An
ASCII Image is simply an Ordinary Photographic Picture that has been reduced to
a Gray Scale of 16 Shades of Gray, Then Converted to A Set of Characters from a
Given Alphabet.
The
ASCII Conversation Program ( written
for this series in ‘Scratch’ ) then Samples The Picture at Various Points to
Determine which Gray Shade is at that location— Then it Selects from a List of
Characters, One which Ostensibly Conveys that Same ‘Sense’ of ‘Darkness’ or
‘Lightness’ —
e.g.; A very Light Swatch ( Pixel ) in The
Photograph will be rendered on The ASCII Image as a : . , ‘ or a -
A
Dark Swatch ( Pixel ) would be Translated into a : W M R or K
The
Internet has many such ASCII Conversation Programs Available for you to use;
But
i have found that many of them; Nearly All in fact, Are very lazy about
choosing which Characters are used for Which Shades of Gray. Even if you were
to find an ASCII Conversion Program that Uses more than 16 Shades of Grey; Say—
24, 36, 128 or 256; The Generic Approach is to Use One Single Character for
Each Shade of Gray; So that there is a Consistent 1::1 Correspondence with Grey
Swatch to Character Used.
The
Problem; As i Perceived it; Was that this approach created Regions in which a
Single Character was Used Exclusively, & Worse, Far Worse ( !!! ) — Was
that when a Given Gray Region would Buttress another Gray Region; The Boundary
of these Two Regions would create an UnSightly & UnNatural Looking Ridgeway
( ! )
When
there are Several of these Ridgeways Defining a Larger Swatch which in The
Photograph looks like a Smooth Gradation from Very Light to Very Dark— When it
is Translated into as ASCII Region, It becomes a Garish Series of Stripes ( !!!
) That is Unbearable !
So
The Solution that i incorporated into my ASCII Program; Originally Pioneered on
my HP48sx Programmable Calculator— Was to allow The Program to PseudoRandomly
Choose a Character from A List or Region in A Continuum of Characters for Each
Gray Shade !
e.g.;
A Given Shade of Middling Gray may be Either a : y 2 3 6 9 or P
This
Results in a ‘Noisier’ Gray Region,
Which
seems to ‘Imply’ a Greater Resolution !
Such
a Rendering Looks Something Like This :
The
Chances of Two of The Same Characters Appearing Side by Side given that The Two
Gray Swatches are Equal is The First Character x The Second Character from a
List of Six Characters or 1 in 6.
Three
Characters Sides by Side would be 1 x 6 x 6 or : 1 in 36
But.
i
thought that this would be ‘Too Often !’
So
i wrote a little Loop into my Program that would try to Find a Unique Character
3 Times.
So
that; Now— Two Characters side by side would be 1 x ( 6 x 6 x 6 )
or
1 in 216.
But
in a Very Large ASCII Picture, is is bound to happen many times.
But
3 Characters side by side would 1 in 432.
&
that would happen many times still, When— The Three Gray Sampled Patches are
all Equal, which happens pretty often.
So
for there to be a Truly Anomalous Region to Exist;
There
has to be Quite a Bit of Repetition, & Additionally; By my Reckoning, There
has to be Several Lines, Vertically Arranged, in a Particular Horizontal
Region—
In
Other Words; A 2D Blob of Repetition !
The
Bigger this Blob is; The More Anomalous it is. ( !!! ) ( ? )
The
Chances of this Occurring, According to my Calculation is :
Something
like 1.0 with 57,132 Zeros after that.
In
other Words; Very Unlikely for this Tiny Swatch.
This
one as shorter Horizontal Repetition Sequences,
Each
only Bucking The Odds to Something like 10_million per line,
But
then they are arranged Vertically, about 7 high;
So
that makes it about ( ? ) 1 with 300,000 Zeros to 1.
Very
Unlikely.
So
when one of these shows up :
What
does this mean ?
The
Fact that these are occurring very irregularly;
Means
that The Program is Working as it should, most of The Time.
Nearly
all of The Time, The Distribution looks exactly as it should.
What
is Going on ?
i’m
thinking that it -Must Be- somekind of weird anomaly in The Scratch
Application, but How or What is it doing ( ? )
This
is The Whole ‘Kaden’ Image that The Really Weird Patch Came from; With The
Other ‘Anomalous’ Regions in Red.
No comments:
Post a Comment