Thorsten Froehlich wrote:

> 
> Sorry to say, but your guessing is not only wrong but also superfluous: 
> Photoshop has two modes of operation for changing the resolution.  This 
> is easily controlled in the dialog you mentioned with the "Resample 
> Image" option.  Disable it, and Photoshop just changes the resolution, 
> which effectively changes the image size in real-world units (i.e. 
> centimeters, inch, etc) without changing any pixel data.  

Yes thanks for clarifying that, and that calculation seems to be 
straight forward.  Using a .bmp image I rendered in POV that is 800 
pixels wide, if I enter 300 pixels/inch in the box instead of 72, it 
automatically updates the document width box to read 800/300 or 2.667 
inches.


> hand, if you enable "Resample Image", Photoshop will interpolate (using 
> the specified method in the popup menu next to the resampling checkbox) 
> the pixel data maintaining the same image size in real-world units.

Yes, and it produces a new image (which
can be resaved) at the increased resolution needed to support the
increased ppi.  So my 800 pixel wide image is resized to 800*300/72 or
3333 pixels wide.  Whether or not this new higher resolution image
offers any truely new information can be debated. My
inclination is to say it does not.

But what is mysterious to me is the calculation that happens when
you click "Auto" on that dialog box with resample turned on.  That
allows you to set a "lines/inch" value which seems to default to 133
lines/inch.  What is opaque to me is two things.  What is the
significance of 133? And how is it used to produce the suggested
pixels/inch value?  For my original 800 pixel wide image, with resample
on, and using the default lpi value, the "auto" ppi value becomes 266 
and the new image resolution  becomes 2956 pixels wide.  I don't see the 
connection between these numbers.



> 
> About pixels:
> 
> A pixel is nothing more than an atomic unit of color information.  It 
> has no resolution whatsoever, it does not have a unit it is measured in.
> 
> The resolution is implicitly defined by taking a (usually rectangular) 
> block of many pixels, commonly called an image, which is a specified 
> number of pixels wide and high.  You then can define that the width of 
> the image, lets say it is 1000 pixels, shall be interpreted as being 2 
> real-world units in width.
> 
> A real world unit could for example be an inch.  In this case, there 
> would be 1000 pixels  to be fitted such that they are two inches in 
> total.  Thus, each pixel is interpreted to be 2 inch divided by 1000 
> pixel, which equals 500 pixels per inch.  You can just as well replace 
> the two by 200, and then  you get 5 pixels per inch.  Or you use some 
> other unit to map the pixel to.  The essence is, a pixel has no 
> dimensions, and thus no unit can measure its size.
> 
> In summary: A pixel is just a color information you can interpret to 
> have any width and any height you like.  And that is what is referred to 
> as resolution.
> 
Yes I believe I understand these concepts to the extent you have
explained them.

For if you have a printer that is *finally* going to produce the real
world image at the specified real world size, the ppi can tell it the
intended density of the *color information*.  Because how the printer 
translates this information into real world ink dots is still flexible. 
  It might use a ratio of 1 pixel = 1 dot, but it doesn't have to.  It 
could also use many dots to translate one pixel of color information, 
lets say 1 pixel = 20 dots.

If this were not true, it seems to me that very high resolution 
printers, or even printers produced for the consumer market which boast 
dpi resolutions of up to 5600, would only be usefull for producing very 
find grained images beyond what is necessary for the unaided eye.



Are these valid calculations?

A 6 magapixel camera could produce a picture in a 4x6 aspect ratio that
has a resolution of 2000x3000=6000000 pixels maximum.

This resolution could produce:
an image of 20 x 30 inches  at up to 100 ppi density
an image of 6.6 x 10 inches at up to 300 ppi density
an image of 4 x 6 inches    at up to 500 ppi density
an image of .4 x .6 inches  at up to 5000 ppi density

But when it comes to making the print, the upper limit also is affected 
by the "resolution" that the printer is capable of, expressed in dpi. 
So if a printer has a maximum resolution of 5000 dpi, it could produce 
the 4 x 6 image and use 10 dots per pixel to improve its own color 
handling, or it could produce a .4 x .6 inch image using 1 dot per pixel.

Post a reply to this message