[Blindmath] grayscale braille

Wed Mar 21 12:20:18 UTC 2012

Thanks for that, it just sparked of a logical way of assigning in my mind.

However before discussing the system, may be quick thoughts on colours and 
how someone who cannot see relates to them. I have seen a little in the past 
so do understand what colours are and can imagine what a colour might be 
like if someone can start from a point I know and then describe how it 
varies (eg. darker, a bit greener, etc). However if I don't have a decent 
start point close to where we need to get I can actually struggle a bit to 
understand what it might look like (eg. I am getting an artificial eye, I 
get to choose my eye colour how great, however when discussing it I was 
struggling at times to know what colour precisely we were discussing and 
whether it suits me as I have never seen well enough to pick out eye 
colours). So I guess my thought is simply, colour in tactile diagrams will 
only be useful for area identification rather than to convey how something 
visually appears.

Back to how to show colour in a tactile form. We will stick with your 27 
levels to start with. Each row of the Braille cell (6 dot cell) could 
represent each colour, and the number of dots on that row tells you the 
level for the colour. Lets just assign the rows logically, RGB, top, middle 
bottom rows respectively.

Now the above also fits with your previous suggestion of filling up the cell 
means more intense and the overall view of the cell relates to that.

However may be at the cost of loosing that relationship of more dots more 
intense, we could actually extend up to 64 colours by using the position of 
just one dot on a row (left or right side of the cell) so allowing each 
colour to have four intensities.

Whether 27 or 64 is needed I don't know, in either case its still an 
improvement over what sighted people had with computers back when I started 
with a BBC micro (8 or 16 depending on what computer system you went with).

Out of interest, having discussed how colours are represented, what was 256 
colours about on some of the early windows computers, 256 is not anything 
raised to the three.

Michael Whapples

-----Original Message----- 
From: Richard Baldwin
Sent: Wednesday, March 21, 2012 12:57 AM
To: Blind Math list for those interested in mathematics
Subject: Re: [Blindmath] grayscale braille

I'm going to add a few more comments to those that I made earlier. As you
may have gathered, I have had a long career in computer technology. That
career included a large dose of digital signal processing. I will get to
the reason for that comment later.

Once we learn of a new technology, it is interesting to speculate on how
that technology might be put to use. John indicated earlier that his ASCII
Gray Scale technology will provide 26 levels of gray scale. Since I know
little to nothing about Braille, I will take him at his word on that.

My digital signal processing activities included working for several years
in the area of undersea sonar technology. At that time, the general
consensus was that having only a gray scale display, an experienced sonar
operator could only make use of about seven levels of gray. Thus, while 26
levels of gray might be useful for some applications, it might be overkill
for others.

It was also the general consensus that an experienced sonar operator using
a color display could make use of far more than 7 different colors. This
caused me to wonder if it might make sense to apply those 26 levels to
color as an alternative to gray scale. I wonder if 26 levels of color
information might be more useful than 26 levels of gray scale information.

I am going to assume that such a possibility might be worth looking into
and for reasons that I will explain later, I am going to assume that by
including one non-alphabetic character, it would be possible to expand to
27 levels instead of 26.

A typical full-color bitmap image contains more than 16 million colors.
Therefore, converting those colors to either 26 gray scale values or 27
color values will necessarily cause a large amount of information to be
lost.

However, in many cases, converting to 26 gray scale levels will cause much
more information to be lost than converting to 26 or 27 colors.

Converting to gray scale means throwing away the color information and
preserving only intensity information without making any attempt to make
use of the color information.

Reducing the number of colors down to 27 preserves both color information
and intensity information and may produce a more useful result.

Having decided that it might be useful to re-quantize an image into 27
unique colors leaves open the question of how best to do that. There are a
multitude of possibilities in this regard and some may prove more useful
than others. I will describe one such approach below.

First some background information. Color in a modern computer is typically
maintained as a weighted sum of three primary colors: red, green, and blue.

Displaying those three primary colors along with three secondary colors
produced by adding the three primary colors together in pairs produces the
six colors of a typical rainbow: red, yellow, green, cyan, blue, and
magenta.

While this RGB approach to dealing with color is particularly efficient for
computational purposes, it does not describe how humans think about color.
For example, it is not likely that a human would go to a furniture store
and ask to see a sofa with upholstery that is X-percent red, Y-percent
green, and Z-percent blue. Instead, in Austin, Texas, where burnt orange is
the color of the day, (at least for UT students) a UT student might ask for
a color that is a subdued version of orange that is a little on the dark
side and closer to red than green, or words to that effect.

