WPCN  
      2 
      B    V   P    Z       
      Courier 10cpi    ? x x x 
     Sx 6X   @ 8	 X@DOS Text Printer                     DOTEXPRI.PRS x 
   @           q X@    2         <      L                       3| x                                 HOW FAST IS IT REALLY?
                        by David Andrews

     From the Editor: One of the most exciting aspects of the
recent technology revolution for the blind is our increasing
access to Braille. Using a personal computer, a Braille-
translation program, and a printer that generates Braille rather
than ink-print, almost anyone can now produce readable Grade II
Braille. The result is that virtually any computer text file can
today be produced in Braille by the insertion of a few formatting
commands that do not require any knowledge of Braille.
     The quality of the Braille turned out by this process
depends in large part on the printer (called an embosser). How
clear is the Braille it produces? How well does it stand up to
the demands made on it? And, of course, how rapidly does it print
a document? As the price of embossers drops, more and more
individuals and organizations can consider actually investing in
such equipment. Deciding which one to buy gets harder as the
choices multiply. Every manufacturer should be prepared to
provide samples of the Braille produced by its embossers, and
with a little investigation one can learn about the reliability
of the equipment other people have purchased and the dealers in
the area. These are subjective measures, but they are important
matters to consider.
     The question of the speed of Braille production is a
different matter. For years manufacturers have been reporting
embosser speed in characters produced per second. How accurate
are their claims? How does each embosser stack up against its
competition? Potential buyers need objective information in order
to make sound decisions.
     Enter the International Braille and Technology Center for
the Blind (IBTC). While David Andrews was still director of the
IBTC (he is now with the Minnesota Agency for the Blind), he
conducted tests to determine the accuracy of manufacturer ratings
of the speed of their embossers. The following article is the
result. Here it is:

     People who are buying a Braille embosser (printer) and who
consult the International Braille and Technology Center for the
Blind always ask us two questions: how much does it cost, and how
fast does it print--not necessarily in that order. Until now, the
speed figures we have given them are those published by the
manufacturers or distributors themselves. While we thought that
some of those figures might be exaggerated slightly, we did not   h)         0*0*0*  Ԍknow by how much, so we decided to find out.

