F0). In this study, natural speech recorded by five speakers (2 female, 3 male) was used to provide a more realistic test. The context /hVd/ was used with vowels V = {/i, ae, u, a/}, in which formant frequencies could be expected to be relatively constant; declarative intonation was elicited so that F0 would be falling throughout the vowel, eliciting evidence of F0 bias. Formants were estimated throughout using LPC Burg, LP closed-phase covariance, Weighted Linear Prediction with Attenuated Main Excitation (WLP-AME) [Alku et al., JASA 134(2), 1295–1313 (2013)], AVG, CEPS, and reassigned spectrogram with pruning (RS). The LPC methods show evidence of F0 bias, with WLP-AME having the smallest errors (Alku et al. compared it to many LPC variants, including DAP); RS has the smallest errors overall, but is not fully automatic. Accuracy during vocal fry and for low vs. high F0s will be discussed. As with synthetic speech, WLP-AME can be recommended as the analysis of choice until RS becomes more automatic and less subjective. [Work supported by NIH.]

Do Red Deer Stags (Cervus elaphus) Use Roar Fundamental Frequency (F0) to Assess Rivals?

It is well established that in humans, male voices are disproportionately lower pitched than female voices, and recent studies suggest that this dimorphism in fundamental frequency (F0) results from both intrasexual (male competition) and intersexual (female mate choice) selection for lower pitched voices in men. However, comparative investigations indicate that sexual dimorphism in F0 is not universal in terrestrial mammals. In the highly polygynous and sexually dimorphic Scottish red deer Cervus elaphus scoticus, more successful males give sexually-selected calls (roars) with higher minimum F0s, suggesting that high, rather than low F0s advertise quality in this subspecies. While playback experiments demonstrated that oestrous females prefer higher pitched roars, the potential role of roar F0 in male competition remains untested. Here we examined the response of rutting red deer stags to playbacks of re-synthesized male roars with different median F0s. Our results show that stags’ responses (latencies and durations of attention, vocal and approach responses) were not affected by the F0 of the roar. This suggests that intrasexual selection is unlikely to strongly influence the evolution of roar F0 in Scottish red deer stags, and illustrates how the F0 of terrestrial mammal vocal sexual signals may be subject to different selection pressures across species. Further investigations on species characterized by different F0 profiles are needed to provide a comparative background for evolutionary interpretations of sex differences in mammalian vocalizations.

Neural Representation of Harmonic Complex Tones in Primary Auditory Cortex of the Awake Monkey

Many natural sounds are periodic and consist of frequencies (harmonics) that are integer multiples of a common fundamental frequency (F0). Such harmonic complex tones (HCTs) evoke a pitch corresponding to their F0, which plays a key role in the perception of speech and music. "Pitch-selective" neurons have been identified in non-primary auditory cortex of marmoset monkeys. Noninvasive studies point to a putative "pitch center" located in a homologous cortical region in humans. It remains unclear whether there is sufficient spectral and temporal information available at the level of primary auditory cortex (A1) to enable reliable pitch extraction in non-primary auditory cortex. Here we evaluated multiunit responses to HCTs in A1 of awake macaques using a stimulus design employed in auditory nerve studies of pitch encoding. The F0 of the HCTs was varied in small increments, such that harmonics of the HCTs fell either on the peak or on the sides of the neuronal pure tone tuning functions. Resultant response-amplitude-versus-harmonic-number functions ("rate-place profiles") displayed a periodic pattern reflecting the neuronal representation of individual HCT harmonics. Consistent with psychoacoustic findings in humans, lower harmonics were better resolved in rate-place profiles than higher harmonics. Lower F0s were also temporally represented by neuronal phase-locking to the periodic waveform of the HCTs. Findings indicate that population responses in A1 contain sufficient spectral and temporal information for extracting the pitch of HCTs by neurons in downstream cortical areas that receive their input from A1.

Training listeners to report fundamental frequency and formant range information independently.

The vowels of speakers of different sizes vary in terms of their average f0 and formant frequencies. In general, larger speakers produce vowels with lower formant frequencies and lower f0s. Listeners have demonstrated the ability to estimate the approximate size of a speaker and previous experiments have shown that these judgments are based on the joint consideration of f0 and formant range information. Thus both lower f0s and lower formant frequency ranges are associated with larger speakers by listeners. Studies which have asked listeners to evaluate voices have focused on the extraction of apparent speaker characteristics (which are informed by f0 and formant frequencies) rather than asking speakers to report f0 and formant range information directly. The current study consists of a training procedure by which participants will learn to report the f0 and formant range of voices independently. The training consists of a voice matching game in which participants hear a pair of vowels produced by a voice and are asked to indicate which of the candidate voices they just heard. Results will be analyzed in light of current theories of vowel perception and normalization.

