KeyStone ScreenSPeaker
and SPeechMaster:
Frequently Asked Quesitons
KeyStone ScreenSpeaker and SpeechMaster: Frequently Asked
Questions
[Note: Dragon NaturallySpeaking is referred to throughout this
document as DNS. Microsoft Word is referred to as Word and
relates to Word 97 (Word 8) and Word 2000 in the Office
97 or Office 2000 suite, respectively.]
Q: What is the difference between
ScreenSpeaker and SpeechMaster?
A: ScreenSpeaker is a special type of screen reader which is designed
specifically to assist dyslexic people when they use voice recognition
equipment. It is also useful for visually handicapped people to use as a more
general screen reader.
SpeechMaster is a boxed product containing ScreenSpeaker and
the DNS Preferred (version 4) voice recognition system. The
microphone is included with the DNS software. The package includes 60 days of
e-mail support for ScreenSpeaker. Support for DNS is provided by Dragon Systems.
ScreenSpeaker is ideal for people with reading problems and DNS
helps those with writing problems; put them together and you have the
ideal package for dyslexic people.
Q: What has happened to KeyStone Professional, Keystone 99 and KeyStone
2000?
A: KeyStone Professional was designed for Windows 3.11 and was
modified slightly in 1997 to operate in Windows 95 and Windows 98.
The program was completely rewritten in 1999, was renamed KeyStone 99 and
was made available last summer. As the Millennium date approached it was
considered prudent to rename it KeyStone 2000. Due to confusion over
whether the KeyStone product included voice recognition it was decided to
establish two new product names: KeyStone SpeechMaster and KeyStone
ScreenSpeaker, supplied with and without DNS voice recognition
respectively. Some of our existing leaflets still refer to KeyStone 2000.
Q: What are ScreenSpeaker Plus and SpeechMaster Plus?
A: ScreenSpeaker Plus and SpeechMaster Plus are
both supplied with KeySpell at a discounted rate. They are primarily
intended for dyslexic users who find that they need additional help in checking
and correcting spelling.
ScreenSpeaker Plus is primarily for those typing and using Word .
A highlighted word or phrase in a Word document can be passed to KeySpell
where it is checked. An alternative word or phrase can be passed back to the
document in Word by a simple mouse click. KeySpell can reinforce Word’s
own spelling checker or it can be used instead of it.
SpeechMaster Plus has the same function as ScreenSpeaker Plus.
However in addition, a word displayed in the DNS correction box can be
passed to KeySpell for checking. An alternative word or phrase can be
passed back to the correction box and selected in the normal way, completing the
correction process.
Q: Can I use DNS Professional instead of Preferred and,
if so, how can I purchase them?
A: Simply ask for KeyStone SpeechMaster Professional.
The cost of KeyStone SpeechMaster Professional is adjusted
for the more costly version of DNS.
Q: If I already have a version of DNS can I use it with ScreenSpeaker
and if so, will it work in the same way as SpeechMaster?
A: Yes, provided that you have version 4 of DNS Preferred or
Professional, it will work in the same way as SpeechMaster once you load ScreenSpeaker.
Previous versions of DNS are not recommended but should operate
satisfactorily. The benefits of version 4 over version 3 are significant; an
upgrade pack is available.
Q: What do I gain if I obtain DNS Preferred or Professional
instead of Standard?
A: A separate leaflet is available which tells you the difference between the
various versions of DNS [or click here
to see the difference in DNS products]. KeyStone users will see little
difference between Preferred and Professional. Preferred
has two additional features not present in Standard: text-to-speech and
voice playback. KeyStone has its own text-to-speech system which is much
more versatile than the one which is provided with DNS. If you require
voice playback then you should order DNS Preferred.
The main difference between Preferred and Professional is that
the latter provides the facility to set up macros (special voice-driven
programs). These enable the user to speak a command and initiate an entire
sequence of operations, e.g., initiate a mail merge from a specific file into a
specified document and print the results onto pre-selected labels, run automated
tasks by voice, insert blocks of formatted text, etc..
Q: What is the minimum machine specification to use ScreenSpeaker and SpeechMaster?
A: You will need a PC having a Pentium I processor with a speed of at least
166 MHz and 32 Mb RAM to use ScreenSpeaker. For SpeechMaster you
will require a Pentium II processor with a speed of at least 500 MHz and 64 Mb
RAM.
You will need at least 96 Mb RAM on a Pentium processor running at 500 MHz if
you are working with Word. However, The Literacy Center currently recommends
a Pentium III processor with a speed of at least 500 MHz and 128-256 Mb RAM for
good performance of SpeechMaster with MS Word.
