logo
Speech Recordings  |  Speech Characteristics  |  Normative Studies  |  Depression  |  Coping Skills  |  Program Package  |  Home
unisig
logo

FDFREQ — Analysis of mean vocal pitch, overtone distributions

Determines the voice sound characteristics, separately for male and female speakers, in terms of mean vocal pitch (fundamental frequency F0, F0-amplitude) and intonation (F0-variability, F0-contour, and overtone distribution). The respective speech parameters are estimated from a sequence of consecutive spectra computed at a quartertone resolution over 7 octaves between 64 Hz and 8192 Hz (cf. program TRANS). The distributions of the above frequency-related speech parameters within a given population can be computed and plotted.

 
            Specificationlist:      FDFREQ
            ------------------------------
            R4 CUT                  16.000  Default-value
            R4 DMIN                  4.000  Default-value
            I4 FRST                      1  Default-value
            I4 NSPK                      0  Default-value
            I4 PROT                      0  Default-value
            I4 PLOT                      0  Default-value
            I4 PMAX                      0  Default-value
            I4 PROZ                      0  Default-value
            I4 LPRT                      6  Default-value
            I4 SAVE                      0  Default-value
 
            01 CUT  Specifies minimal spectral amplitude for formants
            02 DMIN Specifies signal attenuation for bandwidth calculations
            03 FRST Specifies first spectrum to be used (out of maximal 44)
            04 NSPK Specifies number of spectra to be used for F0 estimation
            05 PROT Controls output to display/printer
            06 PLOT Controls graphic output
            07 PMAX Specifies maximum number of plot pages
            08 PROZ Specifies global maximum for frequency distributions in [%]
            09 LPRT Logical unit number of plot-device
            10 SAVE Controls storage of data in databank
            11 DEMO Examples that illustrate program function
 
            - CUT  = c: Minimum spectral amplitude required for the
                        estimation of formants (default=16)
 
            - DMIN = d: In cases where the F0-6db-bandwidth cannot be
                        determined the program looks for a "DMIN"db-bandwidth
                        (default=4db)
 
            - FRST = e: First spectrum to be used (out of a maximum of 44)
 
            - NSPK = n: Number of spectra to be used for F0 estimation
 
 
            - PROT = 0: No print output
                   = 1: Analysis of mean vocal pitch F0 (in Hz)
                   = 2: Analysis of mean vocal pitch F0 (quartertones)
                   = 3: Analysis of F0-amplitudes
                   = 4: Analysis of F0-variability (intonation)
                   = 5: Analysis of F0-contour
 
            - PLOT = 0: No plot output
                   = 1: Spectra with markings for formants F0,F1,F2
                   = 2: F0 distribution (single cases)
                   = 3: F0-Amplitude distribution (single cases)
                   = 4: F0 distributions of entire sample (males vs. females)
 
            - PMAX = q: Maximum number of plot pages
                   = 0: Unlimited number
 
            - PROZ = p: Global maximum for frequency distributions in [%]
                   = 0: Will be optimized for each individual plot
 
            - LPRT = q: Logical unit number of plot-device (standard=6;
                        valid numbers are 46-96)
 
            - SAVE = 0: No effect
                   = 1: Data are to be stored in databank
 
 
            - DEMO: Spectral analysis of speech recordings
 
           

Example

 
 
            &&START OUTPUT
            &&START CSELECT=Normative speech study zurich  (study 600/1+2)
             STUD=600,ATAG=0,PROT=1,POPU=STUD600,RSET=1
            &&START FDFREQ=Analysis of converted bli-tapes (study 600/1+2)
             FRST=2,NSPK=32,PROT=1,PLOT=0,SAVE=1
            &&START CSELECT=Normative speech study zurich  (study 601)
             STUD=601,ATAG=0,PROT=1,POPU=STUD601,RSET=1
            &&START FDFREQ=Analysis of converted bli-tapes (study 601)
             FRST=2,NSPK=32,PROT=1,PLOT=0,SAVE=1
            &&START CSELECT=Normative speech study zurich  (study 602/1+2+3+7)
             STUD=602,ATAG=0,PROT=1,POPU=STUD602,RSET=1
            &&START FDFREQ=Analysis of converted bli-tapes (study 602/1+2+3+7)
             FRST=2,NSPK=32,PROT=1,PLOT=0,SAVE=1
 
 
           
voxFig23
Fig. 23: Voice sound characteristics ("timbre") of a female speaker as quantified through spectral analyses. Spectral intensities are plotted along the y-axis on log-proportional scales and as a function of frequency (x-axis: 7 octaves covering the frequency range of 64-8192Hz). The maxima represent the "overtones" at fixed, physically well-defined intervals from the fundamental frequency F0 ("octave": frequency ratio 2:1; "fifth": frequency ratio 3:2; "forth": frequency ratio 4:3; "major third": frequency ratio 5:4; "minor third": frequency ratio 6:5; etc).

Partners:
Everis, Spain
ETH, Switzerland
UZH, Switzerland
Freiburg, Germany
MA Systems, UK
Bristol, UK
Xiwrite, Italy
Ultrasis, UK
Jaume, Spain
Valencia, Spain
Lanzhou, China

 

EU-Grant (FP7):
248544

[ Comments and suggestions to Webmaster ] k454910@bli.uzh.ch
[ Home ]
Impressum  |  Acknowledgements