Hi Maghi
Just to throw my tuppence worth into the ring.
I have used (or tried to use) just about every dictation product on the
market. At times i can achieve teh 95% or so accuracy they claim.
Yet I cannot get used to them - and as soon as I have to correct something
using the keypad my fingers remain there!
I have also heard claims that at their current accuracy levels they are
about 3 times slower than a poor typist. This is simply because any time
gained in inputting text is more than lost when the text has to be
corrected.
In any case comparing typing speed is probably irrelevant at this stage.
Probably the most important question concerns the perceived effort of using
such systems.
This would be extremely easy to measure - I would set it out as follows:
1. Select 12 people with a range of typing skills (e.g. 4 non typists), 4
people untrained as typists but frequent user sof keybaords and 4 touch
typists
2. Select 12 tasks - 4 copy typing tasks, 4 'creative tasks' (e.g. type a
short essay on your holidays) and 4 copy dictation tasks
3. Select 3 input methods - dictation system, oridinary wordprocessor and as
a 3rd I would suggest a real Wizard of Oz Touchtypist who could actually
understand what was required (and possibly even speak back to query).
After each task ask for an opinion on that task. A scale along the lines of
(this would need to be refined but gives the gist).
A. Extremely easy - no effort required at all
B. Easy to complete but required a little effort.
C. Required effort but was not too difficult to use.
D. Required a great deal of effort.
E. Impossible to use
Once the results are obtained and scored (A = 4; B = 3; C = 2; D = 1; E = 0)
then opinions can be compared.
With this design for example you will have a total of 144 opinions which can
be partitioned into
3 averages from user types - with 48 examples of each (averaged over tasks
and devices)
3 averages from devices types - 48 examples of each averaged over tasks and
users
3 averages from taks types - 48 each averaged over users and devices.
Could be a very telling experiment.
Regards,
Denis johnston
> -----Original Message-----
> From: Maghi King [SMTP:Margaret.King@issco.unige.ch]
> Sent: Sunday, May 16, 1999 12:57 PM
> To: Pascale Fung
> Cc: EAGLESeval@cst.dk
> Subject: Re: EAGLESeval: a question
>
> Hello!
>
> I think the question is implicit: if comparing typing speed to dictating
> speed directly doesn't give you a measure, and if it's hard to separate
> out time spent translating from time spent typing/dictating, how do you
> find out whether being able to dictate instead of type increases
> productivity?
>
> Maghi
>
>
> Pascale Fung wrote:
> >
> > There does not seem to have any question but just their own
> > answers????
> >
> > pascale
> >
> > Maghi King wrote:
> >
> > > Anybody got any good ideas for these two?
> > >
> > > Thank you,
> > > Maghi
> > > --
> > > Please note my new e-mail address (old address was
> king@divsun.unige.ch)
> > > Maghi King | E-mail: Margaret.King@issco.unige.ch
> > > ISSCO, University of Geneva | WWW: http://issco-www.unige.ch/
> > > 54 route des Acacias | Tel: +41/22/705 71 14
> > > CH-1227 GENEVA (Switzerland) | Fax: +41/22/300 10 86
> > >
> > >
> ------------------------------------------------------------------------
> > >
> > > Subject: the letter for the forum!
> > > Date: Fri, 14 May 1999 14:22:04 +0200
> > > From: Daniele Grasso <grasso5@etu.unige.ch>
> > > Reply-To: danielegrasso@mail.com
> > > To: Margaret.King@issco.unige.ch
> > >
> > > Question related to speech recognition programs evaluations.
> > >
> > > We are two translation students preparing a CAT certificate and we
> are
> > > working on an evaluation plan for speech recognition programs. Given
> the
> > > fact that we are aiming at determining whether or not the Italian
> > > section of the Swiss Confederation would benefit from such an
> > > investment, we enquired about translators' touchtyping speed. Infact
> we
> > > thought that knowing how many words they typed a minute, we would soon
> > > discover whether the acquisition of speech recognition programs would
> be
> > > useful to them.
> > >
> > > We were able to get in touch with the translators on the phone, but
> > >
> > > 1) We were not given any hints on how many words were typed a minute;
> > > 2) We were told that translators [and it is far from being surprising
> !]
> > > translate while typing their translation
> > > a) Even if we were to know how many words were typewritten per hour,
> we
> > > would still come up against the dedalus of splitting pure touchtyping
> > > time from translation timel.
> > > b) The simple solution that would be that of translating while
> dictating
> > > would not reflect reality. In fact, recreating a working environment
> > > that had to take into account all the working habits and the know how
> of
> > > these translators would never be faithful to the real ones.
> > >
> > > Thank you for your attention.
> > >
> > > --
> > > Mr. Daniele Grasso
> > > 11 Allee des Cyclamens
> > > F-74240 Gaillard
> > > 0033 4 50 31 85 98(from outside France)
> > > 059 4 50 31 85 98 (from Geneva)
> > > New email address : danielegrasso@mail.com
> > > To send a fax through your email : http://www.tpc.int/
> >
> > --
> > =================================================
> > Pascale Fung
> > Assistant Professor
> > Department of Electrical & Electronic Engineering
> > University of Science & Technology
> > Clear Water Bay, Kowloon
> > Hong Kong
> >
> > http://www.ee.ust.hk/~pascale
> > tel: +852 2358 8537
> > sec: +852 2358 7087
> > fax: +852 2358 1485
> > =================================================
>
> --
> Please note my new e-mail address (old address was king@divsun.unige.ch)
> Maghi King | E-mail: Margaret.King@issco.unige.ch
> ISSCO, University of Geneva | WWW: http://issco-www.unige.ch/
> 54 route des Acacias | Tel: +41/22/705 71 14
> CH-1227 GENEVA (Switzerland) | Fax: +41/22/300 10 86