Preserved Topic: Hypodermic sitamication (Page 1 of 1)

Ok, just for the sake of argument, lets say I create a 400x400 px square. Then I take that square and stroke a 1 px line around the outside of it.

Now lets say I resize that puppy to 200x200 px. How does Photoshop now create a 1/2 px line around the square?

Just something that popped into my unoccupied mind earlier today.

It uses half tones.

Off the top of my head, I'd say it gives you a 50% tranparent anti-aliased black line around the square.

You mean to tell me you calculated that off the top of your head? Sheesh.

I'm saying without Photoshop in front of me.

I hope that wasn't I serious answer.

1 px thin line with RGB values of 143 a piece.

It looks to me as if it leaves the line one pixel wide and reduces its color to 25%.

quote:

---

I hope that wasn't I serious answer.

---

I'm not exactly sure what's going on here, so forgive me while I ignore you.

I tried the idea... made a 400 px square, put a 1 px black line around the outside of it, then resized it 50%

What I had in the end was a 200 px square with a grey value of 137 line around the inside of the square, and a 50% transparent line down the right side and across the bottom, outside the square, meeting at a 25% px in the bottom right. I verified the transparency by ctrl-clicking the layer and then filling a black channel with white. The outside line to the right and bottom were filled with 127 grey, which indicates a transparency of 50%. Feel free to check it yourself.



[This message has been edited by Human Shield (edited 02-08-2001).]

How you resample makes a difference. Nearest Neighbor, Bilinear, Bicubic. All work differently. It's at the bottom of the Image > Image Size dialoge.

Image size was not used in my experiment. He did not say to resize the image, but just the square, and I was following his steps. There's no choice for resampling with the scale feature. I just tried it with image size, and I ended up with a 200 px square with a light line (138 grey) around the inside and a 76 grey line around the outside. No transparency. So it's all in how you do it.

There's no choice for what??

Go to FILE>>>PREFERENCES>>>GENERAL

Set your Interpolation to Nearest Neighbor....Now go back to your PSD and scale the layer.

Guess what happens?


Peter





ICQ:# 10237808

Okay, I was sort of correct.

BTW, I just checked my Preferences and that is set to Bicubic -- my favorite! Everytime I set up PS, I go through and set things like that and then forget about 'em.

Ah, that. It's not exactly a part of the scale feature itself, but it will have to do in a pinch.

Now, changing it to what DG said, when I scaled the square down, I ended up with a 200 px square with a black line around the right and bottom side, and no line around the other...

I know this is a very long post. I probably should have taken the time to reduce the amount of information here, but, I' getting ready to leave for the weekend and wnated to get this out before I left and this pos gets too old:


Follow this procedure:

1. Create a new image:
2. Select the entire image:
3. Select PhotoShop Default Colors.
4. Stroke the Selection
5. Resize the image.

The PhotoShop commands to do this are:
1. Ctrl-N: File->New...(400 pixels, 400 pixels, 72 dot/inch, Mode(RGB),
Contents(White) ).

2. Ctrl-A: Select --> Select All

3. D: (Click on the little black and white squares under the foreground
color square in the tool palette)

4. Edit --> Stroke...(1 pixel, Inside, 100%, Normal)

5. Image --> ImageSize ... (200 pixels, 200 pixels, _, _, _,
Constrained, Resample(Bicubic) )
- The '_'s in the line above mean don't change these values. In the
future I will denote this command with ImageSize(200, 200, Bicubic)


Notes on commands:
1. The size and resolution of the image make no difference for this
experiment. I tried several images with different sizes and different
resolutions and got the same results for each.

3. Set the foreground to 100% Black and the Background to 0% White.

4. The center of your image is now White RGB(255,255,255) = K(0%) and the one
pixel line around the image is Black RGB(0,0,0) = K(100%).

5. There are several options for resizing the image. You can resize the image
using Image-->ImageSize... or you can resize all or part of the image using
Edit->Transform.

There are actually several Transform methods. I chose to use
Edit-->Transform->Numberic for this experiment just so I knew I was getting
the same amount of scaling each time.

There are three interpolation methods used with each method of scaling:
Bicubic, Bilinear, and Nearest Neighbor.

