This paper deals with the analysis of cry melodies (time variations of the fundamental frequency) as well as vocal tract resonance frequencies (formants) from infant cry signals. The increase of complexity of cry melodies is a good indicator for neuro-muscular maturation as well as for the evaluation of pre-speech development. The variation of formant frequencies allows an estimation of articulatory activity during pre-speech vocalization. Subjects are three pairs of healthy identical twins (monocygozity determined by DNA-fingerprint). Spontaneous cries of these six children were recorded at different ages: 8th-9th week, 15th-17th week and 23rd-24th week. Analysis of 136 cry melodies and intensity contours was made using KAY-CSL 4300/MDVP. For formant estimation a spectral parametric technique was applied, which was based on autoregressive models (Digital spectral analysis with applications, 1987) whose order is adaptively estimated on subsequent signal frames by means of a new method (Med. Eng. Phys. 20 (1998) 432; Utras. Med. Biol. 21 (1995) 793). Cry melodies exhibited an increasing complexity during the observation period. Beginning with the second observation period (15th-17th week) an increasing coupling and tuning between melody and resonance frequencies was observed, which was interpreted as "intentional" articulatory activity. Possible applications are in cry diagnosis as well as in the evaluation of pre-speech development.

One of the basic parameters characterising voiced phonation is the fundamental frequency, named pitch, which is the rate of vibration of the folds. In pathological voices. pitch variations within an utterance are indicative of the patient status. As such voices are corrupted by 'noise', robust pitch estimation methods are required in order to track its variations. This paper aims to compare some pitch estimation methods, pointing out their main advantages and drawbacks for present application. For each method, modifications are proposed in order to enhance performance. The methods are tested on simulated signals and then applied to real signals, coming both from healthy and pathological voices. The latter we obtained from patients who have undergone surgery for vocal folds via laser or traditional lancet techniques.

This paper aims at estimating the fundamental frequency (pitch) and the vocal tract resonant frequencies (formants) from newborn infant cry signals. Such parameters are of interest in exploring brain function at early stages of child development, for the timely diagnosis of neonatal disease and malformation. The paper compares a spectral parametric technique and the cepstrum approach, extending previous results. The parametric technique is based on autoregressive models whose order is adaptively estimated on subsequent signal frames by means of a new method. This allows the correct tracking of pitch and formant variations with time. The traditional cepstrum approach is modified in order to follow signal variability. In particular, the cepstrum spectral resolution is improved by applying the chirp Z-transform (CZT) and by adaptively varying the 'lifter' length. The two methods are tested on simulated data, as far as robustness to noise and spectral resolution are concerned, and are then applied to real baby cry data.

For pathological voices, spectral noise is closely related to the degree of perceived hoarseness. In this paper, noise variations are tracked during an utterance by means of an adaptive version of the normalized noise energy method [1]. A first step is devoted to pitch estimation, which allows defining the varying optimal time window length for noise retrieval, avoiding empty "dip" regions. The approach is tested on synthetic data and applied to real data coming from cordectomised adult male patients.