SpeechDat-Car Fixed Platform

SpeechDat-Car aims to develop a set of speech databases to support training and testing of multilingual speech recognition applications in the car environment. Two types of recordings compose the database. The first type consist of wideband audio signals recorded directly in the car while the second type is composed by GSM signals transmitted from the car and recorded simultaneously in a far-end. Therefore, two recording platforms were used, a 'mobile' recording platform installed inside the car and a 'fixed' recording platform located at the far-end fixed side of the GSM communications system. This paper describes the fixed platform software developed by the Universitat Politècnica de Catalunya (ADA-K). This software is able to work with standard inexpensive PC cards for ISDN lines.. The telephone server presented in this paper to automate the recording of speech databases was developed by the authors in the framework of the SpeechDat-Car EC project LE4-8334 [Moreno (2000)]. Automatic speech recognition (ASR) appears to be a particularly well-adapted technology for providing voice-based interfaces (based on hands-free mode) that will enable new in-car applications to develop while taking care of safety aspects. However, the car environment is known to be particularly noisy (street noise, car engine noise, vibration noises, bubble noise, etc...). To obtain an optimal performance for speech recognition, it is necessary to train the system on large corpora of speech data recorded in context (i.e. directly in the car). The European project SpeechDat-Car 1 aims at providing a set of uniform, coherent databases for nine European languages and for American English. SpeechDat-Car continues the success of the SpeechDat project in developing large-scale speech resources for a wide range of languages and for in-car applications (voice dialling, car accessories control, etc.). It will produce resources for participation of external partners to the original consortium is also possible. Siemens is an 'external' partner. It is also important to note that SpeechDat-Car commits itself to a strict validation protocol to ensure optimal quality and exchangeability of the databases. 1 SpeechDat-Car started in April 1998 in the 4th EC framework under project code LE4-8334 with a 30 months' project duration. 5HFRUGLQJJSODWIRUPV Two types of recordings compose the database. The first type consist of wideband audio signals recorded directly in the car and the second type is composed by a GSM signal transmitted from the car and recorded simultaneously in a far-end. Two recording platforms were used, a 'mobile' recording platform …