I followed the above procedure for ImageSize with Bicubic, Bilinear, and
Nearest Neighbor. Then I did the same thing with
Transform->Numberic...(No Position, Scale(50%, 50%), No Skew, No Rotation),
using File-->Preferences-->General...(Interpolation(*)) to set the
interpolation method to Bicubic, Bilinear, and Nearest Neighbor. These
experiments were run with the line drawn on the white of the Background
layer. Next I re-ran all experiments with the White background, but this
time I created a new layer for the Black outline to live in. Another set
of experiments creates the White Background and a new layer, which I
filled with white before stroking the Black Square on that layer.

Note: In all the Bicubic cases the resize operation creates three
different colors: The lines that make up the edges of the square
became one color, the corner pixels became another color, and the
pixels on the vertical lines just above and just below the corner
pixels became a third color.

Note: In all the Bilinear cases the resize operation creates a new
square with the edge pixels equal to one color and the corner
pixels equal to another.

Note: In all the Nearest Neighbor cases the resize operation creates a
solid black line along the right and bottom edges of the square
and the lines along the top and left edges disappeared.

Note: In the cases where the Black Square and the White Background are
on two separate layers, there are two readings for each set of
pixels. The first reading indicates the color of the pixels with
all layers visible. The second indicates the color of the pixels
with only the Black Square's layer visible.

As a variation on this experiment, I created a new layer between the
Background layer and the Black Square layer, after resizing the image.
I filled just the part of the second layer that was covered by the
black pixels in the Black Square layer with White. The results were
exactly the same as with the White Background.

The following shows the results of these experiments:

ImageSize(200, 200, Bicubic) Black square on White Background layer.
lines RGB(128, 128, 128) = K(50%)
corners RGB( 64, 64, 64) = K(75%)
next to RGB(137, 137, 137) = K(46%)

ImageSize(200, 200, Bilinear) Black Square on White Background layer.
lines RGB(128, 128, 128) = K(50%)
corners RGB( 64, 64, 64) = K(75%)

ImageSize(200, 200, Nearest Neighbor) Black Square on White Background layer.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0,) = K(100%)

ImageSize(200, 200, Bicubic) Black Square on a White layer.
lines RGB(135, 135, 135) = K(47%) / RGB(128,128,128) = K(50%)
corners RGB( 86, 86, 86) = K(66%) / RGB( 64, 64, 64) = K(75%)
next to RGB(142, 142, 142) = K(44%) / RGB(137,137,137) = K(46%)

ImageSize(200, 200, Bilinear) Black Square on a White layer.
lines RGB(128, 128, 128) = K(50%) / RGB(128,128,128) = K(50%)
corners RGB( 64, 64, 64) = K(75%) / RGB( 64, 64, 64) = K(75%)

ImageSize(200, 200, Nearest Neighbor) Black Square on a White layer.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0) = K(100%) / RGB(0, 0, 0) = K(100%)

ImageSize(200, 200, Bicubic) Black Square and White Background
on separate layers.
lines RGB(143, 143, 143) = K(44%) / RGB(0,0,0) = K(100%)
corners RGB( 94, 94, 94) = K(63%) / RGB(0,0,0) = K(100%)
next to RGB(148, 148, 148) = K(42%) / RGB(0,0,0) = K(100%)

ImageSize(200, 200, Bilinear) Black Square and White Background
on separate layers.
lines RGB(127, 127, 127) = K(50%) / RGB(0,0,0) = K(100%)
corners RGB( 63, 63, 63) = K(75%) / RGB(0,0,0) = K(100%)

ImageSize(200, 200, Nearest Neighbor) Black Square and White Background
on separate layers.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0) = K(100%) / RGB(0, 0, 0) = K(100%)

Scale(50%, 50%, Bicubic) Black Square on White Background layer.
lines RGB(135, 135, 135) = K(47%)
corners RGB( 86, 86, 86) = K(66%)
next to RGB(142, 142, 142) = K(44%)

Scale(50%, 50%, Bilinear) Black Square on White Background layer.
lines RGB(128, 128, 128) = K(50%)
corners RGB( 64, 64, 64) = K(75%)

Scale(50%, 50%, Nearest Neighbor) Black Square on White Background layer.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0) = K(100%)

