Person:
Carlosena García, Alfonso

Profile Picture

Email Address

Birth Date

Research Projects

Organizational Units

Job Title

Last Name

Carlosena García

First Name

Alfonso

person.page.departamento

ORCID

0000-0002-7146-4043

person.page.upna

370

Name

Filters

2002 - 20066
Reset filters

Settings

Author: Arroabarren Alemán, Ixone ×

Search Results

Now showing 1 - 6 of 6
  • Thumbnail Image
    PublicationOpen Access
    Vibrato in singing voice: the link between source-filter and sinusoidal models
    (Springer Open, 2004) Carlosena García, Alfonso; Arroabarren Alemán, Ixone; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    The application of inverse filtering techniques for high-quality singing voice analysis/synthesis is discussed. In the context of source-filter models, inverse filtering provides a noninvasive method to extract the voice source, and thus to study voice quality. Although this approach is widely used in speech synthesis, this is not the case in singing voice. Several studies have proved that inverse filtering techniques fail in the case of singing voice, the reasons being unclear. In order to shed light on this problem, we will consider here an additional feature of singing voice, not present in speech: the vibrato. Vibrato has been traditionally studied by sinusoidal modeling. As an alternative, we will introduce here a novel noninteractive source filter model that incorporates the mechanisms of vibrato generation. This model will also allow the comparison of the results produced by inverse filtering techniques and by sinusoidal modeling, as they apply to singing voice and not to speech. In this way, the limitations of these conventional techniques, described in previous literature, will be explained. Both synthetic signals and singer recordings are used to validate and compare the techniques presented in the paper.
  • Thumbnail Image
    PublicationOpen Access
    Glottal spectrum based inverse filtering
    (ISCA, 2003) Arroabarren Alemán, Ixone; Carlosena García, Alfonso; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    In this paper a new inverse filtering technique for the time-domain estimation of the glottal excitation is presented. This approach uses the DAP modeling for the vocal tract characterization, and a spectral model for the derivative of the glottal flow. This spectral model is based on the spectrum of the KLGLOTT88 model for the glottal source. The proposed procedure removes the glottal source from the spectrum of the speech signal in an accurate manner, particularly for highpitched signals and singing voice, and the estimated glottal waveforms present less amount of formant ripple.
  • Thumbnail Image
    PublicationOpen Access
    Unified analysis of glottal source spectrum
    (ISCA, 2003) Arroabarren Alemán, Ixone; Zivanovic, Miroslav; Carlosena García, Alfonso; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    The spectral study of the glottal excitation has traditionally been based on a single time-domain mathematical model of the signal, and the spectral dependence on its time domain parameters. Opposite to this approach, in this work the two most widely used time domain models have been studied jointly, namely the KLGLOTT88 and the LF models. Their spectra are analyzed in terms of their dependence on the general glottal source parameters: Open quotient, asymmetry coefficient and spectral tilt. As a result, it has been proved that even though the mathematical expressions for both models are quite different, they can be made to converge. The main difference found is that in the KLGLOTT88 model the asymmetry coefficient is not independent of the open quotient and the spectral tilt. Once this relationship has been identified and translated to LF model, both models are shown to be equivalent in both time and frequency domains.
  • Thumbnail Image
    PublicationOpen Access
    Modelling of vibrato production
    (2004) Arroabarren Alemán, Ixone; Carlosena García, Alfonso; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    In this paper, a novel Non-Interactive Source-Filter model for singing voice, incorporating vibrato, is proposed. In this way the periodic frequency variation, peculiar to vibrato is included in a simplified signal model for singing voice production. The inclusion of vibrato will further allow to relate two distinct signal models for the voice: The Non Interactive Source-Filter and the Sinusoidal Models. In this way, the Instantaneous Amplitude and Instantaneous Frequency of the harmonics, which are intrinsically Sinusoidal Model parameters, will be related to the glottal source and vocal tract response, typical of the Source-Filter model. Thanks to the proposed model, the most relevant acoustic features of vibrato will be related to the voice production mechanisms.
  • Thumbnail Image
    PublicationOpen Access
    Analysis and synthesis of vibrato in lyric singers
    (2002) Arroabarren Alemán, Ixone; Carlosena García, Alfonso; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    In this paper two aspects of the vibrato signal characterisation are presented. First, an analysis method, which breaks down the musical signal into its different partials, and decomposes them into two separate contributions: AM and FM. This allows an objective characterization of the vibrato according to musical parameters. Second, a procedure for vibrato synthesis is proposed, based on the previous analysis. Even though the results of the synthesis are not fully satisfactory, they shed new light on the vibrato modeling.
  • Thumbnail Image
    PublicationOpen Access
    Effect of the glottal source and the vocal tract on the partials amplitude of vibrato in male voices
    (Acoustical Society of America, 2006) Arroabarren Alemán, Ixone; Carlosena García, Alfonso; Ingeniería Eléctrica y Electrónica; Ingeniaritza Elektrikoa eta Elektronikoa
    In this paper the production of vocal vibrato is investigated. The most relevant features of the acoustical vibrato signal, frequency and amplitude variations of the partials, will be related to the voice production features, glottal source GS and vocal tract response VTR . Unlike previous related works, in this approach, the effect on the amplitude variations of the partials of each one of the above-mentioned voice production features will be identified in recordings of natural singing voice. Moreover, we will take special care of the reliability of the measurements, and, to this aim, a noninteractive vibrato production model will be also proposed in order to describe the vibrato production process and, more importantly, validate the measurements carried out in natural vibrato. Based on this study, it will be shown that during a few vibrato cycles, the glottal pulse characteristics, as well as the VTR, do not significantly change, and only the fundamental frequency of the GS varies. As a result, the pitch variations can be attributed to the GS, and these variations, along with the vocal tract filtering effect, will result in frequency and amplitude variations of the acoustic signal partials.