Marcel Wassink, Managing Director, Philips Speech Processing
Tell us a little bit about yourself. How long have you been with Phillips and in the speech industry?
Marcel Wassink I have a master's degree in Computer Science, but have always been more interested in the benefits of technology rather than in the technology as such. Therefore, I have spent most of my career in sales and business management, first within Fujitsu Software, and then with Philips Speech Processing. I joined Philips in 1999 to start the professional speech business, in which we have achieved significant growth of revenues.
A year ago Philips sold its speech processing telephony and voice control businesses to ScanSoft. How has this affected the Philips Speech Processing dictation business?
MW Large vocabulary dictation recognition has always been the focus for PSP. From a technological point of view it is probably the most difficult deployment of speech recognition technology, since you have to enable users to dictate in a natural manner, without having to adapt, for example, the speed at which they speak or their accent, to the speech recognition system. Given these facts, the sale of the telephony and voice control businesses was a logical step that allowed us to fully concentrate resources on our core segment, which is to provide specialized and tailored digital dictation and speech recognition solutions for professional users. The result of our focused strategy is the market leadership in the medical and legal dictation recognition markets. In addition to SpeechMagic, which is our speech recognition and digital dictation software for professional document creation, we also offer a Speech SDK to enable other companies to integrate the latest Philips speech recognition features into their own applications for deployment in various fields and industries.
Philips Speech Processing (PSP) has recently increased its activities in the U.S. and has established a network of partners. What are PSP's goals in the U.S.?
MW As well as in Europe, we want to become the market leader in the professional dictation market in the U.S. We work with all big dictation vendors like Lanier Healthcare, DVI, Dolbey, WinScribe and Crescendo. We also used to work with Dictaphone, but when they were acquired by L&H, unfortunately, that relationship had to end. On the other hand, in support of transcription services, we are working closely with the market leader in the U.S., MedQuist. They have a large customer base in the medical market, and in close co-operation we are setting up the biggest ASP Speech Recognition service in the U.S. medical market. All our partners benefit from the strength and dedication of a global player with its own speech recognition engine that ensures that market developments and user specific engine requirements are quickly reflected in the technology.
Could you give us an update on PSP's partnership with WinScribe? How vital is it for the speech industry to form partnerships like this?
MW Similar to Lanier Healthcare, DVI, Dolbey and Crescendo, WinScribe has integrated speech recognition into its digital dictation solution and offers it to the medical and legal industry. Philips and WinScribe Europe have been working together for quite a while. Following the success in Europe, WinScribe and Philips formed a global integration and distribution partnership for SpeechMagic. As WinScribe has more than 50 resellers in the U.S. alone, our partnership with them as well as with the other dictation vendors are important as they enable us to quickly penetrate the U.S. market. By integrating speech recognition, dictation vendors can offer their customer base a more efficient and cost-effective way to create documents - so speech recognition is the natural "next step" for dictation vendors.
What is PSP doing to explain its technologies to end-users? Is this a difficult task?
MW We work jointly with our partners to explain the benefits of our technology. It is not difficult to explain this as intuitively every heavy typist or author understands the benefits of speech recognition. We also offer an ROI analysis to our customers so they can rationally calculate the ROI time and see how quickly they will benefit from speech recognition. But one of the most powerful tools for proving how straightforward speech recognition is and how far it has developed is to give potential customers live experience by letting them dictate anonymous reports at trade shows and events - without training or any other preparation.
PSP has been involved in dictation technologies for 40 years. What do you predict for the future of the dictation and speech recognition market?
MW We think this market is still in its early stages and will continue to grow year on year. The basic speech recognition technology is mature, however, over time, we will be adding more intelligence to the system. This is to support professionals who are not used to dictating and who often make mistakes and correct themselves while dictating. Also, when a patient record shows that there has been a problem with the right wrist of the patient, and suddenly the author starts talking about the left wrist, the system will be able to warn the user of the discrepancy. It comes down to "recognizing what a person meant," rather than literally recognizing what somebody said.
Can you explain the characteristics of professional speech recognition systems and how PSP differentiates from other professional systems?
MW While for home-users it is mainly the ease of use and the price that influences purchasing decisions, things look different for professional systems. There are three major differentiation aspects between the two. Firstly, integration and network capability allow the speech recognition software to melt with existing IT systems. This results in a homogenous IT infrastructure. Further networked administration saves resources and reduces costs, which is significant in large organizations with hundreds of speech recognition users. Secondly, the availability of specialized recognition vocabularies. The quality of these vocabularies determines the recognition rate of specialized terminology. Home users normally just use a general correspondence ConText, which is not sufficient for professional applications. And the third factor is the scalability of the system. Some clients start with speech recognition in one department and when they see the success, they want to rollout throughout the organization. SpeechMagic is scalable for up to 4,000 users and 1,000 hours of dictation throughput a day. With SpeechMagic being developed exclusively for professional users, Philips provides all the capabilities a professional system should offer. It supports various input devices and correction workflows, giving organizations the flexibility to choose the most efficient and cost-effective way of working. In addition to the speech recognition software, Philips also provides the hardware - from input devices to foot pedals - so that the customer receives a complete range of dictation and speech recognition products from one specialist.
Last month PSP announced it received an order for the installation of SpeechMagic in nine German convalescent clinics. What is the significance of such a deployment?
MW In 1999 we started to focus on radiology in Germany, France, the UK and the Benelux countries. In the meantime we have made SpeechMagic available in 22 languages and have made a lot of progress in this niche: in some countries, over 60 percent of radiologists use our speech recognition engine. Later on, in the mature radiology markets, we have started to expand onto other specialties like pathology, surgery, cardiology and now we see this moving into a pull for hospital-wide usage of speech recognition. In that sense, this order was another sign of our, and our partner's, success in the medical market and also of the success of speech recognition as such - because based on practical experience and word-of-mouth, speech recognition has proved that it has matured and that it leads to a measurable productivity increase and cost-saving.
Companies and Suppliers Mentioned