Philips Electronics NV’s Philips Dictation Systems is claiming a breakthrough in speech-to-text technology, saying that it has the world’s first digital dictation system using truly continuous speech recognition. In a claim that is likely to ruffle feathers at IBM Corp, which makes very similar protestations about its own technology, Philips says the system recognises natural speech spoken at normal speed and transforms it into text; it was developed with Philips Research Laboratories in Aachen, Germany. Called the Speech Processing System 6000 or SP 6000, it will come with vertical market vocabularies, and the first version incorporates vocabulary and language models for radiologists. The SP 6000 features an initial English vocabulary of 25,000 words; it is a speaker-specific system, and it needs at least 45 minutes of training to the individual’s speech patterns. It includes word processing functionality and the usual dictation machine features. The dedicated Speech Processing Unit runs on an 80486-based personal computer running Windows and the SP 6000 is designed as a shared logic system with a network of personal computers linked by NetWare. The computers can serve as dictation or editing stations when outfitted with either special Philips microphones or conventional headphones and foot controls. No indication of any price was given for it.