Here's an extract from the VoiceAttack log;
08:16:24 - Unrecognized catch-all : 'But it doesn't seem enough'
08:16:27 - Unrecognized catch-all : 'but whatever and'
08:16:28 - Catch-all for 'book now have any' invoked, but rejected with confidence level 1/10.
08:16:29 - Unrecognized catch-all : 'both'
08:16:31 - Catch-all for 'below' invoked, but rejected with confidence level 6/10.
08:16:34 - Catch-all for 'below' invoked, but rejected with confidence level 2/10.
08:16:36 - Unrecognized catch-all : 'below'
08:16:37 - Catch-all for 'by the' invoked, but rejected with confidence level 0/10.
08:16:38 - Unrecognized catch-all : 'bone'
08:16:44 - Unrecognized catch-all : 'but have have'
08:16:46 - Unrecognized catch-all : 'no thing no of'
08:16:49 - Catch-all for 'Bolton' invoked, but rejected with confidence level 1/10.
08:16:56 - Unrecognized catch-all : 'no the'
08:16:58 - Unrecognized catch-all : 'bone'
08:16:59 - Recognized : 'no' (Confidence 96)
08:16:59 - Listening suspended
All I was actually trying to say (as you can see) was "no". If it can recognise that (eventually!) with 96% confidence - what's with all the other rubbish? It has similar problems with "yes" where it will say it has heard gibberish like "BS" and "GS".
Any ideas on how to improve accuracy? I've done all the training for the voice recognition but it doesn't seem to help VA much.
Thanks!