On certain irregularities of voiced-speech waveforms

Dolansky, Ladislav; Tjernlund, P.

doi:10.1109/tau.1968.1161962

Cited by 14 publications

(7 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is attributed to the fact that the number of basic periods extracted from the speech of subject ao is smaller than those for other speakers (as indicated in Table 1), so that differences in the correlation structures of the pitch fluctuations are not easily perceived. In fact, a similar experiment was carried out by repeating the basic pitch series {T 1 Fig. 7.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Higher‐order correlation analysis of pitch fluctuations in sustained normal vowels by the method of surrogate data

Tokuda

Miyano

Aihara

2004

Electron Comm Jpn Pt III

View full text Add to dashboard Cite

SUMMARYThe pitch fluctuations observed in the sustained pronunciation of voiced sounds are known to be indispensable for perception of voiced sound as natural human pronunciation. In speech synthesis, it is necessary to carefully identify the statistical properties of the pitch fluctuations that are required in order for the reproduced speech to maintain the quality of the original speech. From research to date, it is known that maintaining the second-order correlation structure is important for the realization of natural pitch fluctuations. No information is available on the importance of higher-order correlations above third order. In the present research, a higher-order correlation analysis of pitch fluctuations is carried out by the surrogate method. The effect of the correlation structures of third and higher orders on the naturalness of vowels is studied. The results show that reproduction of the correlation configuration of third and higher orders is important in addition to the second-order correlation in the synthesis of pitch fluctuations. Possible approaches based on nonlinear models are discussed.

show abstract

Section: Methodsmentioning

confidence: 99%

“…Although at first glance this pitch appears to be periodic, it is known that the basic period and the waveform of the pitch fluctuate subtly over time [1]. Various experiments have shown that these pitch fluctuations are indispensable for speech to be perceived as natural human voice.…”

Section: Introductionmentioning

confidence: 99%

Higher‐order correlation analysis of pitch fluctuations in sustained normal vowels by the method of surrogate data

Tokuda

Miyano

Aihara

2004

Electron Comm Jpn Pt III

View full text Add to dashboard Cite

show abstract

“…Instead, [11] used a cost proportional to the derivative of log pitch and also applied a reduced cost to octave jumps. A complication is that, particularly at the end of voiced segments, the true pitch of speech may become irregular, make abrupt octave jumps or show bicyclic behavior in which odd and even larynx cycles have different periods [29]. Although DP can compensate for pitch estimation errors at the frame level, the use of a strong continuity constraint may itself introduce errors and is no substitute for high accuracy in the raw pitch estimation.…”

Section: Temporal Continuity Constraintsmentioning

confidence: 99%

PEFAC - A Pitch Estimation Algorithm Robust to High Levels of Noise

Gonzalez

Brookes

2014

IEEE/ACM Trans. Audio Speech Lang. Process.

210

125

View full text Add to dashboard Cite

We present PEFAC, a fundamental frequency estimation algorithm for speech that is able to identify voiced frames and estimate pitch reliably even at negative signal-to-noise ratios. The algorithm combines a normalization stage, to remove channel dependency and to attenuate strong noise components, with a harmonic summing filter applied in the log-frequency power spectral domain, the impulse response of which is chosen to sum the energy of the fundamental frequency harmonics while attenuating smoothly-varying noise components. Temporal continuity constraints are applied to the selected pitch candidates and a voiced speech probability is computed from the likelihood ratio of two classifiers, one for voiced speech and one for unvoiced speech/silence. We compare the performance of our algorithm with that of other widely used algorithms and demonstrate that it performs well in both high and low levels of additive noise.Index Terms-Fundamental frequency, noisy speech, pitch, speech processing. 2329-9290

show abstract

“…Interestingly, the whole plethora of nonlinear phenomena have been observed in a wide variety of vocalisations: newborn cries [101], pathological voices [99], normal speech [102], Russian lament [103], animal communication [104,105], and in contemporary vocal music [106].…”

Section: Subharmonics Biphonation and Deterministic Chaosmentioning

confidence: 99%

Analysing and Understanding the Singing Voice: Recent Progress and Open Questions

Kob¹,

Henrich²,

Herzel³

et al. 2011

CBIO

View full text Add to dashboard Cite

Abstract:The breadth of expression in singing depends on fine control of physiology and acoustics. In this review, the basic concepts from speech acoustics, including the source-filter model, models of the glottal source and source-filter interactions, are described. The precise control, the extended pitch range, the timbre control and, in some cases, the uses of alternate phonation modes all merit further attention and explanation. Here we review features of the singing voice and the understanding that has been delivered by new measurement techniques. We describe the glottal mechanisms and the control of vocal tract resonances used in singing. We review linear and nonlinear components of the voice and the way in which they are measured and modelled and discuss the aero-acoustic models. We conclude with a list of open questions and active fields of research.

show abstract

On certain irregularities of voiced-speech waveforms

Cited by 14 publications

References 1 publication

Higher‐order correlation analysis of pitch fluctuations in sustained normal vowels by the method of surrogate data

Higher‐order correlation analysis of pitch fluctuations in sustained normal vowels by the method of surrogate data

PEFAC - A Pitch Estimation Algorithm Robust to High Levels of Noise

Analysing and Understanding the Singing Voice: Recent Progress and Open Questions

Contact Info

Product

Resources

About