Posted on 14-Mar-2004 12:40
Filed under: News
Telstra Mobile post-paid and pre-paid customers can now send Talking Text messages to most Telstra fixed home and business phones. The service converts the text into speech and reads it to the person answering the fixed phone, who will soon be able to choose to reply from a series of simple, pre-set text responses.
Telstra Mobile and fixed phone customers pay no extra monthly fee for access to the Talking Text service and fixed phone customers pay no charge to listen to Talking Text messages.
Telstra's Consumer Segment Head, Jenny Young, said Talking Text messages were easy to send and added "Text messaging is enormously popular. Our mobile users send more than 100 million SMS messages a month and now our fixed phone customers can be part of the phenomenon.
"The computer-generated, female voice reading out the message sounds natural and clear, thanks to cutting-edge technology. The service can even recognise many of the symbols and abbreviations used to represent words in text messages.
"Fixed phone SMS is the next step in the evolution of text messaging. It's one of Telstra's key innovations as voice and text communications converge," Ms Young said.
The ability to receive Talking Text messages has been added to most Telstra fixed phones, but customers can cancel the service, while Telstra Mobile customers sending a Talking Text message to a Telstra fixed phone are charged at their standard SMS rate.
When a Talking Text message is delivered, the fixed phone rings as usual. The person answering hears a message advising they have an SMS. If they choose to listen, the Talking Text service reads the message.
The service uses Speechify, “[an] advanced text processing technology [that] allows it to accurately predict textual ambiguities in a sentence. For example, it can distinguish between words with identical spellings, but different pronunciations. “It can analyse an abbreviation such as ‘St.’ and distinguish ‘street’ from ‘saint’. It can detect the basic pattern of a street address and apply a tag activating a special set of processing rules to greatly improve the rendering of addresses, especially for location and navigation applications.”
The system also increases deployment speed by cutting back on the time and expense of studio recording. The ScanSoft R&D team analysed the voice of Queensland-born singer Karen Jacobsen for the basic phonetic units of human sound - phonemes. They used these to create more than half a million synthetic sounds that make up Speechify Karen’s standard Australian voice.