View Full Version : Plugin thought: vocal interface

2005-03-08, 16:38
On Radio 4's "In Touch" program this evening (known in our house as "Does He
Ever Stop Whinging?") there was excited discussion of a new DAB radio with a
speaking interface. It obviously represents a considerable improvement on
existing radios - but seemed very limited. (In particular it sounded as if
it had a limited set of pre-recorded phrases, and beyond that resorted to
spelling everything out - eg "You are listening to. BBC Radio 4. O N A I R
COLON A F T E R N O O N P L A Y"! Also, not everything that is shown on the
display can be read out.) None the less, all the presenters and users were
bowled over by it.
(it's the first item)

The presenter said that they'd spoken to 14 other major manufacturers of DAB
radios and found only one that had plans to bring out a speaking model.

The SqueezeBox, I guess, is already pretty good for those with limited
vision, since the display is so bright compared to the LCDs on most digital
radios, and the size of the text can be so simply adjusted up to to huge.
And perhaps it would be worth promoting it in appropriate forums (especially
if more of the AlienBBC type functionality is rolled in to the standard
installation!) on this basis alone.

But it occurred to me that it might be relatively simple (easy for me to
say, with no intention of getting my hands dirty with the code) to create a
hack for the Squeezebox that could use the host server's speech facility to
read out the display. Obviously there'd be a tricky bit capturing the
output of the native speech function and directing it over the network to
the SqueezeBox. But given that this could be conquered, I would think the
basic functionality would be that when a control was pressed on the remote,
after the the command had been enacted, the next text to be sent to the
display would also be vocalised. So it wouldn't constantly read out
whatever happened to be on the display; but as you navigated through, it
would read the menu command, the name of the album, the name of the station
etc that was being displayed in response to the last key press; and pressing
'now playing' would read out the data on the current track/station etc.

Obviously, I'm speaking out of pure ignorance in making assumptions about
what would be simple; but it seems to me that this effort, expended once,
would then bear fruit all over the shop. So all the functionality of the
squeezebox would be vocalisable, and all the text from the internet radio
stations, thus at a stroke leaping over the new speaking DAB radio.

On top of that, when you return to the SB's original purpose and note what
an incredible access to their music collection a talking SqueezeBox would
give to a blind or partially sighted music lover, there would be a very
powerful story to tell.

Er... there, I've said it. Comments?

Ben Rubinstein | Email: benr_ml (AT) cogapp (DOT) com
Cognitive Applications Ltd | Phone: +44 (0)1273-821600
http://www.cogapp.com | Fax : +44 (0)1273-728866