This is supplemental material for The role of creaky voice in Cantonese tone perception. It includes sample wavefiles from perceptual stimuli for Experiment 2, for creaky and non-creaky Tone 4 exemplars. We also use a few of these wavefiles to demonstrate criteria for coding a token file as creaky or not creaky (Experiment 1, 2). Finally, we include sample wavefiles from the perceptual experiment, Experiment 3.
Supplemental data is presented in the following sections:
- Experiments 1/2: Criteria for creak coding
- Experiment 2: Sample perceptual stimuli
- Experiment 3: Sample perceptual stimuli
The criteria for coding uttered syllables as "creaky" for Experiments 1 and 2 were given in the paper as follows:
Target syllables in individual tokens were determined to be creaky by listening and visual inspection of the waveform and spectrogram in Praat (Boersma 2010). A token was defined to be creaky if it had the auditory percept of creaky voice, as determined by the authors and if:
- there were alternating cycles of amplitude and/or frequency or irregular glottal pulses in the waveform or wide-band spectrogram,
- missing values or discontinuities in the f0 track determined by Praat's autocorrelation algorithm with default settings1, or
- the appearance of strong subharmonics or lack of harmonic structure in the narrow-band spectrogram. Generally these three indicators occurred simultaneously.
Below are some samples of implementing the criteria:
Using criteria 1 and 2
Using criteria 2 and 3
Note: all sample audio files provided are in WAV format.
These audio files were drawn from in creating perceptual stimuli for Experiment 2. The subset provided below was used in creating the Tone 4 stimuli.
- Play Tone 4 stimuli coded as "creaky": [wav]
- Play Tone 4 stimuli coded as "non-creaky": [wav]
- Zipped archives of wav files: [creaky], [non-creaky]
Note: all sample audio files provided are in WAV format and are the resynthesized stimuli for "0" f0 shift, in the monosyllabic experiment (no preceding syllable).
Sources for Figures 1 and 2 in paper:
Figure 1: waveforms
- (top left) sex: female, creak type: narrow, creak proportion: heavy
- (top right) sex: male, creak type: narrow, creak proportion: heavy
- (bottom left) sex: female, creak type: wide, creak proportion: heavy
- (bottom right) sex: male, creak type: wide, creak proportion: heavy
Figure 2: spectrograms
- (left) sex: female, creak type: narrow, creak proportion: heavy
- (right) sex: female, creak type: pitched, creak proportion: heavy
silence threshold = 0.03, voicing threshold = 0.45, octave cost = 0.01, octave-jump cost = 0.35, voiced/unvoiced cost = 0.14 ↩