Another system for dealing with color that is more in keeping with how
humans think about color is often called HSV or HSB. This stands for Hue,
Saturation, and either Value or Brightness depending on the use of a V or a
B in the abbreviation.

Fortunately, it is relatively easy to write a computer program that will
transform a set of unique RGB color values to corresponding unique HSB
values and back again.

In the RGB system, the values of red, green, and blue for any individual
pixel can range from 0 to 255. Thus, the number of possible combinations is
equal to 255 raised to the third power, which is where the "more than 16
million" comes from. In fact, a typical modern computer system using 24-bit
RGB color can mathematically describe 16,581,375 different colors.

Similarly, these 16 million plus RGB colors can be transformed into a
corresponding set of 16 million unique colors in the HSB nomenclature. In
the HSB nomenclature, the values for each of the three parameters ranges
from 0 to 1.0.

One approach to converting from more than 16 million unique colors to 27
unique colors would be to:

1. Transform the RGB color for each pixel to the corresponding HSB color.
2. Re-quantize each of the three HSB parameter values into three unique
values: 0.333, 0.666, and 1.0. This would produce 3-cubed or 27 unique
colors.
3. Transform the modified HSB values back to the corresponding RGB values
and replace the original pixel color with the new color.

Given the 27 possible color values, a table lookup procedure could be used
to generate an ASCII value corresponding to each color value and that ASCII
value could be used to produce the Braille characters.

Food for thought.

Dick Baldwin

On Tue, Mar 20, 2012 at 1:42 PM, Richard Baldwin 
<baldwin at dickbaldwin.com>wrote:

> Hi Michael,
>
> You wrote:
>
> "may not really give the same impression as the visual things you
> describe as one never reads/views Braille at a distance so won't get that
> point of not really resolving individual dots but rather a general effect.
> "
>
> In this respect, I was thinking in terms of embossed images as opposed to
> single line Braille displays. It seems to me that brushing your hand 
> across
> an embossed image and being aware of the dot density at different 
> locations
> on the image might be somewhat analogous to viewing a printer-art image
> from a distance.
>
> Dick B.
>
> On Tue, Mar 20, 2012 at 1:26 PM, Michael Whapples 
> <mwhapples at aim.com>wrote:
>
>> Wow, what a lot of information.
>>
>> I hadn't heard of the making a grayscale image with people before, I knew
>> people have used lots of people to create images by having people wearing
>> different colours though.
>>
>> Back to the actual subject, I thought you might have been interested as I
>> know with some of your image processing you have commented on how low the
>> grayscale resolution (shades of gray resolution) is with existing
>> technology. However the increase in gray levels comes at a very high cost
>> of spatial resolution, is 40 pixels wide, may be 80 for very rich/lucky
>> people with such a Braille display, really good enough? I have my doubts
>> even for those with 80 cell displays, the display is so long will one
>> really get a proper awareness of what relates to what because of the 
>> spread?
>>
>> Another question is, why only use letters? Surely one could potentially
>> use up to 64 for a 6-dot cell and on a Braille display one could get full
>> 256 level representation!
>>
>> I like your thought of number of dots for level as it would make it
>> intuitive, but does then bring down the number of levels and may not 
>> really
>> give the same impression as the visual things you describe as one never
>> reads/views Braille at a distance so won't get that point of not really
>> resolving individual dots but rather a general effect.
>>
>> To try to keep to some sort of logical assignment, if trying for the 256
>> levels, then I would just use binary around the Braille cell (eg. 
>> lightest
>> being no dots, next coming dot-1, next being dot-2, next being 
>> dots-12,...
>> very nearly black dots-2345678, darkest dots-12345678).
>>
>> Michael Whapples
>>
>> -----Original Message----- From: Richard Baldwin
>> Sent: Tuesday, March 20, 2012 4:37 PM
>>
>> To: Blind Math list for those interested in mathematics
>> Subject: Re: [Blindmath] grayscale braille
>>
>> I have been following this conversation with interest. For the record, I
>> am
>> not blind and know very little about Braille. However, I do know quite a
>> lot about image processing.
>>
>> In the sighted world, a character printer can definitely be used to
>> produce
>> gray scale images, but not in the way that is described here.
>>
>> In the 1960s, a typical data processing printer weighed several hundred
>> pounds, stood chest high from the floor, printed upper-case letters,
>> numbers, and a set of special characters at 10 characters per inch with a
>> line length of 132 characters on 14-inch wide fan-fold paper. A typical
>> data processing printer could print 600 lines per minute or more. With
>> some
>> printers, the paper came out so fast that special mechanisms were 
>> required
>> to prevent it from flying across the room and to refold itself in the
>> output bin.
>>
>> Many data centers had various examples of printer art posted on the walls
>> with the most common being a reasonably good gray scale replica of the
>> Mona
>> Lisa.
>>
>> However, unlike the scheme that is described here, there was no intent 
>> for
>> the viewer to assign special meaning to any individual character. In 
>> fact,
>> the intent was for the characters to visually run together is such a way
>> that they would not be perceived as characters at all. Instead, the big
>> picture view of the printout would give the impression of a gray scale
>> image with individual characters fading into the background.
>>
>> Someone came up with a sequence of characters based on the amount of ink
>> deposited within the 0.1-inch wide cell by each character. That was a 
>> long
>> time ago and I don't recall the specific sequence of characters that was
>> used. I am guessing that the period character was used to convey light
>> gray. Moving from there through the sequence, each character deposited
>> more
>> ink and therefore produced a darker cell. I'm also guessing that the
>> sequence probably consisted of eight to ten different characters making 
>> it
>> possible to produce the illusion of eight to ten levels of gray.
>>
>> Characters were chosen such that when a person stepped away from the
>> printout and viewed it as a whole, that person didn't see individual
>> characters. Instead, the result was an illusion of a large gray scale
>> image.
>>
>> I used this scheme myself in the days before the invention of the CalComp
>> incremental plotter to produce images of contour maps.
>>
>> Perhaps a similar scheme could be used with braille with each cell
>> containing from zero to six dots (or perhaps eight dots). This might make
>> it possible for a blind person to perceive white plus six (or eight)
>> levels
>> of gray without the requirement to mentally associate specific characters
>> with specific shades of gray.
>>
>> For the six-dot case, the following sequence of characters might provide
>> the illusion of increasing darkness (but a different selection might
>> produce better tactile results):
>>
>> hex 41, A, 1 dot
>> hex 42, B, 2 dots
>> hex 44, D, 3 dots
>> hex 47, G, 4 dots
>> hex 51, Q, 5 dots
>> hex 3D, =, 6 dots
>>
>> Dots have long been used to produce the illusion of gray scale images.
