[Tactile-Talk] Tactile Talk Grayscale braille versus position

[Seeing Hands]

I have done similar things to deal with intensity, but I used a slightly 
different encoding mechanism which can be used to get higher precision 
in intensity levels and had a couple other advantages, so it might be of 
interest here.

My approach was to encode the intensity value to a binary number, then 
represent that binary number as a cell. For example, a six-dot cell can 
have 64 different combinations, so if they are all used, you can see 64 
different levels of intensity. In my testing, I found that the order of 
the dots made a major difference in how easy it was to detect the 
differences. I used the top row, dots 1 and 4, for the first two levels 
of intensity. That made it easy to scan the image at different levels of 
precision. When looking for the bright areas versus the dark areas of 
the image, you only have to check each cell for the presence of dot 1. 
Once you've found an area of interest, you can do the same by looking 
for dot 4, drilling down until you have the fine detail with dot 6 
representing the smallest change in intensity. The reason I chose that 
order is that I often identify the presence of dot 1 and 4 quickly when 
scanning, which means that if it changes, I notice quickly. The default, 
numeric ordering of the dots would be just as capable of this, but I 
found it more difficult to read quickly.

There are two other advantages of this approach. If you needed all 256 
levels, you could do the same thing with an eight-dot cell and encode 
all of them. The main downside of all these approaches is that you can't 
represent a large intensity graph or heat map without a massive sheet of 
paper, but the benefit of binary encoding is that you can compress your 
graph at the cost of precision much more flexibly. You could use only 
one vertical side of a cell per point, get eight levels of precision, 
and represent two times as much data horizontally. For embossers where 
you don't have to use standard-sized cells, you could use two-by-two 
cells, get 16 levels of precision, and fit a square map onto a square 
sheet of Braille paper. That approach could also work on the Monarch or 
possibly the Graphiti.


