Is Asia edging ahead in the translation technology innovation stakes? News from Japan and Korea suggests that automatic speech translation (i.e. not written) will go live quite soon. And India is experimenting with language interpretation by radio. In Europe, 2005 kickoff news sounds distinctly underwhelming.

In the wake of the announcement that Korean mobile maker Pantech&Curitel’s latest handset features text to speech technology (whose, I wonder):

The TTS capabilities of the handset will convert incoming text messages to voice messages, and read them aloud as they arrive in loudspeaker mode. Also, users can control and have the tiny clamshell read the contents of its menu, phone book and unanswered calls.

and extends the fad for cameras with Optical Character Reader software to manage the local biz card ritual:

which can recognize text from JPEG files. This functionality is specifically aimed at those handling large numbers of business cards, which rather than requiring storage can simply be photographed with the built-in camera and parsed into the phone book of the device.

comes the suggestion in the Yomiuri Shimbun that Japan’s government-led “multilingual audio-text translation system” is likely to start testing soon and be completed during 2005. The plan is to offer a translation service probably for business folk to start with between Japanese, Chinese, Korean, and English over mobile devices such as phones and PDAs, using a database of ‘500,000 phrases or 5 million words’ (presumably meaning a million or so in each language). I can’t find any information on who is actually doing the development but suspect that ATR will be playing a key role for the ‘audio’ segment.

Meanwhile in Korea,

VoiceBuzz reports that two telecoms firms, SK Telecom and KTF, are to introduce in March two separate services that can translate Chinese, Japanese and English. Will this actually be based on the same underlying technology that will drive the Japanese government project? Although the German Verbmobil program looked well placed to deliver an operational speech trans system when it began in 1996, Japanese, Korean, US and EU speech translation teams have been working loosely together and sharing ideas in the C-STAR consortium. The ongoing European Union effort called TC-STAR for Technology and Corpora for Speech to Speech Translation started work more recently on new ways of building a talking translation machine. So while in Europe it’s been research, research and more research, it looks as if Japan and Korea – and via them China – will be the first to actually market an operational system.

In a democratic and massively multilingual country like India, access to linguistic services in public debate forums for people who can’t always read is now being provided by an innovative use of FM radio. According to this report, audiences at conferences can listen to simultaneous translations of the speeches by an interpreter, transmitted over a low band frequency from a control room in the conference center. This solution was invented by Hemant Babu and pioneered by members of Nomad, a grassroots organization that is developing tools for digital translation and archiving for non-commercial translation requirements.

The FM based translation system was first used during the World Social Forum 2004 in Mumbai, reputedly bringing down the cost of translations from US$223,000 to US$13,000. It was used again at Social Forum for South America at Quito, Ecuador in July 2004, and will be implemented next week at the World Social Forum at Porto Alegre, Brazil.

By comparison, gobsmacked you probably won’t be at the news from equally multilingual Europe that a project called TransType 2 , “an innovative computer-aided system that allows rapid and efficient high quality translations” is well…nearing completion. Using a mix of translation and machine translation technologies,

The TransType2 prototype is currently designed to assist translations from English to French, Spanish and German and vice versa, although additional European languages can be incorporated relatively simply. “To add Chinese or Arabic, for example, would take more research but it is possible,” José Esteban (project manager) says.