Speech resynthesis
Contents
All the stimuli below were created using the excellent speech synthesizer MBROLA (T. Dutoit), the pitch extractor in BLISS (J. Mertus, Brown University), as well as home-made programs.
Selective degradation/preservation of phonetic/prosodic properties
The sound files below are taken from the ones I used in my experiments described in the following paper:Ramus, F., & Mehler, J. (1999). Language identification with suprasegmental cues: A study based on speech resynthesis. Journal of the Acoustical Society of America, 105(1), 512-521.
These are .au files, 16-bit, 16000 Hz. Sorry if your browser doesn't support these.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Explanations:
- saltanaj: fricatives were turned into /s/, vowels into /a/, liquids into /l/, occlusives into /t/, nasals into /n/, and semi-vowels into /j/.
- sasasa: consonants were turned into /s/, vowels into /a/.
- aaaaaa: all phonemes were turned into /a/.
- flat sasasa: same as sasasa, but with a flat intonation.
The original intonation is preserved in all the above versions.
Normal and backward Dutch and Japanese sentences
The sound files below are taken from the ones I used in my experiments described in the following papers:Ramus, F., Hauser, M. D., Miller, C., Morris, D., & Mehler, J. (2000). Language discrimination by human newborns and by cotton-top tamarin monkeys. Science, 288, 349-351.
Ramus, F. (2000). L'étude comparative de la perception de la parole : développements récents. Primatologie, 3, 421-444.
They are also available from
Science
on-line.
| Dutch | Japanese | |
| Original (Exp. 1A) | click here! | click here! |
| Original backwards (Exp. 2A) | click here! | click here! |
| saltanaj (Exp. 1B, 2B) | click here! | click here! |
| saltanaj backwards (Exp. 1C, 2B) | click here! | click here! |
Explanations:
- backwards: amounts to digitally reversing the order of the samples. It is equivalent to playing a tape in the wrong direction.
- saltanaj is as explained above.
- saltanaj backwards means the sentence is first resynthesized in the saltanaj manner, then played backwards.
Pure rhythm for babies
The sound files below are taken from the ones I used in my experiments described in the following paper:Ramus, F. (2002). Language discrimination by newborns: Teasing apart phonotactic, rhythmic, and intonational cues. Annual Review of Language Acquisition, 2.
| Dutch | Japanese | |
| Original (Exp. 1) | click here! | click here! |
| saltanaj (Exp. 2) | click here! | click here! |
| saltanaj with
artificial intonation (the same for both languages) (Exp. 4) |
click here! | click here! |
| sasasa with
artificial intonation (the same for both languages) (Exp. 3) |
click here! | click here! |
Explanations:
When testing babies on pure rhythm, I feared that a flat intonation (as in the flat sasasa) might be too boring for them. This is why I used an artificial intonation contour instead.
-
saltanaj is
as explained above.
sasasa is as explained above
in the last two versions, the same artificial intonation contour is applied to the sentences of both languages, regardless of their original intonation.
Comparing Catalan and Polish with English and Spanish
The sound files below are taken from the ones I used in the experiments described in the following paper:Ramus, F., Dupoux, E., & Mehler, J. (2003). The psychological reality of rhythm classes: Perceptual studies. Paper presented at the 15th International Congress of Phonetic Sciences, Barcelona.
| English | Spanish | Polish | Catalan | |
| Original | click here! | click here! | click here! | click here! |
| flat sasasa | click here! | click here! | click here! | click here! |
