Resize question

MS
Posted By
Matt Silberstein
Apr 11, 2005
Views
290
Replies
5
Status
Closed
Suppose I have an image that is 800×600. If I resize this to be 801×601 every pixel is modified. In my naiveté it seems to me that this would produce a worse result than, say, taking it to 1600×1200, where each pixel just becomes 4 pixels. Is there something were there are preferred % to expand or are the algorithms good enough that I don’t have to game it? My specific application involves taking photos and essentially making a collage. I have some control over the original size of the item in the photo, but I can’t make them the right size.

—
Matt Silberstein

All in all, if I could be any animal, I would want to be a duck or a goose. They can fly, walk, and swim. Plus,
there there is a certain satisfaction knowing that at the end of your life you will taste good with an orange sauce or, in the case of a goose, a chestnut stuffing.
Related Tags
question resize

Must-have mockup pack for every graphic designer 🔥🔥🔥

Easy-to-use drag-n-drop Photoshop scene creator with more than 2800 items.

LJ
Lurkis Jerkis
Apr 11, 2005
"Matt Silberstein" wrote in
message
Suppose I have an image that is 800×600. If I resize this to be 801×601 every pixel is modified. In my naivet
MS
Matt Silberstein
Apr 11, 2005
On Mon, 11 Apr 2005 13:36:23 -0400, in comp.graphics.apps.photoshop , "Lurkis Jerkis"
wrote:

"Matt Silberstein" wrote in
message
Suppose I have an image that is 800×600. If I resize this to be 801×601 every pixel is modified. In my naiveté it seems to me that this would produce a worse result than, say, taking it to 1600×1200, where each pixel just becomes 4 pixels. Is there something were there are preferred % to expand or are the algorithms good enough that I don’t have to game it? My specific application involves taking photos and essentially making a collage. I have some control over the original size of the item in the photo, but I can’t make them the right size.
why are you adding the extra pixel, and why not simple adjust the canvas size instead?
I did not explain that well. I gave some extreme cases, adding just a single pixel vs a double, as examples. I was wondering if adding in "whole" amounts led to better results than adding "odd" amounts. Take a more reasonable example, suppose I have some control over the initial image, would taking a somewhat smaller image then doubling the result be better than taking a somewhat larger image and making it 50% larger? I am trying to make some reasonable test cases, but I wondered if anyone had already worked this out. And, I suppose, the same question goes for shrinking an image.

—
Matt Silberstein

All in all, if I could be any animal, I would want to be a duck or a goose. They can fly, walk, and swim. Plus,
there there is a certain satisfaction knowing that at the end of your life you will taste good with an orange sauce or, in the case of a goose, a chestnut stuffing.
LJ
Lurkis Jerkis
Apr 11, 2005
"Matt Silberstein" wrote in
message
On Mon, 11 Apr 2005 13:36:23 -0400, in comp.graphics.apps.photoshop , "Lurkis Jerkis"
wrote:

"Matt Silberstein" wrote in
message
Suppose I have an image that is 800×600. If I resize this to be 801×601 every pixel is modified. In my naivet
T
Tacit
Apr 11, 2005
In article ,
Matt Silberstein wrote:

I did not explain that well. I gave some extreme cases, adding just a single pixel vs a double, as examples. I was wondering if adding in "whole" amounts led to better results than adding "odd" amounts. Take a more reasonable example, suppose I have some control over the initial image, would taking a somewhat smaller image then doubling the result be better than taking a somewhat larger image and making it 50% larger?

You’re thinking along the wrong line; instead of asking yourself which is better based on the ease of the math behind the interpolation, you should be thinking about the total amount of information contained in the image.

Because one pixel is the smallest unit of a raster image–a pixel can have only one color value–the amount of information in an image is fixed and quantifiable. If you start with a large image and you sample it up by 50%, you are spreading that same amount of information across a larger number of pixels; but if you start with a smaller image and double the number of pixels, you are spreading a smaller amount of information across those pixels.

When you double the size of an image, you do not just take each pixel and make it into four pixels (at least not if you’re using anything other than nearest-neighbor interpolation). It’s always best to start with the greatest amount of information you can, since you can not add new information by interpolating the image.

This can be proven mathematically, provided you have sufficient background in signal processing theory, but proving it mathematically is unnecessary; you can determine this for yourself quite easily. 🙂

Scan an image twice, with one scan smaller than the other, then interpolate both up to the same size. Regardless of whether your target size is an even multiple of the resolution of the smaller scan or not, the results from the larger scan will look better than the results from the smaller scan.

—
Art, photography, shareware, polyamory, literature, kink: all at http://www.xeromag.com/franklin.html
M
MorituriMax
Apr 14, 2005
"tacit" wrote in message

much good stuff snipped for brevity

Wow.. good information.

Must-have mockup pack for every graphic designer 🔥🔥🔥

Easy-to-use drag-n-drop Photoshop scene creator with more than 2800 items.

Related Discussion Topics

Nice and short text about related topics in discussion sections