A sawtooth waveform inspired pitch estimator (SWIPE) has been developed for speech and music. SWIPE estimates the pitch as the fundamental frequency of the sawtooth waveform whose spectrum best matches the spectrum of the input signal. The comparison of the spectra is done by computing a normalized inner product between the spectrum of the signal and a modified cosine. The size of the analysis window is chosen appropriately to make the width of the main lobes of the spectrum match the width of the positive lobes of the cosine. SWIPE('), a variation of SWIPE, utilizes only the first and prime harmonics of the signal, which significantly reduces subharmonic errors commonly found in other pitch estimation algorithms. The authors' tests indicate that SWIPE and SWIPE(') performed better on two spoken speech and one disordered voice database and one musical instrument database consisting of single notes performed at a variety of pitches.

No abstract

Ramos-Campo, DJ, Rubio-Arias, JÁ, Freitas, TT, Camacho, A, Jiménez-Diaz, JF, and Alcaraz, PE. Acute physiological and performance responses to high-intensity resistance circuit training in hypoxic and normoxic conditions. J Strength Cond Res 31(4): 1040-1047, 2017-The aim of this study was to analyze physical performance and physiological variables during high-intensity resistance circuit training (HRC) with the addition of 2 levels (moderate and high) of systemic hypoxia. Twelve resistance-trained young male subjects participated in the study. After a 6 repetition maximum testing session, participants performed 3 randomized trials of HRC: normoxia (NORM: fraction of inspired oxygen [FiO2] = 0.21; ∼0 m altitude), moderate hypoxia (MH: FiO2 = 0.16; ∼2.100 m altitude), or high hypoxia (HH: FiO2 = 0.13; ∼3.800 m altitude), as controlled by a hypoxic generator. Bench press force, heart rate and heart rate variability, rating of perceived exertion, resting metabolic rate, energy cost, and countermovement jump were assessed in each session. Heart rate variability in HH was significantly lower (standard deviation of all normal NN intervals [intervals between two "normal" beats] = 111.9 vs. 86.7 milliseconds; standard deviation of the difference between consecutive NN intervals = 19.5 vs. 17.0 milliseconds; p ≤ 0.05) in comparison with NORM. There were significant differences in rating of perceived exertion between NORM and HH (11.6 vs. 13.8 points). Peak and mean force on the bench press were significantly lower (p ≤ 0.05) in HH when compared with MH (peak: 725 vs. 488 N; mean: 574 vs. 373 N). Energy cost was significantly higher (p ≤ 0.01) in both hypoxic conditions compared with NORM (NORM: 10.4; MH: 11.7; HH: 13.3 kJ·min). There were no differences between conditions in heart rate and countermovement jump variables. These results indicate that hypoxic stimuli during HRC exercise alter physical performance and physiological variables and affect how strenuous the exercise is perceived to be. High-intensity resistance circuit training in hypoxia increases the stress on the performance and physiological responses, and these differences must be taken into account to avoid an excessive overload.

Perception of breathy voice quality is cued by a number of acoustic changes including an increase in aspiration noise level (AH) and spectral slope [1]. Changes in AH in a vowel may be evaluated through measures such as the harmonic-to-noise ratio (HNR), cepstral peak prominence (CPP) or via auditory measures such as the partial loudness of harmonic energy (PL) and loudness of aspiration noise (NL). Although a number of experiments have reported high correlation between such measures and ratings of perceived breathiness, a formal model to predict breathiness of a vowel has not been proposed. This research describes two computational models to predict changes in breathiness resulting from variations in AH. One model uses auditory measures while the other uses CPP as independent variables to predict breathiness. For both cases, a translated and truncated power function is required to predict breathiness. Some parameters in both of these models were observed to be pitch-dependent. The "unified" model based on auditory measures was observed to be more accurate than one based on CPP.