17. Februar 2006
Trying to cut your way through rush hour Paris, you’re running late for an important meeting. With all these traffic and intersections, thinks are getting pretty hectic. Fortunately your navigation system, prevents you form missing exits. It supports sign posts and says “Turn left on A13 heading towards Bois de Boulogne”, making it easy for you to chose the right exit. Another message is spoken: “Traffic jam close de ‘l’Arc de Triomphe”. Your route is then recalculated. Your navigation system helps you avoid the jam.
Making it easy for drivers to conquer rush hour traffic in unfamiliar cities or warning them about potential delays and traffic jams is just one application of Text-to-Speech (TTS) -based speech output solutions
In contrast to navigation systems that rely on pre-recorded speech, systems featuring TTS can make full use of dynamic content provided by services such as TMC or DAB.
Clearly, TTS opens up a whole new world of customer benefits in terms of usability, convenience, comfort and safety. TTS enabled phone book and communication applications like SMS or email are getting even more popular. Text-to-speech technology is no longer a niche product restricted to upper class models. Increasingly it has been showing up in the high volume middle class models. DaimlerChrysler, PSA and Nissan have all been deploying speech output solutions featuring TTS. Customers are demanding it!
Many OEM’s and after market players finished quite complex Text-to-Speech projects. To successfully finish the projects they all followed some common principles:
One single voice for all applications:
End customers prefer the same voice for each in-car device, such as email and SMS reading, navigation system and board computer. Ending up with multiple different voices for each application will damaging the customer’s brand experience.
System architecture with central speech output device:
To have full control over the speech output a single device being able to handle dynamic and fixed announcements (e.g. TTS and prompts) at the some time is preferred. This contributes to reduced costs by keeping memory requirements low and furthermore contributes to increased system stability. Furthermore, a single speech output device supports a simple update during phase lift.
Strict control of development times and resources:
Last minute changes are a reality. Thus, an experienced TTS partner provides means to act flexible in all project phases! Development processes but also the software and system architecture need to be designed for the required flexibility to avoid delays in the start of production or even non-availability of the TTS and speech features. Key is the early availability of prototypes for the OEM to communicate at early project stages how the final speech output system will sound.
The outlined principles contribute to an efficient development process with clear responsibilities. Thus, quality control is easy to implement: Errors are avoided or become obvious at early stages. Error analysis benefits from the clear system architecture.
Successful projects contribute and also drive medium and long-term product development cycles. They also have their effects on partnerships and in most cases emphasize a Mutual Innovation Strategy.
These reasons illustrate the importance of partnering with a TTS supplier experienced in auto development processes. Taking advantage of the learnings their supplier has gained by integrating similar projects will help through the inherently innovative process and ensure steep learning curves.
Martin Reber, SVOX’s Director of Embedded Solutions is a veteran in terms of automotive TTS projects. He gives three hard and fast rules to be followed when planning the switch to a TTS-based speech output system:
1. Share the know-how: Before any project work begins clearly state and demonstrate what TTS technology can and can’t do for all applications at each project phase. As a result of this phase key innovations and characteristics of the new speech enabled product can be defined by the OEM.
2. Implement the speech output system: Based on precedent first step, a clear definition of objectives and system specifications can be created. Today’s best practice favours one central speech layer for all applications, hence reducing memory requirements and system risks, and also ensuring brand consistency - one voice.
3. Control the project: In spite of clearly defined procedures and interfaces for the project flexibility remains important. Think about possible change requests and upgrades in early project phases to minimize resource requirements and costly last-minute changes.
Of course, this requires the selection of an experienced speech output component partner with a track record of several large scale automotive projects. Along with the educational function that such supplier can provide, they should also be willing to become the long-term speech component partner. This involves taking on responsibility for conception and implementation of detail specifications, as well as product and process innovation. The result: a lower overall cost and higher quality as the partner works from a more global point of view.
Zurich-based SVOX believes that by defining best practice standard engineering processes, they have helped their partners cope with the uncertainty and complexity of integrating TTS speech output solutions.
More infomation under www.svox.com
lluminating the road to in-car speech integration