                            The Test

     We began by devising a test--that is, producing a Braille
file that could be sent to all the printers. Our goal was to
devise a test that, while hard, was also a real-life document.
     We chose the July, 1995, issue of the Braille Monitor. We
used a leading Braille translation program, the Duxbury Braille
Translator Version 10.1, to produce a Grade II translated and
formatted file. Our test file contained straight text and other
styles that occurred naturally or that we added. These included
regular and outdented paragraphs, partially and fully blank
pages, centered and right-justified lines, two columns of words,
print and Braille page numbering, lines of dots and table-of-
contents guide dots, and more. We chose to include a range of
formatting and layout variations because different embossers
handle these complications in different ways, some more
efficiently than others.
     We also decided to place twenty-five 40-character lines of
text on each page. While some printers will print up to 27 lines
and/or 44 characters a line on an 11-by-11.5-inch page, we chose
to use less ambitious requirements. First, the 25-line-by-40-cell
page is commonly used, so it represents a real-life test. Second,
while some printers can emboss more than 40 cells on a line, not
all can. But with the exception of the Braille Blazer, which can
print only 34 cells a line, all units can produce 40-cell lines.
Consequently, by using a 40-cell line, we had a basis of
comparison. We formatted the Blazer test for 32 cells, which
allowed room for binding, as was done in the other tests. We also
had to format the Ohtsuki BT-5000 file for 19 lines since it will
not print 25 on a page. This machine embosses both print and
Braille on the same page, which takes more room.
     Finally, using the 25-line by 40-cell page, the file we
generated using Duxbury was 187 pages in length--long enough to
even things out. Of course the Blazer and Ohtsuki files contained
more pages, but the same amount of data.
     The appropriate test file was then sent to the embosser. The
"PRINT" command was used from the MS-DOS command line, and the
already-translated file was sent. This combination allowed a more
accurate measurement of embossing time than other methods that
could have been used. The timer was started at the same instant
that the "Return" ("Enter") key was pressed on the computer. One
of the Braille 'n Speak family of products from Blazie
Engineering--Braille 'n Speak, Type 'n Speak, or Braille Lite--
was used to time all tests. The timer was stopped as the embosser
ceased printing. The resulting time was converted into seconds
and divided by the size of the file, yielding a characters-per-
second rating.
     What follows is an alphabetical list of all the tested
embossers and what they did. The "CPS" acronym stands for
"characters per second" and "IBTC" stands for the International
Braille and Technology Center for the Blind."
   h)         0*0*0*  ԌBraille Blazer: manufacturer-rated speed, 15 cps; IBTC-rated
     speed, 8.0 cps; percent of variance, 47 percent; price,
     $1,695; price per cps, $211.88
Braille BookMaker: manufacturer-rated speed, 80 cps; IBTC-rated
     speed, 59.9 cps; percent of variance, 25 percent; price,
     $7,995; price per cps, $133.47
Braillo 200: manufacturer-rated speed, 200 cps; IBTC-rated speed,
     193.1 cps; percent of variance, 3 percent; price, $37,995;
     price per cps, $196.76
Braillo 400S: manufacturer-rated speed, 400 cps; IBTC-rated
     speed, 274.4 cps; percent of variance, 31 percent; price,
     $77,995; price per cps, $284.24
Braillo Comet: manufacturer-rated speed, 75 cps; IBTC-rated
     speed, 54.6 cps; percent of variance, 27 percent; price,
     $3,795; price per cps, $69.51
Braille Express: manufacturer-rated speed, 150 cps; IBTC-rated
     speed, 109.0 cps; percent of variance, 27 percent; price,
     $14,995; price per cps, $137.57
Elekul-03: manufacturer-rated speed, 300 cps; IBTC-rated speed,
     263.2 cps; percent of variance, 12 percent; price, $42,500
     approximately; price per cps, $161.47
Express 100: manufacturer-rated speed, 100 cps; IBTC-rated speed,
     78.6 cps; percent of variance, 21 percent; price, $10,500;
     price per cps, $133.59
Everest (Telesensory): manufacturer-rated speed, 100/60 cps;
     IBTC-rated speed, 51.7 cps; percent of variance, 48
     percent/14 percent; price, $5,995; price per cps, $115.96
Everest-D: manufacturer-rated speed, 79 cps; IBTC-rated speed,
     68.0 cps; percent of variance, 14 percent; price, $3,695;
     price per cps, $54.34
Index Basic: manufacturer-rated speed, 50/40 cps; IBTC-rated
     speed, 28.5 cps; percent of variance, 43 percent/29 percent;
     price, $2,795; price per cps, $98.07
Juliet: manufacturer-rated speed, 40 to 56 CPS (one line per
     second); IBTC-rated speed, 33.5 cps; percent of variance, 16
     percent; price, $3,995; price per cps, $119.25
Marathon: manufacturer-rated speed, 200 cps; IBTC-rated speed,
     167.0 cps; percent of variance, 17 percent; price, $12,995;
     price per cps, $77.81
Ohtsuki BT5000: manufacturer-rated speed, 8 cps (print and
     Braille); IBTC-rated speed, 6.4 cps; percent of variance, 20
     percent; price, $4,695; price per cps, $733.59
Porta-Thiel: manufacturer-rated speed, 10 cps; IBTC-rated speed,
     7.8 cps; percent of variance, 22 percent; price, $1,895;
     price per cps, $242.95
Resus RS-214: manufacturer-rated speed, 140 cps; IBTC-rated
     speed, 103.3 cps; percent of variance, 26 percent; price,
     $15,995; price per CPS, $154.84
Romeo RB-20: manufacturer-rated speed, 20 cps; IBTC-rated speed,
     17.2 cps; percent of variance, 14 percent; price, $2,495;
     price per cps, $145.05
Romeo RB-25: manufacturer-rated speed, 25 cps; IBTC-rated speed,
     23.8 cps; percent of variance, 5 percent; price, $2,295;
     price per cps, $96.43   h)         0*0*0*  ԌRomeo RB-40: manufacturer-rated speed, 40 cps; IBTC-rated speed,
     28.1 cps; percent of variance, 30 percent; price, $3,495;
     price per cps, $124.38
Ted 600: manufacturer-rated speed, 350 CPS approximately; IBTC-
     rated speed, 295.2 cps; percent of variance, 16 percent;
     price, $37,500; price per cps, $127.03
Thiel BAX-10: manufacturer-rated speed, 300 cps; IBTC-rated
     speed, 183.0 cps; percent of variance, 39 percent; price,
     $66,000; price per cps, $360.66
Thiel Beta X/3: manufacturer-rated speed, 130 cps; IBTC-rated
     speed, 116 cps; percent of variance, 10 percent; price,
     $13,995; price per cps, $120.64
Thomas: manufacturer-rated speed, 40 cps; IBTC-rated speed, 37.0
     cps; percent of variance, 8 percent; price, $2,995; price
     per cps, $80.95
VersaPoint: manufacturer-rated speed, 40 cps; IBTC-rated speed,
     27.4 cps; percent of variance, 32 percent; price, $3,795;
     price per cps, $138.50.