Identification of speaker sex from one vowel across a range of fundamental frequencies

To identify a speaker's sex, listeners may rely on sex-based differences in average fundamental frequency (F0), but overlap in male and female F0 ranges undermines such judgments. To test accuracy of sex-identification throughout the F0 range, listeners were asked to judge sex based on audio recordings of /ɑ/ spoken on a number of overlapping steady F0s by 10 male and 10 female English speakers. In general, listeners performed above chance (71.6% correct). However, near range extrema, listeners followed an apparent bias toward hearing high F0s as female and low as male; confidence was high when accuracy was high and vice-versa. At mid-range, listeners identified sex fairly accurately but were not very confident in their judgments. In a forced-choice task, vowels close in F0 (but beyond the difference limen) were presented in male-female or female-male pairs. Listeners weakly identified speaker sex (63.3% correct). Identification of the male voice was considerably above chance only when the male had the lower F0 of the pair. Reliance on stereotypes of speaking F0 may bias listeners to hear low F0s as male and high F0s as female, perhaps with a contribution from vocal-tract length information. No strong evidence for a contribution of voice quality obtained.

Detection and F discrimination of harmonic complex tones in the presence of competing tones or noise

Normal-hearing listeners' ability to "hear out" the pitch of a target harmonic complex tone (HCT) was tested with simultaneous HCT or noise maskers, all bandpass-filtered into the same spectral region (1200-3600 Hz). Target-to-masker ratios (TMRs) necessary to discriminate fixed fundamental-frequency (F0) differences were measured for target F0s between 100 and 400 Hz. At high F0s (400 Hz), asynchronous gating of masker and signal, presenting the masker in a different F0 range, and reducing the F0 rove of the masker, all resulted in improved performance. At the low F0s (100 Hz), none of these manipulations improved performance significantly. The findings are generally consistent with the idea that the ability to segregate sounds based on cues such as F0 differences and onset/offset asynchronies can be strongly limited by peripheral harmonic resolvability. However, some cases were observed where perceptual segregation appeared possible, even when no peripherally resolved harmonics were present in the mixture of target and masker. A final experiment, comparing TMRs necessary for detection and F0 discrimination, showed that F0 discrimination of the target was possible with noise maskers at only a few decibels above detection threshold, whereas similar performance with HCT maskers was only possible 15-25 dB above detection threshold.

The Effects of Fundamental Frequency Level on Voice Onset Time in Normal Adult Male Speakers

The purpose of this study was to examine the effect of fundamental frequency (F(0)) on stop consonant voice onset time (VOT). VOT was measured from the recordings of 56 young men reading phrases containing all 6 English voiced and voiceless stops in word-initial position across high-, medium-, and low-F(0) levels. Separate analyses of variance for the voiced and voiceless stops revealed no significant main effect for F(0) for the voiced stops but a significant F(0) effect for the voiceless stops. Across the voiceless stops, productions at high F(0)s displayed significantly shorter VOTs than productions at low or mid F(0)s. The findings indicated that researchers must take into account the F(0) level at which voiceless stop VOT is measured.

Fundamental frequency discrimination and harmonic resolvability in sensorineural hearing impariment

Sensorineural hearing loss (SNHL) often results in impaired fundamental frequency (F0) processing. Three experiments tested the hypothesis that this deficit is related to a loss of frequency selectivity, which may result in fewer peripherally resolved harmonics. F0 difference limens (F0DLs) were measured as a function of F0 for bandpass filtered sine- and random-phase harmonic complexes in listeners with mild-to-moderate SNHL. All listeners showed a transition between small (good) F0DLs at high F0s and large (poor) F0DLs at low F0s, although the transition point varied across subjects. Two measures that are thought to reflect frequency selectivity generally corresponded to the F0DL transition point: the maximum F0 for which F0DLs were phase-dependent, and the minimum modulation frequency required to discriminate amplitude modulation from quasi-frequency modulation. These results provide some support for the idea that the F0-processing deficit associated with SNHL is related to poor frequency selectivity. However, contrary to our hypothesis, poor correspondence was observed between the F0DL transition point and auditory filter bandwidths estimated using the notched-noise method in the same subjects. This suggests a possible discrepancy between traditional spectral-masking estimates of frequency selectivity and estimates based on temporal interactions between unresolved components. [Work supported by NIH grants R01DC05216 and 5T32DC00038.]