>> When
>> I was a youngster, pictures in most small-town newspapers were presented
>> in
>> gray scale because printing presses that could print in color were very
>> expensive. If you looked closely at a newspaper photo, you could see that
>> the picture was simply an array of dots. I seem to recall that the gray
>> scale effect was achieved by producing an array of black dots on a 
>> uniform
>> grid using different sized dots.
>>
>> At one point in time, I had some very interesting photographs from "Life"
>> magazine involving very unique gray scale images. In those photos, a
>> photographer produced images of various things, including a portrait of
>> Woodrow Wilson and a picture of the U.S. Marine insignia by taking
>> photographs of thousands of troops in formation wearing white shirts and
>> black shirts. In effect, each person was one dot in the image.
>>
>> Apparently the photographer would place the camera on top of a building 
>> or
>> tower and take of picture of the troops in formation down below. He even
>> took perspective into account. For example, moving away from the camera,
>> each row of troops was wider than the one before it. In some cases, the
>> row
>> of troops closest to the camera contained 20 or 30 troops while the most
>> distant row would contain 200 to 300 troops. Thus, the "dot density"
>> increased as you viewed the image going from bottom to top.
>>
>> I was able to find an image of a printer generated Mona Lisa on the web,
>> but was unable to find any images of the human-dot photographs.
>>
>> Dick Baldwin
>>
>>
>> On Mon, Mar 19, 2012 at 6:13 PM, Michael Whapples <mwhapples at aim.com>
>> wrote:
>>
>>  I don't know if John is on the blindmath list, I'll forward the message
>>> on
>>> in case he isn't, however it would be better if you could somehow
>>> communicate direct with him (eg. by posting to the NFB-science list if 
>>> he
>>> isn't on blindmath).
>>>
>>> Michael Whapples
>>>
>>> -----Original Message----- From: Pranav Lal
>>> Sent: Monday, March 19, 2012 10:58 PM
>>> To: 'Blind Math list for those interested in mathematics'
>>> Subject: Re: [Blindmath] grayscale braille
>>>
>>>
>>> Hi John,
>>>
>>> Can I use this plotting technique for any image? Your example seems to
>>> work
>>> only for functions.
>>>
>>> Pranav
>>>
>>>
>>> ______________________________****_________________
>>> Blindmath mailing list
>>> Blindmath at nfbnet.org
>>> http://nfbnet.org/mailman/****listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org>
>>> <**http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org>
>>> >
>>>
>>> To unsubscribe, change your list options or get your account info for
>>> Blindmath:
>>> http://nfbnet.org/mailman/****options/blindmath_nfbnet.org/****<http://nfbnet.org/mailman/**options/blindmath_nfbnet.org/**>
>>> mwhapples%40aim.com<http://**nfbnet.org/mailman/options/**
>>> blindmath_nfbnet.org/**mwhapples%40aim.com<http://nfbnet.org/mailman/options/blindmath_nfbnet.org/mwhapples%40aim.com>
>>> >
>>>
>>> ______________________________****_________________
>>> Blindmath mailing list
>>> Blindmath at nfbnet.org
>>> http://nfbnet.org/mailman/****listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org>
>>> <**http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org>
>>> >
>>>
>>> To unsubscribe, change your list options or get your account info for
>>> Blindmath:
>>> http://nfbnet.org/mailman/****options/blindmath_nfbnet.org/****<http://nfbnet.org/mailman/**options/blindmath_nfbnet.org/**>
>>> baldwin%40dickbaldwin.com<http**://nfbnet.org/mailman/options/**
>>> blindmath_nfbnet.org/baldwin%**40dickbaldwin.com<http://nfbnet.org/mailman/options/blindmath_nfbnet.org/baldwin%40dickbaldwin.com>
>>> >
>>>
>>>
>>
>>
>> --
>> Richard G. Baldwin (Dick Baldwin)
>> Home of Baldwin's on-line Java Tutorials
>> http://www.DickBaldwin.com
>>
>> Professor of Computer Information Technology
>> Austin Community College
>> (512) 223-4758
>> mailto:Baldwin at DickBaldwin.com
>> http://www.austincc.edu/**baldwin/ <http://www.austincc.edu/baldwin/>
>> ______________________________**_________________
>> Blindmath mailing list
>> Blindmath at nfbnet.org
>> http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org>
>> To unsubscribe, change your list options or get your account info for
>> Blindmath:
>> http://nfbnet.org/mailman/**options/blindmath_nfbnet.org/**
>> mwhapples%40aim.com<http://nfbnet.org/mailman/options/blindmath_nfbnet.org/mwhapples%40aim.com>
>>
>> ______________________________**_________________
>> Blindmath mailing list
>> Blindmath at nfbnet.org
>> http://nfbnet.org/mailman/**listinfo/blindmath_nfbnet.org<http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org>
>> To unsubscribe, change your list options or get your account info for
>> Blindmath:
>> http://nfbnet.org/mailman/**options/blindmath_nfbnet.org/**
>> baldwin%40dickbaldwin.com<http://nfbnet.org/mailman/options/blindmath_nfbnet.org/baldwin%40dickbaldwin.com>
>>
>
>
>
> --
> Richard G. Baldwin (Dick Baldwin)
> Home of Baldwin's on-line Java Tutorials
> http://www.DickBaldwin.com
>
> Professor of Computer Information Technology
> Austin Community College
> (512) 223-4758
> mailto:Baldwin at DickBaldwin.com
> http://www.austincc.edu/baldwin/
>

-- 
Richard G. Baldwin (Dick Baldwin)
Home of Baldwin's on-line Java Tutorials
http://www.DickBaldwin.com

Professor of Computer Information Technology
Austin Community College
(512) 223-4758
mailto:Baldwin at DickBaldwin.com
http://www.austincc.edu/baldwin/
_______________________________________________
Blindmath mailing list
Blindmath at nfbnet.org
http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
To unsubscribe, change your list options or get your account info for 
Blindmath:
http://nfbnet.org/mailman/options/blindmath_nfbnet.org/mwhapples%40aim.com 