Q: What is the recommended PC specification to use ScreenSpeaker and SpeechMaster?
A: With ScreenSpeaker alone, you will not see any significant change
in performance with a more powerful processor or extra memory (RAM) unless your
PC is running several programs at the same time, when the screen reading
performance may improve.
SpeechMaster will benefit considerably from extra processing power and
memory. You are recommended to have at least 132 Mb RAM on a Pentium III
processor running at 500 MHz if you are working with Word.
Q: What problems might I experience if I do not use the recommended
specification of PC to use SpeechMaster?
A: You will find that the PC may well run very slowly if you have several
programs loaded at once. In extreme cases, the machine may appear to
"freeze" so that although it has not "crashed" in the strict
sense, the computer processing is effectively suspended.
If you have less than 96 Mb RAM, SpeechMaster will tend to run very
slowly if you use it with Word. The recommended amount of memory for
smooth operation is a minimum of 132 Mb RAM.
Speech recognition performance may suffer if your PC falls well below the
recommended specification.
Q: Is there anything I can do to speed up the operation of SpeechMaster
and make it more accurate with a given machine specification?
A: Yes, to a limited extent. In general, if you are experiencing problems
with the default settings, you would be well advised to make the small extra
financial outlay and upgrade your PC. However, there are settings you can change
which will make the processing faster at the expense of accuracy of recognition
and vice versa. You cannot improve both at once.
Q: I understand that I can restrict myself to speaking text into the DNS
document window when using SpeechMaster instead of using Word.
What are the benefits or disadvantages of this?
A: The benefit is that on a less powerful machine the display of text on the
screen will be quicker and may be slightly more accurate than if Word is
used. The disadvantage is that you will have to transfer text into other
application packages such as Word rather than dictating into them
directly. Many of the very useful functions such as "Bullet that"
which are available when working with Word are not available when working
in the DNS document window.
Q: I have heard that voice recognition systems are susceptible to background
noise. What can I do about this?
A: If you use a noise-canceling microphone and or a microphone with USB, [click
here and click here for more
information] and your PC has a sound card which is recommended for use with SpeechMaster
you should not experience problems. If you are experiencing difficulties and you
are entitled to technical support, these can be resolved through contacting Words
Worldwide.
Q: What is meant by "continuous speech"? Can I speak as fast as I
can in conversation?
A: Continuous speech systems enable a competent user to dictate without
pausing between words and phrases. The earlier discrete systems required a small
pause of about a tenth of a second between words, resulting in a slightly
stilted delivery.
Speeds of over 140 words a minute of accurately recognized text can be
achieved. Dictation must be well articulated but not artificially so. Any
tendency to mumble, gabble or slur the speech will reduce the accuracy of
recognition considerably.
Q: I believe discrete speech systems have been around much longer than
continuous systems. Are they therefore more accurate and more stable in computer
terms?
A: DragonDictate was launched in 1988 and is still available. Other
than DragonDictate and Kurzweil VoicePad discrete systems have
disappeared from the market. Discrete systems are extremely robust and will work
on almost any modern IBM-compatible PC. For certain types of use it is still the
most effective: for example when use is restricted to issuing terse commands
such as "Load file XYZ", "Bold that", "Switch to
Microsoft Excel", "Move down 6 lines", "Delete next 6
words". Its advantage for this type of use compared to continuous speech
systems is that it requires minimal amounts of memory and modest processing
power.
The accuracy tends to be rather lower for an experienced user than would be
found if a continuous system was being used. Dictation will be about half as
fast and will be slightly more of an effort. The systems are arguably rather
more robust than continuous systems from the processing point of view, i.e. one
tends to experience rather fewer system-related issues.
Q: What level of accuracy can I expect from SpeechMaster?
A: DNS will provide recognition accuracy between 95% and 99% from the
outset, depending on the level of previous experience that the user has in using
recognition technology and depending upon the system that is being used to run
the software. As the user continues to correct errors, accuracy rates increase
steadily.
Q: I understand that some users, particularly dyslexics, may have difficulty
in speaking in continuous mode. What can be done about this?
A: It is impossible to determine at the outset whether a dyslexic person or a
person with learning difficulties will find it more appropriate to use discrete
or continuous recognition. If the user does prefer to speak in discrete word
mode, i.e. a word at a time with a pause between them, DNS will permit
this without penalty. However, many dyslexic people will find that the
correction technique for adjusting for recognition errors falls short of what
was possible with DragonDictate.
