This is Interesting: Free Magazines for Graphics designers and webmasters
Home > Archive > PainShop Pro Scripting > November 2007 > script to make clear 3 value images
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
script to make clear 3 value images
|
|
| Cathy Hasty 2007-11-14, 10:18 pm |
| underprocessable | |
| Spandex Rutabaga 2007-11-15, 3:17 am |
| Cathy Hasty wrote:
>
> Hello,
> I am trying to make value studies of photos to help me paint.
You are using "value", which is an expression from art. It would
help if you explained the art part of it to the rest of us. I
think it is just another way of saying lightness or the greyscale
level corresponding to a color. Correct me if it means something
else to you. I'm not familiar with all this tone, tint, shade
and value stuff.
> I know some
> ways to do this using tools like the scripts in PSP, grayscale, Black and
> white but none of them get at what I want.
It's kind of hard for us to give you what you want when you haven't
explained exactly what it is you want. Assume we are idiots who
know nothing about how artists talk about color or what they might
do in preparing to paint and explain step by step what it is you
want.
> So far I can't get three good
> definitions. Can anyone help me build a script that would have gray white
> and black?
What do you mean by "have gray white and black"? It's easy to create
a script that puts splodges of grey, white and black in an image
but I doubt that's what you want. I suspect that these blacks, greys
and whites are somehow supposed to be derived from the colors of the
image but you haven't described this part of what you want. Also, in
PSP black is a color in which red = green = blue = 0 and white is a
color in which red = green = blue = 255. In other words, each of
them is a *single* *unique* color in up to 16.7 million possible
image colors. However, you make blacks and whites be plural words
so you need to explain by what rules you want to make colors ranges
you can call blacks and whites.
> I have tried a few myself with little success.
You've attached two (unnecessarily large) images. One is apparently
an original from which you want to extract some kind of information
about blacks, greys and whites. I don't know what the other is
supposed to be. Have a go at explaining it since a vegetable like
me is not smart enough to figure it out. I can make a range of
various suggestions of how to isolate different lightness ranges
in the image but without knowing what you want to end up with it
would be a waste of your time and mine. The attached shows some
options for displaying the lightness ranges of color.
| |
| Cathy Hasty 2007-11-16, 3:18 am |
| underprocessable | |
| Spandex Rutabaga 2007-11-16, 3:18 am |
| Cathy Hasty wrote:
> Whenever you, SR, write the book on Corel PSP X*.*, I want to be first in
> line! I see so many Adobe resources but not enough in print about PSP. Am
> I missing something?
Yes, writing books doesn't pay well but takes hard work, and
the vast majority of people doesn't want in-depth information
anyway. They want the image editing equivalent of "Neurosurgery
for Dummies", or "Being a brain surgeon is only a scalpel click
away", or "Microsoft Neurohack 2.0103 Home Edition".
> I am sorry that my question was not clear and my illustrations were too
> large. I don't know what a reasonable size is for sending a clear picture
> to view on email. I noticed that 50 worked well for a complex subject sent
> so I hope these smaller ones work for less complexity.
I'm sorry but 50 what? What units? In which part of the program?
I'll take a wild guess and say you are talking about the JPEG
compression level. If so, a level of 50 is too high and will
cause artifacts in the image. Stay in the range 15 to 30. There
are two things that determine the file size of an attachment:
(1) the JPEG compression level and (2) the size of the image in
pixels. Use Image > Resize to make the larger of your image
width or height be no more than 900 pixels. Resize using Smart
Size. After resizing apply the compression level I suggested.
> I will present my question again with smaller illustrations from someone who
> is teaching the idea of the value sketch.
>
> I attached two value sketches and their corresponding photo.
> Admittedly the "Photo sketchbook-gramere-lane" is a pretty bad photo. The
> value sketch from that photo "Value Sketch sketchbook-gramere-lane.." shows
> significant revisions.
> When I work I do the revisions on the computer before I make the value or
> tone sketch.
>
> There is a great value sketch of fruit from the photo of fruit.
>
> What I want to do is have the computer make my photo (after I manipulate
> it ) into a value sketch with three tones of white, gray and Black. There
> will not be much graduation between the three tones or values.