     As you can see from these data, the results we obtained from
some of the printers varied greatly from their manufacturers'
published figures. There are a number of reasons for this
variation, some of which are understandable and acceptable, and
some of which are not. First, remember that manufacturers
naturally rate their machines in the way that shows them in the
best light. While this is understandable and mostly acceptable,
we think that some of them have gone too far and should revise
their figures. One long-time marketing executive in the printer
field said to us, "Lots of mud is going to be slung here, and
some of it will stick to us all, but we [the embosser producers]
will be better off for it in the long run."
     A number of embossers, such as the BookMaker and the Braille
Express from Enabling Technologies, take approximately the same
time to print a line, whether there are 30 characters on it or
44. Since these printers will print up to 44 characters on a
line, the manufacturer naturally computes the CPS rating using
the longer line, which raises the rating. Thus our results, which
are based on a 40-character line length, should legitimately be
increased by at least 10 percent, making our figures very close
to Enabling's.
     Other manufacturers may measure only a single Braille page
of text because the page change takes time. We suspect that this
is in part how Index arrived at its initially unrealistically
high rating of the original Everest. Still others use a burst
rate--timing the printing of just one line. This is what Blazie
Engineering did with the Blazer, according to Deane Blazie.

                   Putting it into Perspective

     What follow are our observations concerning individual
embossers as well as information we believe necessary to
interpret the individual results. Please remember that printing
speed is only one aspect of choosing a Braille embosser. Others
include price; quality of Braille produced; reputation of   h)         0*0*0*  Ԍmanufacturer; reliability; reputation of local dealer; past
experience; availability of timely service and support; ease of
operation; clear and understandable documentation; manuals
available in well-formatted, Grade II Braille; and more.
     The manufacturer-provided speed figure for the Braille
Blazer is a bit optimistic and should be reduced in our opinion.
While a good deal of variance from our test can be expected, a 47
percent difference seems too much to us. Prior to testing, our
hypothesis was that a 10 percent variance would be acceptable.
After conducting the tests, we now conclude that a 20 percent
variation is probably acceptable, although less is better. As
already pointed out, at least a 10 percent variance can
legitimately be accounted for in some cases.
     The Braille BookMaker from Enabling Technologies is one of
the units that was rated by its manufacturer using a 44-cell
line, so its 80 CPS rating is not unrealistic. We were surprised
by how close the Braillo 200 came to its rated speed and how far
off the Braillo 400S was from its rated speed. With a 31 percent
variance, a re-rating by Braillo Norway might be in order. A 350
CPS rating would make more sense. The Braille Comet missed its
mark because it is slow in moving paper when going to a new page
from a partially printed page. The Braille Express and the
Express 100 were both rated by Enabling Technologies using a
44-cell line, so their figures are not as different from ours as
would first appear. The Elekul-03 came fairly close to its
manufacturer's rating; in fact its speed is directly related to
the voltage of the electrical supply. Since ours is on the low
side of average, the speed could be increased by raising the
voltage.
     The Everest-D (originally sold exclusively by Telesensory)
has two figures listed: 100 and 60 CPS. When it was introduced in
1992, Index and Telesensory both widely touted its 100 CPS
figure. As you can see from our results (51.7 CPS and a 48
percent variance) its actual performance doesn't even come close.
In June of 1995 Index published a newsletter in which it provided
speed figures for all their printers. Interestingly enough, all
figures were lower--considerably so in some cases--than those
previously published. Even figures for embossers the company no
longer manufactures were adjusted downward. Further, Index has
developed its own speed test and has been circulating it to other
printer manufacturers in an attempt to get all of them to use the
test. At least one company (Enabling Technologies) has resisted,
feeling, they say, that the testing should be done by an
independent entity like the International Braille and Technology
Center for the Blind. They also believe that the test was devised
so as to put Index products in the best light.
     We have looked at the Index test and have even run it on a
few machines. It yields faster times than the IBTC-developed
test. While it is not a bad one, we concluded that it wasn't very
realistic. It was almost entirely straight text with most lines
filled. It also included no form feeds, leaving page breaks to
the embosser which, with most machines, is usually faster.
However, most Braille translation programs insert form feeds so
that they can control page size, accurately supply page numbers,   h)         0*0*0*  Ԍand easily change page sizes.
     One Braille translator author described Index's revised
figures as "preemptive revisionism." Index knew we were working
on printer speed ratings and has very recently publicly revised
its figures to be more realistic. While these are now quite
accurate and more realistic than those of many and though the
efforts of Index to get a handle on measured printer speeds are
laudable, the fact still remains that in the past, and for a long
time, Index used some very inflated figures. We can only hope
that everyone will start using more realistic figures now that
attention has been drawn to this issue.
     Other Index embossers tested include the new Everest-D and
the Index Basic. The new Everest came relatively close to its
revised rating of 79 CPS. The Index Basic missed its rating by an
unacceptable level. Index has marketed the embosser as a 50-CPS
machine for years, touting it as faster than the competition, the
Romeo RB-40 and the VersaPoint. At 28.5 CPS, it is little faster
than the others, particularly the Romeo, which we rated at 28.1
CPS. Its revised rating of 40 CPS is more realistic than before,
but its 29 percent variance is still too high.
     According to Enabling Technologies, the Juliet prints one
line per second. Since it can print up to a 56-character line,
its potential speed can be as high as 56 CPS. We found that the
machine comes acceptably close to its published rating.
     Also the Marathon, Ohtsuki, and Porta-Thiel come reasonably
close to their published specifications. As an interesting aside,
the Ohtsuki took over 7 hours to run the printing test. The
quickest machine ran it in just under 9.5 minutes.
     The Resus RS-214, which is no longer available, missed its
specifications by a little more than we would have preferred. It
seems a bit slow in moving paper through blank lines. The Romeo
RB-40 is not rated accurately. The machine misses its mark by 30
percent. A longer line might have helped it a little, but not
much. The RB-20 and the RB-25, on the other hand, came quite
close to their published specifications. The RB-25 in particular
represents a good bargain. For $1,100 less than the RB-40, you
sacrifice only 4.3 CPS, and the Braille quality is quite good.
The RB-20 is primarily marketed to sign makers now. Some
components have been strengthened to allow it to emboss on light
metal and plastic materials reliably.
     The Ted 600 was originally rated at 450 CPS. The
manufacturer later slowed down the machine to make it more
reliable. They estimate the slowed-down version at 350 CPS, and
at 295.2 CPS we consider our rating to be in line with theirs.
     The Thiel BAX-10 was a disappointment, while the Beta X/3
was quite near its published figure. Considering its 183 CPS and
a 39 percent variance figure, the BAX-10 should be re-rated.
     The Thomas is another one of Enabling's new printers that
comes quite close to its published figure, produces good Braille,
and is a good bargain.
     Finally, the VersaPoint missed its mark by a good deal. It
is slower than the Romeo RB-40, which is its main competition. At
27.4 CPS and a 32 percent variance, it should be reclassified by
Telesensory.   h)         0*0*0*  Ԍ                           Conclusions

