In photographic material (after scanning at 24bits) or in created art
which has effects like gradients etc(24 bit), each pixel is a 24bit
pixel. This can be seperated into, for example, R, G and B channels
and each channel seperately viewed.
The question is whether when an image is enlarged, say to 100 times
its size, will each original pixel become transformed into a large
chunky glob (100 x 100 of the original pixel) or will photoshop
intelligently enlarge it so that there are no 'globs'. For example,
will a simple linear gradient remain the same (except for total size)
even after getting enlarged by a factor of 1000 or will it become
granular?
I have tried and it seems like photoshop does this 'intelligent;y'. If
this is so, can anyone point out software or a technique that will NOT
do this 'intelligently'

Hi there,

When you use the Image: Image Size dialog box, there is an option at
the bottom to "Resample Image".

If you choose the option "Nearest Neighbor" (as opposed to the default
"Bicubic", it will enlarge the image pixel-by-pixel verbatim without
any "intelligent" anti-aliasing or interpolative smoothing.

For an example, see http://www.eyewire.com/tips/mini/nearest.html

Please let me know if you have any questions.

--Joey

For more sites on this, check the following straegy: "nearest neighbor" photoshop

---

Clarification of Answer by joey-ga on 06 Aug 2004 22:41 PDT

In the last line, I meant "Google searching strategy".

--Joey

excelleent! thanks!

Hey there,

I a student in computer science and have worked on ray-tracers and
other graphics related projects.  I have written software specifically
to handle image resizing.

To answer your question, you first need to understand the difference
between bitmap-based graphics and vector-based graphics.  In
bitmap-based graphics each pixel is saved with an RGB value.  This is
currently the most common form of graphics editors.  Vector based
graphics instead use math formulas to represent shapes and objects. 
Because they consist of points and objects rather than pixels on a
grid, vector-based graphics are free from the confines of pixel
resolution.

You can read more about the difference between the two types of image
editors here: http://graphicssoft.about.com/cs/findsoftware/a/twotypes.htm

More and more graphics programs are blending the features of both
bitmap-based image editing software and vector-based software.  This
is exactly where Photoshop is right now.  When Photoshop enlarges an
image, it will first use vector based technology if available,
otherwise it will enlarge it using a standard resize matrix.  After
enlarging the image it may adjust the image or apply other effects
such as blur to hide the distortion.

I'm sorry that I don't know more specifics about Photoshop.  I don't
use the program much.  Let me know if I can clarify anything and I
hope this helps at least a little!

Josh Fraser

hi,
I'm a graphic student and i got this little advice on how to enlarge
small photos in poor resolution and size to bigger better photos from
a friend.

first if the photo is really in poor quality the best is to print it
on photo paper in best quality. then scan it.

now lets say you have a 10x15 photo and you want to make it a 40x60.
go to image size and instead of resizing the photo to 40x60 make it
bigger only by 2%. keep ding that over and over again (you can make it
and action) until the photo is about 45x67.5 now do another image size
and reduce it to 40x60.

by doing this you trick photshop to better handle the pixel resize.
since in bicubic mode photoshop sample the pixels and creates an
average pixel in-between. by enlarging the photo slowly the pixels you
get in between the 2 originals are more accurate since they have the
average value of the average value and so on and so on.

try it out you won't be disappointed.

michael06-ga

I did as you wrote, enlarging by 2%...for me this trick did not work.
I've got really poor resault....maybe I did something wrong?!

Michael --

First a question . . . how did you enter the 2%?  Using my technique,
you can only enter a destination size in pixels -- you can't use a
percentage to increase.

Second, say you started with something that's 20x20, if you want to
"unintelligently" (e.g. without interpolation) enlarge it, it is not
going to look smooth unless you enlarge it in multiples of the
original file size (e.g. 40x40, 60x60, etc.)  This isn't really a bug,
it's just the nature of avoiding interpolation -- because it isn't
"smoothing" the pixels in-between, it has to choose a solid color for
every pixel, and if it's an odd in-between pixel when enlarging from
20x20 to 23x23 or something), it's going to appear jagged.  However,
above a certain point (4x +), you can choose any value, and it will
still appear decently.

--Joey

Sorry, ignore my last comment . . . I was responding to shakibgd and
thought he was confused about my initial procedure.

Carry on. <grin>