Speech recognition for experts and laymen now "Schwiizerdütsch
SBB is today adding a voice recognition function to its mobile app. The SBB timetable now speaks dialect. In a pilot project, preview app users can make connection queries in their own dialect. The basis is the self-developed speech analysis solution from Spitch AG, which understands Swiss German with its many local dialects.
Speech recognition is only good if it also understands its users. The new SBB app should be able to recognise a rave breaker such as "Vo Büach uf Wöwe" (dialect: from Bülach to Vevey) without any problems. However, the mobile app only recognizes Swiss German for the time being!
Starting today, customers should be able to request information verbally in the preview version of the SBB Mobile App - and thus enter connections more quickly.
Until now, speech service-bound applications could not understand, let alone process, speech input in dialect. Stephan Fehlmann, Business Development Manager Spitch AG, explains: "This project illustrates very well how strongly our Swiss dialects differ from High German. Depending on the region and even area of residence, place names are pronounced quite differently and therefore cannot be understood by a general, High German language solution."
Spitch AG's proprietary speech analysis solution understands dialect and is now implemented in the preview version of the SBB Mobile App. Spitch speech analysis is based on a self-learning technology that recognizes, analyzes and understands verbal input in real time.
Typing and swiping unnecessary
Connection polling via voice input allows app users to get quick journey planning information with a simple "from A to B". Not only is this very fast, it's also convenient - typing or swiping becomes unnecessary with voice input. Spitch AG is an international solutions and professional services provider that develops powerful speech analytics technologies for businesses.
Spitch's solutions come from Switzerland and contribute to continuous improvements in the field of speech recognition. This area is experiencing strong growth due to the latest findings in the field of "Artificial Intelligence" and "Machine Learning", which pave the way for more precise SLT applications.
Further regions desired
The speech recognition or speech analysis solution from Spitch AG currently understands around 18,000 stations in the German-speaking cantons of Switzerland. Further expansion to the other language cantons is in progress. In order to test its use extensively, SBB has decided to launch the voice features first in the preview app, which is available free of charge from today in the App Store and Google Play.
During six months, users can give their feedback on the functionality in order to constantly improve the recognition rate. If the function is used actively during these six months, the voice input will also be launched in French, Italian and English. Other input patterns will also be added, such as "last train from Zurich to Bern".
At a later date, it is also conceivable that a ticket can be purchased directly via voice recognition.
You can find further information on the SBB Mobile Preview here
The core team at Spitch AG has over 50 years of experience in the research and development of speech technologies (SLT), including automatic speech recognition (ASR), text-to-speech synthesis (TTS) and information retrieval (IR) - currently also offering "speech technologies for insurance". In addition, the team has experience with both start-ups and established companies.
You can find further information under www.spitch.ch