For users who wish to have a high level of support in using the correction
technique it is recommended that they purchase KeyStone ScreenSpeaker
and DragonDictate, along with appropriate support arrangements.
Q: Do I qualify for an upgrade if I discover a fault in the program?
A: All products have been rigorously tested but it is in the nature of
computer products for some faults to remain hidden, only to be exposed when a
user carried out an unusual operation.
Any fault which is reported with full documentation will qualify the owner of
the product to a replacement copy of the program free of charge as soon as a new
version becomes available.
Q: How can a large group of users utilise ScreenSpeaker of SpeechMaster?
A: Any number of users can use either product. In the case of ScreenSpeaker,
individual user settings can be retained through password control, accessed via
the Settings Sub-menu in the Windows Start Menu. SpeechMaster user
settings can be retained in the same way but speech files are also identified
specifically for any particular user.
Q: Can young children use ScreenSpeaker and SpeechMaster?
A: Yes, children soon learn how to control the ScreenSpeaker
functions. DNS can cater for voices of all ages and the special SpeechMaster
enrollment scripts help children train their voices in the recognition system.
For practical purposes, a minimum age of six years can be taken as a reasonable
guide.
Q: I often hear about "enrollment" and "training" of
speech systems. What does this mean?
A: "Enrollment" (sometimes also referred to as "training")
of the speech system is where the user provides examples of how he or she speaks
by reading from a series of sections of text displayed on the screen. DNS
collects information from the spoken text and this indicates how the user will
form particular sounds when dictating.
Q: I have heard that some dyslexic users have great difficulty in enrolling
on speech systems. If this is true, can anything be done?
A: SpeechMaster has a number of special enrolment scripts which
consist of short phrases and sentences instead of lengthy paragraphs. The
dyslexic user is therefore not confused by a mass of text. In addition, SpeechMaster
contains extra scripts to those included normally with DNS, giving
greater scope for finding one that suits the user.
SpeechMaster also speaks out the short phrase or sentence so that those
who cannot read the text simply have to repeat what they hear.
The user can speak the enrolment text a word at a time if they are having
difficulty although initial accuracy of recognition may be somewhat affected.
Q: Can a person with speaking difficulties or a strong accent use SpeechMaster?
A: Provided that the user speaks reasonably consistently and smoothly and has
enrolled on DNS carefully, good results can be obtained. In the case of
dysarthric speech it is harder to estimate the likely success at the outset.
Slurred speech is easier to cater for than broken speech, where there is a pause
within the word. A stutter can be coped with but this presents a significant
problem.
Where DNS fails to cater for speech problems, DragonDictate
should be tried. In many cases those with quite severely dysarthric speech and
bad stuttering can succeed in using ScreenSpeaker with DragonDictate.
Q: Do I need support and training? If so, what is available?
A: Sufficient telephone and e-mail support is provided free with the product
to enable a person with reasonable computer knowledge to install and operate the
software satisfactorily. Those who have limited computer knowledge, especially
those using SpeechMaster who have no experience with voice recognition,
are strongly recommended to purchase additional training from an experienced
trainer. It is also recommended that those with minimal computer experience
purchase the annual support package (please call for details).
At least one day’s training is recommended, especially for those who are
managing a group of users, for example at a school or college. Whereas a user
may be able to cope with a lack of knowledge and/or experience, a person
responsible for a number of users will find this inadequacy stressful and will
lead to reduced confidence and inefficiency amongst the user group.
Free telephone or email support for ScreenSpeaker is available The
Literacy Center for 21 working days from the day on which the user installs
the product. A further 60 working days free support is available via e-mail from
The Literacy Center. No specific support is provided for those using a
speech recognition system with ScreenSpeaker unless an appropriate order
has been placed (this is provided by Dragon Systems).
Free telephone or email support for SpeechMaster is available The
Literacy Center for 28 working days from the day on which the user installs
the product. A further 90 working days free support is available via e-mail The
Literacy Center. No specific support is provided for those using a voice
recognition system with ScreenSpeaker unless an appropriate order has
been placed (this is provided by Dragon Systems).
An annual support agreement is available from The Literacy Center for ScreenSpeaker
for SpeechMaster (please call for details).
Training is available for $135 per hour from The Literacy Center on or
off-site. Group trainings can be arranged by negotiation where suitable
equipment can be provided at the trainees’ cost at the training site; per
capita rates will be considerably reduced.
|