Would the following be an accurate description of what you want?
1. First convert the image to greyscale, which is the lightness
corresponding to each color and which ranges from 0 to 255.
2. Divide the greyscale range into three almost exactly equal
intervals, namely 0 to 84 (blacks), 85 to 170 (greys), and
171 to 255 (whites).
3. Mark in black on a white background the pixels in the image
which lie in each of these ranges, either as separate images or
as layers in the original image.
I'm not at all sure about a number of things:
1. Whether you want the ranges to be as nearly equal as possible
or whether you want greys to be a wider range than blacks or
whites.
2. Whether you need the ranges to be adjustable or not.
3. Whether you need more than three ranges.
4. Whether you want the different value (i.e. greyscale range)
regions in separate images as black against white, or all in
the same image colored respectively black (42), grey (128)
and white (213). The numbers in parentheses are the average
greyscale levels in each of the three greyscale ranges.
> What the
> value sketch does is to make the photo as simple as possible so the lights
> and darks are what tell you what the subject is. With less information I
> can see the space divisions and shift them around according to composition
> theory.
I'm afraid I know nothing about composition theory and I don't
understand *how* you plan to shift *what* around. In other words,
I'm fishing for a clue as to what PSP edits you want to do if I
give you some kind of image you want. You seem to be planning on
some kind of trial and error approach and I can't understand
what the implications might be for a script intended to help you.
Is the final idea that the image should contain equal areas of
blacks, greys and whites? Is that what composition theory dictates?
If so, this might be tricky since there doesn't seem to be a way
to access histogram information in scripts, and counting colors
by sampling with the Eyedropper is incredibly slow (since one
PSP command is executed for every single pixel in the image).
As far as histograms are concerned, if you hold the mouse over
the histogram display you can read off what fraction of image
pixels lie below and up to a particular level in a tooltip. If
you find the 33% and 67% points you can threshold at these
brightness values to split the image into three lightness zones,
each with an equal number of pixels. This you can do if you work
by hand. Scripts don't give you access to the information. An
example of what I mean is attached.
> The painting I sent earlier of an iris illustrates the outcome
> after a value sketch is used to work out the details.
Sorry but I didn't make much of that. It didn't seem to match
the color image you included in the same post and there was no
original of the iris image, so it was hard to guess what might
have been done and with what purpose.
> In looking that the images that SR sent, the closest to this division seems
> to be "The areas of the medium colors only shown in black and white. " How
> did you achieve that?
Two duplicate layers of the greyscaled image, one thresholded at
at a low value to mark blacks and the other thresholded at a higher
value to mark blacks together with dark greys. The top layer was
set to the difference blend mode so that what you see is the
difference between these two layers, which is the range of dark
greys without the range of blacks. The final twist is to add a
Negative adjustment layer to the top of the layer stack, which
converts the white differences on a black "no-difference" background
to black areas against a white background.
| |
|
| Cathy Hasty wrote:
> What I want to do is have the computer make my photo (after I manipulate
> it ) into a value sketch with three tones of white, gray and Black. There
> will not be much graduation between the three tones or values. What the
> value sketch does is to make the photo as simple as possible so the lights
> and darks are what tell you what the subject is. With less information I
> can see the space divisions and shift them around according to composition
> theory. The painting I sent earlier of an iris illustrates the outcome
> after a value sketch is used to work out the details.
Do you really need it to be a 3-colour image (1 white, 1 black and 1
shade of grey)? It would be easier to get a recognisable result if there
could be a few more shades of grey. Try converting the photo to
greyscale (Image > Greyscale) and then going to Effects > Artistic
Effects > Posterize - there you can tweak the number of levels so that
there are as few as possible but the image is still recognisable. If the
darkest colour that you get isn't pure black, you can select it
afterwards with the magic wand (contiguous unchecked) and flood fill
with black (match mode none). You can also combine some of the greys by
selecting and flood filling.
--
tape
| |
| Spandex Rutabaga 2007-11-16, 6:23 pm |
| tape wrote:
> Do you really need it to be a 3-colour image (1 white, 1 black and 1
> shade of grey)? It would be easier to get a recognisable result if there
> could be a few more shades of grey.