     While we have devoted a good deal of time and energy to
rating Braille embossers, we reiterate that speed is only one
consideration in making a printer decision. The actual numbers
themselves are probably not as important as their comparisons to
each other. It is more important to know that the Romeo RB-40 is
slightly faster than the VersaPoint or that the Braillo 200 is
faster than the Thiel Bax-10 (which is considerably more
expensive) than it is to remember the exact numbers. Also, as
noted earlier, there are other factors which should be
considered.
     In addition to the speed ratings, we have provided a dollar-
per-CPS figure. While it is hard to resist this measure, it does
involve a couple of dangers. First, there is a tendency
(especially by government types who are writing specifications
for equipment bids) to use measures like this as their sole means
of decision-making. These measures do not take into account
Braille quality and other important factors. Second, when
designing a printer, the manufacturer has to make certain
tradeoffs and compromises. It may be possible to gain some
apparent speed by using a shorter line--at the cost of
flexibility. And an embosser employing a long line and rated by a
conservative but arduous test such as that developed by the
International Braille and Technology Center for the Blind may
look more unattractive than it actually is. In short, a variety
of factors should be considered and balanced when making the
decision to buy an embosser.
     The prices of Braille printers have dropped over the past
few years. There are also many more models available now than
formerly. However, past a certain point, the only way to achieve
more speed or lower price is to sacrifice the quality of the
Braille produced. While we applaud consumer choice and lower
prices, we do not want to see the quality of the Braille produced
by computer-driven embossers reduced dramatically.
     We urge the Braille printer industry to take a new look at
measuring embosser speed. While we may not have designed the
definitive speed test, we do believe that we have designed a
fair, real-life test that can be used to draw valid comparisons.
We also urge consumers, agencies, and government entities not to
look only at our speed figures when making purchasing decisions.
They are a useful comparison and guide, but they should be only
one of many considerations.
     The data presented here, while interesting, do not represent
complete reviews of the Braille printers. However, the NFB's
International Braille and Technology Center for the Blind has
been working as a sub-contractor on a Braille Literacy Training
grant that was awarded to the American Printing House for the
Blind (APH) by the National Institute on Disability and
Rehabilitation Research (NIDRR). Among other activities in
connection with the NIDRR grant, the International Braille and
Technology Center for the Blind has now reviewed all of its
Braille-related technology. The resulting document is being
published in ink-print and Braille by APH and electronically by   h)         0*0*0*  Ԍthe National Federation of the Blind. By the time you read this
article, the entire report should be available.
     For more information about Braille embossers, translation
software, or other technology matters, call the International
Braille and Technology Center for the Blind. You can reach us
afternoons at (410) 659-9314. You can also reach us via NFB NET,
the NFB's computer bulletin board service, by calling (612)
696-1975.