Spectral contributions to intelligibility of sentences with flattened fundamental frequency

Recently, the contribution of fundamental frequency (F0) variation to speech intelligibility in background noise has been studied [J. Laures and G. Weismer, JSLHR 42, 1148–1156 (1999)]. The process used for this analysis is to flatten the frequency contour at the average F0. Results show that sentences with flattened F0 are less intelligible than those with natural F0 variation. However, this technique may reduce the prominence of formant peaks because it excludes F0s that are below the average flattened frequency. As noted by Laures and Weismer (1999), eliminating lower F0s from the analysis results in wider spacing between the harmonics and the available energy to excite the vocal tract resonances is diminished. This study examined the effect of flattening at the average and lowest F0 of each sentence to intelligibility. Twenty subjects listened to sentences in a continuous noise background. Sentences were equated for root-mean square energy. Results showed that the low-flattened condition was significantly more intelligible than the average-flattened condition. However, a condition with natural F0 variation was more intelligible than the two flattened conditions. These findings show that that the frequency of flattening does contribute to intelligibility but that natural F0 variation appears to contribute to a greater degree.

Cortical activities relating to modulation of sound frequency: how to vocalize?

This is the first report to clarify the underlying mechanisms of processing in the modulation of frequencies (tones) in humans using magnetoencephalography (MEG). Volunteers were instructed to vocalize a simple vowel sound (/u/) after receiving a cue (S2) for either one of three (low, middle, or high fundamental frequencies) (F0s). Three tasks, (1) the modulated vocalization task in which the subjects were asked to modulate vocalization tones according to S2, (2) the non-modulated vocalization task in which the subjects were asked to vocalize the same sound (/u/) with a fixed F0, and (3) the image task in which the subjects had to modulate according to S2 and imagine the vowel (/u/) sound, but not vocalize it. In all tasks, two clear components, 1M and 2M, were recorded at approximately 190 and 290 ms after the S2. Since both were identified even in the Image task, they appear to be specifically related to activity for modulation. The equivalent current dipoles of both 1M and 2M were estimated to lie mainly in the inferior frontal lobe or insula in both hemispheres. Therefore, the activity relating to modulation mainly took place in the inferior frontal lobe or insula in both hemispheres starting about 200 ms after the viewing of a cue.

Phase effects in masking in normal-hearing and hearing-impaired listeners

Changes in the phase relationships between components in a harmonic tone complex masker can lead to large changes in masked threshold for long-duration sinusoids in normal-hearing listeners. The effect is reduced or absent in hearing-impaired listeners. This reduction could be due to a loss of peripheral compression, a change in the auditory filters’ phase response, or both. Signal thresholds were measured as a function of masker phase curvature, signal frequency, and masker fundamental frequency (F0). In contrast to previous studies, more than just two phase relationships were considered, allowing us to compare the overall phase response of impaired and normal ears. Peripheral compression was estimated in a separate experiment using a notched-noise masking technique. If changes in filter phase response can account for the differences between normal-hearing and hearing-impaired listeners, then certain phase relationships should produce large threshold differences even in hearing-impaired listeners. If peripheral compression is important, threshold differences in hearing-impaired listeners should emerge at very low F0s, due to the predicted interaction between peripheral compression and temporal processing at higher stages. Preliminary results suggest that the difference between normal-hearing and hearing-impaired listeners can be explained primarily in terms of peripheral compression. [Work supported by DFG and NIDCD.]

Influence of fundamental frequency on stop-consonant voicing perception: A case of learned covariation or auditory enhancement?

Listeners labeling members of an acoustic series modeling VOT (e.g., /ba/-/pa/) are more likely to identify tokens with higher f0 as voiceless than they are for otherwise-identical tokens with lower f0s. This pattern of results may arise because a high f0 enhances perception of voicelessness, in line with auditory enhancement accounts of speech perception. Alternatively, because f0 and VOT covary in English production, it is possible that listeners respond in this manner due to experience with VOT/f0 covariation in the speech signal. The present investigation was designed to tease apart the relative contributions of these two potential mechanisms. Japanese quail (Coturnix coturnix japonica) were trained to ‘‘label’’ stimuli drawn from VOT series by pecking a key. During training, each quail experienced one of three styles of VOT/f0 covariation. For one group of quail, VOT and f0 covaried naturally with voiceless series members having higher f0s than voiced members. Another group of quail heard the inverse, ‘‘unnatural’’ covariation. A final group experienced stimuli for which there was no covariation between VOT and f0. Results indicate that experience with VOT/f0 covariation is the predominant force in shaping perception. Thus, general learning mechanisms may account for this symmetry between perception and production.