That concerned me too along with how to establish the useful
ranges of each of the shades.
> Try converting the photo to
> greyscale (Image > Greyscale) and then going to Effects > Artistic
> Effects > Posterize - there you can tweak the number of levels so that
> there are as few as possible but the image is still recognisable. If the
> darkest colour that you get isn't pure black, you can select it
> afterwards with the magic wand (contiguous unchecked) and flood fill
> with black (match mode none). You can also combine some of the greys by
> selecting and flood filling.
That's an excellent suggestion if you don't need to control the
exact number of colors and the fraction of pixels in the image
that contain a particular color. I wish I had remembered it :)
I've attached an example of the effect, which also illustrates
some of the problems you alluded to (color count and result color).
| |
| Warren Moore 2007-11-16, 6:23 pm |
| Hello Y'all...
As a painter who is trying to become a photographer at a late
stage in life, here's the definition in "vegetable" terms. <grin>
"Value" when talking about pigments is nothing more than
grayscale...either solid black to solid white, or from a color at
it's saturated maximum to white. "Composition" is where on the
canvas (or paper or whatever) the painter wishes to place the
picture elements...nothing more. That's why my wife says I paint
nice pictures, but take lousy photos...I can move things around
on a canvas.
A value sketch such as Cathy is talking about (in three or so
shades) gives an artist a better shot at recognizing the major
elements. The old painter's trick is to squint at the scene,
which de-emphasizes the colors and allows the darks and lights to
"pop" out.
Warren Moore
"Spandex Rutabaga" <SpRu@agabatur.xednaps> wrote in message
news:473D3289.7340C41E@agabatur.xednaps...
> Cathy Hasty wrote:
>
>
> Yes, writing books doesn't pay well but takes hard work, and
> the vast majority of people doesn't want in-depth information
> anyway. They want the image editing equivalent of "Neurosurgery
> for Dummies", or "Being a brain surgeon is only a scalpel click
> away", or "Microsoft Neurohack 2.0103 Home Edition".
>
>
> I'm sorry but 50 what? What units? In which part of the
> program?
> I'll take a wild guess and say you are talking about the JPEG
> compression level. If so, a level of 50 is too high and will
> cause artifacts in the image. Stay in the range 15 to 30. There
> are two things that determine the file size of an attachment:
> (1) the JPEG compression level and (2) the size of the image in
> pixels. Use Image > Resize to make the larger of your image
> width or height be no more than 900 pixels. Resize using Smart
> Size. After resizing apply the compression level I suggested.
>
>
> Would the following be an accurate description of what you
> want?
> 1. First convert the image to greyscale, which is the lightness
> corresponding to each color and which ranges from 0 to 255.
> 2. Divide the greyscale range into three almost exactly equal
> intervals, namely 0 to 84 (blacks), 85 to 170 (greys), and
> 171 to 255 (whites).
> 3. Mark in black on a white background the pixels in the image
> which lie in each of these ranges, either as separate images
> or
> as layers in the original image.
>
> I'm not at all sure about a number of things:
> 1. Whether you want the ranges to be as nearly equal as
> possible
> or whether you want greys to be a wider range than blacks or
> whites.
> 2. Whether you need the ranges to be adjustable or not.
> 3. Whether you need more than three ranges.
> 4. Whether you want the different value (i.e. greyscale range)
> regions in separate images as black against white, or all in
> the same image colored respectively black (42), grey (128)
> and white (213). The numbers in parentheses are the average
> greyscale levels in each of the three greyscale ranges.
>
>
> I'm afraid I know nothing about composition theory and I don't
> understand *how* you plan to shift *what* around. In other
> words,
> I'm fishing for a clue as to what PSP edits you want to do if I
> give you some kind of image you want. You seem to be planning
> on
> some kind of trial and error approach and I can't understand
> what the implications might be for a script intended to help
> you.
>
> Is the final idea that the image should contain equal areas of
> blacks, greys and whites? Is that what composition theory
> dictates?
> If so, this might be tricky since there doesn't seem to be a
> way
> to access histogram information in scripts, and counting colors
> by sampling with the Eyedropper is incredibly slow (since one
> PSP command is executed for every single pixel in the image).
