Methodology

The methodology used

This section describes the methods used to capture data and information in each of the website sections and the process followed to conduct the experiments.

To see the methodology used in each case, select the section you want to see on the menu on the left.

Sounds

Each item is displayed contextualised within a word, with audio sound that can be listened to.The words correspond to informants from different dialects.

The chart for the acoustic analysis performed on each of these words is also shown, consisting of four levels:

  • The upper panel shows the speech waveform, which provides information about time (in msecs.)and breadth (in dBs) of complex waves.
  • The second panel shows the main speech pitch, which provides information on the accent and intonation. When the intonation line is interrupted, then the sound is voiceless (absence of vibration in the vocal cords).
  • The third panel corresponds to the spectrogram (SPG) of the word. A spectrogram shows utterance time on the horizontal axis; frequency (in Hertz) on the vertical axis; amplitude is represented with various tones on the axial axis.
  • The fourth panel includes the IPA transcript of an utterance. The symbols of the International Phonetic Association (https://www.internationalphoneticassociation.org/) are used for this purpose. Each symbol is placed approximately in the centre of the space occupied by each segment in the spectrogram.

There are different types of tables of values, according to the acoustic characteristics of the relevant sounds: 

  1. Duration values for oral occlusive or plosive consonants.
  2. Duration and three acoustic formants for voiced vowels and consonants, that is, nasal, vibrant, lateral and approximant consonants. Formants are responsible for the pitch of each utterance.
  3. Frication duration and onset for non-sibilant fricative consonants.
  4. Frication duration, onset and offset for sibilant fricative consonants.
  5. Duration of plosive and fricative phases, total duration and frication onset and offset for affricate consonants. 

All this information (audio recordings, charts and acoustic values) have been taken from the CD-Rom “Euskararen Hotsak” (Ana Elejabeitia–Alexander Iribar-Rosa Miren Pagola, Bilbao, University of Deusto, 1998). 

2D Images

The 2D image sequences offered have been obtained using the magnetic resonance imaging technique (MRI).The sessions were held at Quirón Bizkaia hospital and GE 1,5T – HealthCare- equipment was used, which obtains Single Shot Fast Spin Eco (SSFSE) dynamic sequences (TR: 4.8, TE: 1.3), 24 frames per minute in the midsagittal plane. The technical protocol for obtaining images was designed and supervised by Dr. Estepan Imanol Gainza Jauregi.

The observable variations in the sequence of the 24 frames that make up the sound for each series are due to the inability to maintain a perfectly steady articulation over a minimum period of time, however short. In addition, occasional movements other than speech articulation (breathing, swallowing, etc.) occur, which should therefore not be considered. 

Although the images correspond to a single informant utterance, images from two other informants are sometimes also provided for clarity purposes.

Animations

Animations are based on actual movements of the vocal tract during the utterance of each sound. They are based on sagittal magnetic resonance images for the articulation of each speech sound in Basque language.

Adobe Flash Professional CS6 tool has been used to develop it, which includes a digitally assisted drawing environment. The base model for all animations is a section of the head where articulatory organs have been highlighted. Subsequently, an animation is generated according to the extreme positions shown by the resonances for each articulation so that the interpolation between rest position and the extreme position generates the simulation of actual movement, based on the inclusion of one or more intermediate positions.

Below is a scheme of articulation places:

3D Images

3D image sequences have been obtained using the magnetic resonance imaging technique (MRI).The recording sessions were conducted at Quirón Bizkaia hospital and GE 1,5T – HealthCare- equipment was used, which obtains Single Shot Fast Spin Eco (SSFSE) dynamic sequences (TR: 4.8, TE: 1.3). The images shown correspond to 56 frames per minute in the sagittal plane. The technical protocol for obtaining images was designed and supervised by Dr. Estepan Imanol Gainza Jauregi.

Although the images correspond to a single informant utterance, images from two other informants are sometimes also provided for clarity purposes.

3D reconstructions

This section describes the three-dimensional reconstructions obtained from 3D magnetic resonance imaging.To do this, the anatomical structures corresponding to the tongue and oral cavities (pharynx, mouth and outside of the oral cavity) have been segmented in each frame. This segmentation was performed using the ITK-SNAP software (http://www.itksnap.org/).

Reconstructions show the tongue and oral cavity. Users can choose to view the combined image or each of its parts separately. They can also stop the rotatory movement at any position and change the perspective offered.

In the case of affricate sounds, two reconstructions are presented corresponding to its two articulatory movements: occlusion and frication.

Electropalatograms

The electropalatography (EPG) tests were conducted at UD’s Phonetics Laboratory. The equipment used was WinEPG Articulate Assistant.

Electropalatography is a technique that monitors the points of contact between the tongue and the palate, using an artificial palate that is fitted into an informant's hard palate.

For each Basque sound, the following is offered:

A video in which the selected sound is heard, synchronised with the chart of the sound spectrum and that of tongue-palate contacts.

  • A static image that shows the contacts during the duration of one of the repetitions of the selected sound in percentage terms.

To interpret electropalatograms, the classification of articulatory zones established by Ana María Fernández Planas (Electropalatographic study of vocalic coarticulation in CVC structures in Spanish, 2001) has been used.

Exercises

The exercises include a number of activities to be solved by users.The answers given will be assessed by the system which, in case of error, will provide additional information that will help users to find the right solution.

The materials used in the preparation of the exercises are provided in the different website sections, in addition to other similar exercises from UD's Phonetics Laboratory. 

We use cookies to improve our website and your experience when using it. Cookies used for the essential operation of this site have already been set. To find out more about the cookies we use and how to delete them, see our privacy policy.

  I accept cookies from this site.
EU Cookie Directive Module Information