Resource: Portuguese SpeechDat(M) database
|Reference||Portuguese SpeechDat(M) database|
|Date of Submission||Jan. 24, 2014, 4:31 p.m.|
|Resource Type||Primary Text|
The Portuguese SpeechDat(M) database contains the recordings of 1,001 speakers (453 males, 548 females). This speech database was collected by Portugal Telecom within the European SpeechDat project.
Speech signals are stored as sequences of 8 kHz, 8-bit A-law. Files are stored according to the file specifications proposed in the SpeechDat database format specification. The file formats and headers follow the SAM recommendations (header files separated from signal files).
This speech database was validated by SPEX (the Netherlands) to assess its compliance with the SpeechDat format and content specifications.
Each speaker uttered the following items:
* 3 natural numbers
The approach adopted for speaker recruitment involved selecting speakers among the employees of Portugal Telecom (about 20,000) and their relatives. The company has a wide geographical coverage, thus guaranteeing a good representation of many regional accents.
The following age distribution has been obtained: 12 speakers are under 16, 345 speakers are between 17 and 30, 436 speakers are between 31 and 45, 196 speakers are between 46 and 60 and 8 speakers are over 60; the age of two speakers is unknown and two others said they were born in 1996.
A pronunciation lexicon with a phonemic transcription in SAMPA is also included.