--

Seeing Hands
Phone: +1 612-219-6271

On 1/8/2025 13:48, Ken Perry via Tactile-Talk wrote:
>
> Have you done much studies on the amount of information that can be 
> done with grey scale.What I am meaning is how many patterns actually 
> work depending on distances between dots.
>
> *From:*Tactile-Talk <tactile-talk-bounces at nfbnet.org> *On Behalf Of 
> *John Miller via Tactile-Talk
> *Sent:* Wednesday, January 8, 2025 2:35 PM
> *To:* David Engebretson via Tactile-Talk <tactile-talk at nfbnet.org>
> *Cc:* John Miller <johnmillerphd at hotmail.com>
> *Subject:* [Tactile-Talk] Tactile Talk Grayscale braille versus position
>
> Hello,
>
> I have a new idea to share with you.
>
> I thought this list was going to be used primarily for how to use a 
> Graphiti or a Monarch or a multi-line braille display to look at 
> tactile position information.
>
> It is true that looking at the output of a 3D printer or looking at a 
> tactile image generated by a traditional braille embosser or reviewing 
> a text file sent to a traditional braille embosser all have merit for 
> showing tactile information versus position.
>
> I use the method described below at my work to look at intensity 
> versus position images.
>
> Sighted engineers look at intensity vs position images from time to 
> time and find them quite useful.
>
> Intensity vs position often shows up as a rainbow of colors from Red 
> to Blue, or also can be different levels of gray when reviewed visually.
>
> I wanted to share with you an example of grayscale braille.
>
> It sometimes is the case that a blind person wishes to review an 
> analytical 3-D function.
>
> Visually these functions are often displayed in grayscale with 
> different locations in a 2-D region having a gray level ranging from 
> black to white.
>
> For visual consumption the grayscale level will often range from 0 
> through 255.
>
> Using letters of the alphabet results in coarser intensity resolution 
> than 256 levels but allows for one braille symbol to be placed in one 
> braille cell position rather than having to use 3 braille cells for a 
> 3-digit number.
>
> With a refreshable braille display or a braille embosser it can be 
> useful to indicate different amplitudes of the image with different 
> letters of the alphabet.
>
> For example, the letters "a" through "t" can be used to indicate level 
> 0 through level 19. This leaves letters "u" through "z" for special 
> emphasis such as using "u" to mark the locations of a route through a 
> topology map showing various altitudes of the land.
>
> With 8-dot braille it is possible to use "a" through "z" followed by 
> "A" followed by "Z" to display 52 levels. Additional symbols such as 
> "1" through "0" can be extended to give 62 different levels of change 
> but in this example only the symbols "a" through "t" are assigned.
>
> I wish to look at the height of a sphere versus position for a sphere 
> that is centered at the origin.
>
> In this case the radius of the sphere is 1.5 and I examine the grid of 
> x ranging from -1.6 through 1.6 and y ranging from -1.6 to 1.6 both 
> with steps of 0.1.
>
> For each (x, y) pair I evaluate the expression z equals the square 
> root of the quantity r squared minus the sum of x squared plus y 
> squared end quantity.
>
> For z in the range of 0 less than or equal to z less than 0.1, assign "a"
>
> for z in the range of 0.1 less than or equal to z less than 0.2, 
> assign "b"
>
> ...
>
> for z in the range of 1.4 less than or equal to z less than 1.5 assign "o"
>
> for z in the range of 1.5 less than or equal to z less than 1.6 assign "p"
>
> I form a 33 row by 33 column grid of letters ranging from "a" to "o".
>
> The plot shows the top hemisphere of the sphere. Note that for the 
> equation of the sphere for each (x, y) there is both a positive z 
> value and a negative  z value with the same amplitude.
>
> By observation the edges of the plot all evaluate to the "a" symbol or 
> to zero.
>
> Each row starts at zero and ends at zero and may visit a nonzero value 
> in the middle.
>
> The same is true for each column.
>
> The letters in each row start with "a" and ascend with some jumps and 
> some duplicates to a high letter and then descend in similar fashion 
> back to "a".
>
> The image appears to have radial symmetry.
>
> The center of the image evaluates to the "o" symbol or to 1.4.
>
> There are about 6 rows with at most 6 columns in each showing the "o" 
> symbol.
>
> Surrounding the "o" region is the "n" region evaluating to 1.3.
>
> There are about 3 "n" symbols adjacent to each other in any row.
>
> Surrounding the "n" region is the "m" region evaluating to 1.2.
>
> You can trace a ring through the "n" region, "m" region, and other 
> regions.
>
> Returning your attention to the "n" region you can see that the band 
> is at most 3 symbols wide or that the band has a width of at most 0.2.
>
> The "g" region evaluating to 0.6 is only 1 symbol wide and on some 
> rows has to be inferred by following the boundary between a letter 
> earlier in the alphabet than "g" such as "f" followed by a letter 
> later in the alphabet than "g" such as "h".
>
> About half way down the image you can examine a row of letters that 
> includes the peak amplitude letter "o".
>
> At the middle of the row of letters the gradient from a "o" to a "j" 
> is very gradual taking about 11 symbols to go from one to the other.
>
> At the beginning and the end of the row it takes about 5 symbols to go 
> from "a" to "j".
>
> This means that an object with some decent traction placed near the 
> middle of this row on the sphere would stay put but if it was placed 
> on the left or the right edge of the sphere it would slide off.
>
> In other words, the left and right sections of this row have a large 
> change in altitude for a small change in horizontal position whereas 
> the middle of the row has a small change in altitude for a small 
> change in horizontal position.
>
> These are observations that are not immediately apparent by just 
> examining a comma delimited CSV file of the data.
>
> On the other hand, these observations can be made directly by using 
> the grayscale braille image.
>
> The grayscale image of a sphere follows:
>
> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
>
> aaaaaaaaaaaabcdddccaaaaaaaaaaaaaa
>
> aaaaaaaaadefgggggggffdbaaaaaaaaaa
>
> aaaaaaabcefghhhhhhhggfdcbaaaaaaaa
>
> aaaaabefghiijjjjjjjiihggedbaaaaaa
>
> aaaabefhhijjkkkkkkkjjiihgfbaaaaaa
>
> aaaaefhiijkkkllllllkkjjihgecaaaaa
>
> aaabfhijjkkllllllllllkjjihfdbaaaa
>
> aaacghijjklllmmmmmmllkkjjigecaaaa
>
> aadehijkkllmmmmmmmmmmlkkjihfeaaaa
>
> aaefijkkllmmnnnnnnnmmmllkjigfbaaa
>
> aafgijkllmmnnnnnnnnnnmllkkihgcaaa
>
> abghjkkllmnnnnnnnnnnnmmllkjhhdbaa
>
> acghjkllmmnnnnooonnnnmmmlkjhhecaa
>
> adghjkllmmnnnoooooonnnmmlkjihedaa
>
> adghjkllmmnnnoooooonnnmmlkjihedaa
>
> adghjkllmmnnnoooooonnnmmlkjihedaa
>
> acghjkllmmnnnnoooonnnmmmlkjihecaa
>
> acghjkllmmnnnnooonnnnmmmlkjhhecaa
>
> aafgijkllmmnnnnnnnnnnmmllkihgdaaa
>
> aafgijkllmmnnnnnnnnnmmllkjihgcaaa
>
> aadfhijkklmmmmnnnmmmmllkkjhgfbaaa
>
> aabdgijjkkllmmmmmmmmllkkjigedbaaa
>
> aaacghijjklllmmmmmmllkkjjigecaaaa
>
> aaabeghijjkkllllllllkkjjihecbaaaa
>
> aaaadfghiijkkkkkkkkkjjiihfdaaaaaa
>
> aaaabbefghiijjjjjjjiihggedbaaaaaa
>
> aaaaaacdefghhhiiiihhhgeecaaaaaaaa
>
> aaaaaaabcefghhhhhhhggfdcbaaaaaaaa
>
> aaaaaaaaaabcdeeeeeedcbaaaaaaaaaaa
>
> aaaaaaaaaaaabcdddccaaaaaaaaaaaaaa
>
> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
>
> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
>
> Very best,
>
> John
>
>
> _______________________________________________
> Tactile-Talk mailing list
> Tactile-Talk at nfbnet.org
> http://nfbnet.org/mailman/listinfo/tactile-talk_nfbnet.org
> To unsubscribe, change your list options or get your account info for Tactile-Talk:
> http://nfbnet.org/mailman/options/tactile-talk_nfbnet.org/community%40seeinghands.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nfbnet.org/pipermail/tactile-talk_nfbnet.org/attachments/20250108/180ac9eb/attachment.htm>