[senit] Voice recognition

Hi Ira

Ann - 

What you suggest is most likely true. The "British English" being an add in. 
That's one positive thing about ViaVoice, you get to choose to load an 
entirely different speech engine if you pick "British." For myself, I found 
it easiest to start there and add the American pronunciations in as I went, 
but obviously I do not do that with my "American" students. 

And though ViaVoice is not as "high-end good" as Dragon, that "tolerance" it 
has has helped me out. We used ViaVoice with kids in Louisiana post-Katrina 
 - now, that accent is a battle recognition-wise, but we got it to work (that 
was using the Mac version). 

But this forces me to go open up my Vista-equipped laptop, and try training 
it in my best British accent. I'll let you now. 

Ira Socol 

Ann writes: 

> Hi Ira
> It's interesting to see that you have a somewhat different experience with
> the versions of VR than most people over here. I wonder if these differences
> may be due, at least in part, to the different pronunciation of sounds/words
> on either side of the Atlantic? We use 'British' versions of the software,
> which are different from the original US versions. I can recall purchasing
> an early version of Via Voice in America and having GREAT difficulty
> training it to my voice despite much time and effort. I found that when it
> utterly refused to produce a sensible result, all I needed to do was to say
> the words in my best attempt at an American accent - and it wrote sense!
> Presumably the software is originally written for an 'American voice' and
> then adapted to take account of the different stress, pronunciation of
> sounds in Britain. How well the software is 'adapted' might account for the
> differences in our success with the software...... 
> 
> Regards
> Ann McDevitt  
> 
> -----Original Message-----
> From: senit-bounces at lists.becta.org.uk
> [mailto:senit-bounces at lists.becta.org.uk] On Behalf Of Ira David Socol
> Sent: 13 September 2006 14:57
> To: senit at lists.becta.org.uk
> Subject: Re: [senit] Voice recognition 
> 
> I have had very good success with ViaVoice (on both Macs and PCs - on PCs 
> the "Pro-USB" edition) with children as young as third grade. I have also 
> had good luck with SpeakQ, and I'm currently having a great time with the 
> Beta version of Windows Vista - which has a great voice-to-text engine built 
> 
> in. I've had much less success below adulthood with any Dragon product - not 
> 
> sure why, although the voice training seems harder and less forgiving.  
> 
> Of course this does not work well in the classroom, except maybe with use of 
> 
> that World War pilot-style "Voice Sylencer" microphone - 
> http://speedchange.blogspot.com/2006/03/csun-2006speech-recognition_23.html
>  - which would embarrass most kids.  
> 
> The key is training, put the training on an iPod (or just a cassette) or 
> whisper each line so they can repeat it - eliminating the need to read 
> component. This is easier in ViaVoice than Dragon just because of the design 
> 
> of the interface, and un-needed in SpeakQ where the software reads it to the 
> 
> person whose voice is being trained.  
> 
> Anyway. This has been great for a lot of children. I find that it seems to 
> improve reading as well, since students learn sightword recognition rapidly 
> when they are speaking and seeing the words appear on the screen.  
> 
> Last thought. Yes, especially among secondary students there are pushy 
> parents more interested in this (especially in Dragon because it costs so 
> much) than their children. But this can sometimes get leveraged to your 
> advantage - if you build the right connection with the child.  
> 
> Ira Socol
> Special Education Technology Scholar
> CEPSE
> College of Education
> Michigan State University  
> 
> Mitchell, Dave (ChS, LBSS) writes:  
> 
>> Chris  
>> 
>> I would say the technology works.... but not in class...  
>> 
>> ...  with ongoing support from adults who understand what is required....
> by students who can take substantial responsibility for setting up their PC
> each time they want to use it (get it out of the bag, open applications, put
> on headset.....) and can correct errors (or use other supportive software to
> assist them..; see WWW Keystone or Clarosoft products)  .... used at home or
> in sessions out of class in school ..... 
>> 
>> you can see I am adding lots of qualifications to my initial
> assertion..... 
>> 
>> I have a report of a small scale study completed for BECTa some years ago
> but not published using portable digital recorders with Dragon Naturally
> Speaking... its a cautionary tale. want a copy? 
>> 
>> Dave Mitchell  
>> 
>> Principal Teacher
>> Assistive ICt & AAC  
>> 
>> -----Original Message-----
>> From: senit-bounces at lists.becta.org.uk
>> [mailto:senit-bounces at lists.becta.org.uk]
>> Sent: 13 September 2006 08:35
>> To: senit at lists.becta.org.uk
>> Subject: [senit] Voice recognition  
>> 
>> 
>> I've been asked about speech recognition as a means for a student to
>> 'write up their work at home'.
>> Has anything happened recently to suggest that the panacea has finally
>> arrived?
>> I know Wrawickshire DISCS ran a project but I can't find any info on
>> their site.
>> Becta doesn't offer anything of note.
>> Has anyone successfully used any systems with mainstream secondary aged
>> pupils?
>> Any suggestions please folks, eother of research to read or first hand
>> experience?  
>> 
>> 
>> Thanks
>> Oh yes and is there any corelation between pushy parents who think ICT
>> is THE solution and disinterested students?  
>> 
>> Chris  
>> 
>> Chris Howles
>> Broadband Sandwell
>> Sandwell TDC
>> Popes Lane
>> Oldbury
>> West Midlands
>> B69 4PJ
>>  
>> Tel: 0121 569 2400
>> Fax: 0121 511 1022
>> **********************************************************************
>   
> 
>  
> 
>  
> 
>  
> 
> 

Ira Socol
Special Education Technology Scholar
CEPSE
College of Education
Michigan State University 