You are here
The nine languages covered by the SpeechDat-Car (SDC) project are: Danish, English, Finnish, Flemish/Dutch, French, German, Greek, Italian, Spanish.
Each database collected is composed by 300 speakers. Each speakers is recorded in two environments, each session is made by 125 items. Items collected are:
Isolated Digit, Digit Strings, Natural Numbers, Money Amounts, Dates, Times, Real Names, Company Names, Spelling, Phonetically Rich Words and Sentences, Application Words, Spontaneous Sentences.
The Italian database (as well as the other DBs) should follow some constrains:
- sex: 50% +-5% males and females
- age: the age distribution follows this table
Age 0 - 15 16 - 30 31 - 45 46 - 60 61+
- Proportion 0 >=20% >=20% >=15% >=0%
- regions: Italy has been divided in four macro regions (Nord Est, Nord Ovest, Centro, Sud), each region has to be represented by at least 50 speakers.
There are seven recording conditions to be covered by the database:
- car stopped by motor running
- town traffic
- town traffic with noise (open windows, fan on)
- low speed rough road
- low speed rough road with noisy conditions (fan on, open windows)
- high speed good road
- high speed good road with audio on
The database is composed by five speech signals. Four are recorded in-car with an high quality (16kHz, 16bit), the last signals is recorded through the GSM channel via a mobile phone with a car-kit (8kHz, 8bit).
The effective database is thus made by two files for each utterance, a multiplexed one with all the in-car signals and the telephone one.
The car is equipped with a PC in the trunk, a LCD display and the microphones:
- a close talk microphone Shure SM10A
- three far talk microphones, Peiker ME15/V520 and AKG Q400 Mk3T
- a Peiker microphone, connected to a Nokia 6110 mobile phone via a car-kit
The far microphones are positioned in three different positions: on the left side column, near the sunvisor and near the rear mirror.
The GSM microphone is installed near the rear mirror, close to another microphone. In this way it's also possible to have an idea of the distorsion introduced by the GSM channel.