GUI of proposed indication tool

Contexts in source publication

Context 1

... et al. provide a vowel map based on the first two formants for long spoken German vowels [11] by means of mean and standard deviation values of the first two formants. A vowel map created based on their data, is visualized in the left subplot of Fig. 3. In order to indicate the vowel, estimated formant frequencies are plotted in the vowel map. From the estimated VT filter coefficientsâcoefficientsˆcoefficientsâ i , formant frequencies and bandwidths are evaluated, which are determined through the location of the corresponding pole in the complex plane. The formant frequency F i and ...

Context 2

... = 90 Hz and a maximum formant frequency F max = 3.5 kHz. Formants which violate one of these criteria are omitted. The detected formants with the lowest two frequencies are considered to be the first two formantsˆFformantsˆ formantsˆF 1 andˆFandˆ andˆF 2 . The estimated formants of several signal blocks are plotted in the vowel map presented in Fig. 3, allowing for a graphical indication of the present vowel based on their ...

Context 3

... For fundamental frequencies f 0 ∈ [70, 320] Hz, prediction scores exceeding 90 % can be achieved. The SVM trained with data for this frequency range is used to classify points of a mesh grid sampling the voice quality feature space in order to visualize the class boundaries, leading to the 2D voice quality map displayed in the right subplot of Fig. 3. ...

Context 4

... the steps depicted in Fig. 1. The features for vowel and voice quality indication are extracted for each signal block, and the results are plotted onto the 2D maps, which were previously created according to subsections 2.4 and 2.5. The results of 15 previous signal blocks are plotted onto the 2D maps as a trace of black dots as visible in Fig. 3. The results shown in Fig. 3 belong to a synthesized, modal /a/ vowel atˆfatˆ atˆf 0 ≈ 300 Hz. The buffer structure allowing block processing was taken from [13]. An additional field in the GUI's lower right corner is indicating the currentˆfcurrentˆ currentˆf 0 , calculated during the preprocessing steps of the analysis stage. The ...

Context 5

... Fig. 1. The features for vowel and voice quality indication are extracted for each signal block, and the results are plotted onto the 2D maps, which were previously created according to subsections 2.4 and 2.5. The results of 15 previous signal blocks are plotted onto the 2D maps as a trace of black dots as visible in Fig. 3. The results shown in Fig. 3 belong to a synthesized, modal /a/ vowel atˆfatˆ atˆf 0 ≈ 300 Hz. The buffer structure allowing block processing was taken from [13]. An additional field in the GUI's lower right corner is indicating the currentˆfcurrentˆ currentˆf 0 , calculated during the preprocessing steps of the analysis stage. The only variable parameter is the ...