Speech Technology Group: Klatt-based samples
GTH

Speech Technology Group: Klatt-based samples


We have selected some samples depending on your interests (and possibly your conection speed). You can also get some samples of our concatenativer approach. Remember all the systems are for Spanish only.

Original converted speech was generated at 10 KHz using 16 bits linear signed samples. If you select formats with lower resolution or sample frequency, sound quality may be degraded. Conversions were done using the Lance Norskog's SOX package. Please, read the warning notes regarding each format before selecting the samples.

Unfortunately, we are not able to check every file format (just Sun AU (8 KHz), raw data and Wavs up today), so please, let us know if you are unable to get them sound properly so that we can fix it properly.


Notes regarding sound file formats given

We are offering the following file formats for your convenience:

Raw data (big-endian) (.rawb)
Raw data directly generated by our TtS software. Sampling rate is 10 KHz, and samples are 16 bits linear signed written for big-endian machines (Sun 4 for example).
Raw data little-endian (.rawl)
Raw data directly generated by our TtS software. Sampling rate is 10 KHz, and samples are 16 bits linear signed written for little-endian machines (PC, VAX for example).
MS-WINDOWS WAV format (.wav)
Standard WAV files recorded at 10 KHz.
Traditional SUN AU format (.au)
AU files using u-law samples at 8 KHz.
Mac snd (.snd)
Mac snd files sampled at 11 KHz.
Mac aiff (.aiff)
Mac hcom (.hcom) at 11025 Hz
Soundblaster VOC format (.voc)
Turtle Beach files (smp) (.smp)


Big paragraph

A big paragraph showing variations in pitch, speed and volume:

Vamos a hacerles una demostración, acerca del funcionamiento, y las posibilidades de nuestro conversor texto habla . En primer lugar, podemos modificar la velocidad. Por ejemplo. [v:80] Vísteme despacio que tengo prisa. [v:250] Así por el contrario puedo ir más rápido. [v:180] A continuación oirán una frase con variaciones de entonación. [t:170] A Dios rogando. [t:60] Y con el mazo dando. [t:100] Se pueden reproducir ficheros de formato DOS y güindous. Por último, entiendo algunas abreviaturas, como excmo., Kg., kHz., km/h., tel., etc..

[rawb 1.1Mb] [rawl 1.1Mb] [wav 1.1Mb] [au 590Kb] [snd 406Kb] [aiff 1.1M] [hcom 150Kb] [voc 590Kb] [smp 1.1Mb]


Reduced paragraph

A somehow smaller paragraph.

Este es el conversor texto voz, del Departamento de Ingeniería Electrónica, adscrito a la Escuela Técnica Superior de Ingenieros de Telecomunicación, de la Universidad Politécnica de Madrid.

[rawb 374Kb] [rawl 374Kb] [wav 374Kb] [au 187Kb] [snd 128Kb] [aiff 374Kb] [hcom 51Kb] [voc 187Kb] [smp 374Kb]


Single sentence in enunciative form

Vas a ir a la Universidad.

[rawb 45Kb] [rawl 45Kb] [wav 45Kb] [au 22Kb] [snd 15Kb] [aiff 45Kb] [hcom 7.5Kb] [voc 22Kb] [smp 45Kb]


Single sentence in interrogative form

żVas a ir a la Universidad?

[rawb 63Kb] [rawl 63Kb] [wav 63Kb] [au 31Kb] [snd 22Kb] [aiff 63Kb] [hcom 8Kb] [voc 31Kb] [smp 64Kb]


Single sentence with nasal sounds

Mi mamá le cantó una nana al niño.

[rawb 66Kb] [rawl 66Kb] [wav 66Kb] [au 33Kb] [snd 22Kb] [aiff 66Kb] [hcom 10Kb] [voc 33Kb] [smp 66Kb]


Single sentence with fricatives

Las fosas no dejan pasar con zuecos.

[rawb 78Kb] [rawl 78Kb] [wav 78Kb] [au 39Kb] [snd 27Kb] [aiff 78Kb] [hcom 11Kb] [voc 39b [smp 78Kb]


Single sentence with BDG sounds

La bodega de Diego.

[rawb 34Kb] [rawl 34Kb] [wav 34Kb] [au 17Kb] [snd 11Kb] [aiff 34Kb] [hcom 5Kb] [voc 17Kb] [smp 34Kb]


PREVIOUS PAGE HOME PAGE TOP

postscript