>
> As far as histograms are concerned, if you hold the mouse over
> the histogram display you can read off what fraction of image
> pixels lie below and up to a particular level in a tooltip. If
> you find the 33% and 67% points you can threshold at these
> brightness values to split the image into three lightness
> zones,
> each with an equal number of pixels. This you can do if you
> work
> by hand. Scripts don't give you access to the information. An
> example of what I mean is attached.
>
>
> Sorry but I didn't make much of that. It didn't seem to match
> the color image you included in the same post and there was no
> original of the iris image, so it was hard to guess what might
> have been done and with what purpose.
>
>
> Two duplicate layers of the greyscaled image, one thresholded
> at
> at a low value to mark blacks and the other thresholded at a
> higher
> value to mark blacks together with dark greys. The top layer
> was
> set to the difference blend mode so that what you see is the
> difference between these two layers, which is the range of dark
> greys without the range of blacks. The final twist is to add a
> Negative adjustment layer to the top of the layer stack, which
> converts the white differences on a black "no-difference"
> background
> to black areas against a white background.
| |
| Spandex Rutabaga 2007-11-16, 6:23 pm |
| Warren Moore wrote:
>
> Hello Y'all...
>
> As a painter who is trying to become a photographer at a late
> stage in life, here's the definition in "vegetable" terms. <grin>
>
> "Value" when talking about pigments is nothing more than
> grayscale...either solid black to solid white, or from a color at
> it's saturated maximum to white.
Yep, that I know.
> "Composition" is where on the
> canvas (or paper or whatever) the painter wishes to place the
> picture elements...nothing more.
That I know too and just a very tiny bit more (and I'm not
exaggerating :) about creating lines the eye will follow and
about the rule of thirds.
> That's why my wife says I paint
> nice pictures, but take lousy photos...I can move things around
> on a canvas.
You can move things around in a photo too. Welcome to the world
of digital editing :)
> A value sketch such as Cathy is talking about (in three or so
> shades) gives an artist a better shot at recognizing the major
> elements.
I understand that sort of, which means I don't know how she wants
to define her shade ranges and whether she wants to adjust them
based on what she see in the value sketch.
> The old painter's trick is to squint at the scene,
> which de-emphasizes the colors and allows the darks and lights to
> "pop" out.
You can greyscale in a digital editor and Gaussian blur to remove
distracting detail. There's a good tutorial in using a similar
technique for selecting matting colors for an image. Here it is:
http://www.fortunecity.com/westwood...orials/matting/
| |
| Warren Moore 2007-11-23, 6:21 pm |
| Hi Spandex. I should have guessed that you'd know more about the
subject than you let on. I think that the problem here is that
Cathy knows precisely what she wants, but can't describe it in
finite terms that translate directly to PSP. It's the old "three
blind men describing an elephant" problem.
The last two examples she sent over (the fruit and grassmere
lane) don't really help a great deal. The fruit value study is
really a grayscale with many more than three values (if it
weren't in watercolor it could be easily used as an underpainting
in opaque media); while the grassmere lane is in pencil, which is
technically only two values: graphite and white (with
intermediate values based on stroke width and pencil
pressure...so again you have more than three values).
I think that your suggestion:
"1. First convert the image to greyscale, which is the lightness
corresponding to each color and which ranges from 0 to 255.
2. Divide the greyscale range into three almost exactly equal
intervals, namely 0 to 84 (blacks), 85 to 170 (greys), and
171 to 255 (whites).
3. Mark in black on a white background the pixels in the image
which lie in each of these ranges, either as separate images or
as layers in the original image."
Fits her original request of only three tones. However, thank
you very much for the link:
http://www.fortunecity.com/westwood...rials/matting/,
which I think could also be used to produce what Cathy is asking
for. "Method 2" with a less-dramatic blur might work very well.
It would show the basic composition...which I also think is what
she's looking for. But, I could be way off-base...I usually am
at least once every four years when I step into the voting booth.
Warren
|
|
|
| | Copyright 2003 - 2009 forum4designers.com Software forum Computer Hardware reviews |
|