Text2Go Blog

Microsoft Anna – Sensual Babe or Just Another Robot?

Microsoft Anna is the new English text to speech voice of Windows Vista. She replaces Microsoft Sam, the robot-like voice that comes with Windows XP. Microsoft Sam has probably single-handedly done more than anything to give text to speech a bad name. The only positive thing I can say about Microsoft Sam is that he reminds me of the Commodore Amiga, possibly the first personal computer to have a built-in text to speech capability.

Does Microsoft Anna sound any better than Sam? Does she have a sensual voice that you could listen to all day or does she have you reaching for the mute button? How does she compare to a voice from a leading speech engine provider, such as one of Nuance®’s RealSpeak™ voices?

To answer these questions, I’ve conducted a simple ‘speak-off’ between Microsoft Sam, Microsoft Anna and RealSpeak™ Samantha. Each voice had to read three passages of text, ranging in difficulty from a simple news story to an abstract from a medical journal that most humans would have trouble pronouncing.

The resulting samples can be listened to using the following link.

	Microsoft Sam	Microsoft Anna	RealSpeak Samantha	Sample Text
Sample 1				SCIENTISTS have explained mathematically why the famous silly walks of Monty Python’s John Cleese have never caught on in the long history of homo sapiens. The giant, leg-twirling strides of silly walks may enable an individual to leap around swiftly but are simply too expensive in metabolic energy compared with conventional locomotion, according to a paper published by Britain’s Royal Society.
Sample 2				The color of a vegetable means something more than what it looks like. The colors of vegetables are affected by the vitamins and minerals contained therein. Having a literally colourful diet is important for your health. A blueberry and a purple carrot both provide your body with antioxidants, blueberries are in anthocyanin pigments which are powerful antioxidants. And carrots contain carotenoid pigments, the darker the colour of a carrot means that this carrot contains a higher concentration of carotenoids than say, an orange carrot.
Sample 3				Similarities in physiological roles of LXR (liver X receptors) and co-repressor RIP140 (receptor-interacting protein 140) in regulating energy homoeostasis and lipid and glucose metabolism suggest that the effects of LXR could at least partly be mediated by recruitment of the co-repressor RIP140. In the present study, we have elucidated the molecular basis for regulation of LXR transcriptional activity by RIP140.

The verdict?

Microsoft Anna certainly has a more natural and lively voice than Sam, however there are a number of problems. The pacing of the speech is very uneven. Some words and syllables are spoken very quickly and others are quite slow. The intonation also seems to be quite wayward at times, with some words being unnaturally emphasised. Microsoft Sam mispronounced the word ‘co-repressor’, pronouncing it instead as ‘company repressor’, confusing ‘co’ as an abbreviation for ‘company’. Microsoft Anna did not fall into the same trap. However when compared to RealSpeak Samantha, you can hear that Microsoft Anna is a long way off state of the art.

Overall I believe that Microsoft Anna is a small step in the right direction. It’s just a shame that it wasn’t a larger step. She’s a long way off sounding like a sensual babe. At best, she’s could be thought of as a cyborg with a malfunctioning speech unit.

Microsoft Sam provided one little gem when trying to pronounce ‘transcriptional activity’ in the 3rd sample, doing a great impersonation of a cylon from Battlestar Galactica!

I’d like to hear your thoughts on Microsoft Anna.

Robot picture courtesy of peyri.