Scale(50%, 50%, Bicubic) Black Square on a White layer.
lines RGB(135, 135, 135) = K(47%) / RGB(128,128,128) = K(50%)
corners RGB( 86, 86, 86) = K(66%) / RGB( 64, 64, 64) = K(75%)
next to RGB(142, 142, 142) = K(44%) / RGB(137,137,137) = K(46%)

Scale(50%, 50%, Bilinear) Black Square on a White layer.
lines RGB(128, 128, 128) = K(50%) / RGB(128,128,128) = K(50%)
corners RGB( 64, 64, 64) = K(75%) / RGB( 64, 64, 64) = K(75%)

Scale(50%, 50%, Nearest Neighbor) Black Square on a White layer.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0) = K(100%) / RGB(0, 0, 0) = K(100%)

Scale(50%, 50%, Bicubic) Black Square and White Background
on separate layers.
lines RGB(143, 143, 143) = K(44%) / RGB(0,0,0) = K(100%)
corners RGB( 94, 94, 94) = K(63%) / RGB(0,0,0) = K(100%)
next to RGB(148, 148, 148) = K(42%) / RGB(0,0,0) = K(100%)

Scale(50%, 50%, Bilinear) Black Square and White Background
on separate layers.
lines RGB(127, 127, 127) = K(50%) / RGB(0,0,0) = K(100%)
corners RGB( 63, 63, 63) = K(75%) / RGB(0,0,0) = K(100%)

Scale(50%, 50%, Nearest Neighbor) Black Square and White Background
on separate layers.
top and left edges: No Lines.
bottom and right edges: RGB(0, 0, 0) = K(100%) / RGB(0, 0, 0) = K(100%)


Note: Increasing the size of the image is a totally different story
and produces very different results.

Note: Using ImageSize you can scale an image down by reducing the number
of pixels in the image, reducing the total height and/or width of
the image, or by increasing the pixel resolution of the image.
All three of these will produce the above results.

Note: Interpolation method and the number of layers involved seem to have the
most effect on the results of resizing an image. However, if you look
carefully at ImageSize(Bicubic) and Scale(Bicubic) for the Black Square
on the White Background layer you will see there are differences there
too.

Note: These experiments were performed with PhotoShop v5.02.


Summary:
ImageSize
Background White Layer Separate Layer
Bicubic 50, 75, 46 47, 66, 42 44, 63, 42
50, 75, 46 100,100,100

Bilinear 50, 75 50, 75 50, 75
50, 75 100,100

Nearest 100 100 100
100 100

Scale
Background White Layer Separate Layer
Bicubic 46, 66, 44 47, 66, 42 44, 63, 42
50, 75, 46 100,100,100

Bilinear 50, 75 50, 75 50, 75
50, 75 100,100

Nearest 100 100 100
100 100


The column labeled Background lists the intensities when the Black Square
resides on the White Background layer.

The column labeled White Layer lists the intensities when the Black Square
resides on a separate layer from the Background. The Background layer and
the separate layer are both filled with white and then the Black Square is
drawn on the layer.

The column labled Separate Layer lists the intensities when the Black
Square is drawn on a layer by itself and the Background layer is filled
with white.

The numbers in each column are the black intensities of the various parts
of the resulting squares. For example, 44 means 44% black
or RGB(142, 142, 142).

Each column has sets of one, two or three numbers (horizontally). If there
is one number in the column it represents the line intensity. Two numbers
represent the line intensity and the intensity of the corner points. Three
numbers represent the line intensity, corner point intensity, and the
intensity of the points on the vertical line just above and just below the
corner points.

Some columns have two sets of numbers (vertically) for each row. When the
Black Square is on a layer separate from the Background Layer, turning the
Background off or on changes the intensity of the pixels in the square
layer. The upper numbers are the square intensities with the background on.
The lower numbers are the intensities with the background off.

way too much time on yer hands hyperbole! *hands ya a weekend pass* ya need a break me thinks!

Now, that` a serious answer, hyperbole. That much typing would have used up my weekend

OK. no hard questions until hy gets some meds...



Strangers have the best candy..

Knowledge. Gotta love it.

Hyperbole, thanks.