How many pixels?

Jerry R

Forum Pro
Messages
12,161
Reaction score
2,337
Location
Venice Court, Naples, FL, US
I just did a calculation for a course I am teaching next quarter on the number of pixels required to represent an 8"x11" pixture and a 4"x6" picture with no visable degradation to the eye when viewed at a distance of 12". Care to guess?

Canon A2E, Sony R1 & Panasonic TZ5
 
I just did a calculation for a course I am teaching next quarter on
the number of pixels required to represent an 8"x11" pixture and a
4"x6" picture with no visable degradation to the eye when viewed at a
distance of 12". Care to guess?

Canon A2E, Sony R1 & Panasonic TZ5
LOL Jerry - that sounds like a loaded question!! Wouldn't a lot depend on the photo's dynamic range, DOF, detail content and ... if I had my glasses on??

Seriously - I have no idea, but I do have a question. Do you think it would be the same for an image viewed on a screen vs an image to be sent for printing an 8 x 11 or 4 x 6 photo?

BG
--

Always open to learn something new - feel free to disagree with me any time as long as you can do it politely!!
 
But I'd feel fairly confident that 5mp would do fine for 8x10, and 2mp for the 4x6 print.

At least, that's based on having made lots of prints from those two sizes!

I've got a few 8x10s from a 2mp camera, NOT upsampled with any special software...and it looks fine from 2 feet. Within 1 foot you'll see some pixel separation, but even that one is sufficient for normal viewing on a wall, and I have yet to have anyone notice that 2mp print standing out as notably worse than my 5, 7, or 10mp prints at the same size.

To the average joe at the normal viewing distance...even 2mp can pass muster. To the camera enthusiasts, you might need to go up to 4mp at normal viewing distances, and 5mp for the close-up scrutiny.

So...what did your research tell you?

--
Justin
galleries: http://www.pbase.com/zackiedawg
 
Now why must you complicate things at this time of the day ? LOL. My brain can not get round the calculation :-).

Regards.

ernest
--





May the Focus be with you
 
Most people are using computer screens with 1 megapixel resolution. If you look close at a white area of the screen you can see the pixels.
--
Canon A2E, Sony R1 & Panasonic TZ5
 
There are 120 cone cells per 290 microns concentrated at the central portion of the retina. This gives rise to a resolution of 60 line pairs (120 pixels) per degree of viewingle angle. An 8.5" x 11" piece of paper subtends a 50x40 degree viewing angle at 12". Therefore the number of pixels is (50deg. x 120 pixels x 40deg. x 120 pixels), about 30 megapixels. for a 4"x 6" picture about 9 megapixels are required to match the resolving power of the eye at 12". Now I have gotten excelent 4x6 and 8x10 prints from 1 meg images (prints are upscaled). This calculation is just the limits of the eyes resolution (20/20 vision).
--
Canon A2E, Sony R1 & Panasonic TZ5
 
Your calculations do not jibe with the information from this source:

http://www.ndt-ed.org/EducationResources/CommunityCollege/PenetrantTest/Introduction/visualacuity.htm

Their formula: X/2 = d (tan Theta/2)

Where: X = normal acuity in inches
d = viewing distance
Theta = 1/60 of a degree, i.e,. normal acuity in degrees

Yields an X value of .00349" for a 12" viewing distance.
That translates into 286.5 pixels per inch. Solving for the pixels
required for an 8.5" x 11" (8.5 x 286.5) x (11 x 286.5) image
yields 7,661,296.5 pixels or a bit more than 7.6 megapixels for
a 12' viewing distance.

Or am I missing something?

--
When a hammer is your only tool, all problems begin to look like nails.
 
not only does that not jibe with the formula and the results from the
source I cited, but it also doesn't fit with my own experience in printing
many pictures at various sizes and resolutions over the past 10 years.
Your numbers are just way too high.

--
When a hammer is your only tool, all problems begin to look like nails.
 
The article you site is speaking about line pairs for the.

"What this means is that if you had alternating black and white lines that were all 0.00349 inch wide, it would appear to most people as a mass of solid gray."

Two pixels are required to produce a line pair. Take the 7 megpixels and multiply it by 4 (2 x per axis) to get the 28 megapixels. I am not claiming that 30 megapixel picture is required to print an acceptable 8.5 x 11 image, this is just the limitation of the eyes resolution.
--
Canon A2E, Sony R1 & Panasonic TZ5
 
Yes, the source I cited does talk about line pairs, made up of lines
that are each 0.00349 inch wide , which is equivalent to 286.5 pixels
per inch. The total width of each line pair is 2 x .00349 or .00698.

Your reasoning would mean that the average human eye can resolve
up to approximately 573 pixels. That doesn't stand up to reality.

Try this experiment, view a camera resolution test card at a distance of
12 inches under good lighting to see how well your eyes can resolve the
lines on it. I did that using my reading glasses, which give me 20/20
vision at that distance, and I could barely discern the line pairs at the
highest density - which on the card I have is about 100 line pairs per
inch - i.e., equivalent to about 200 pixels per inch. That makes it quite
difficult for me to accept that my eyes could resolve line pairs that are
more that 2 1/2 times denser.

--
When a hammer is your only tool, all problems begin to look like nails.
 
Jerry, this is a nice bit of math you have done.

Given the fact that many of us including me will eventually be purchasing a new (Sony?) wide screen TV, the major question is what size screen should I choose for a given viewing distance?

We sit approximately thirteen feet from the TV. What would you or another knowledgeable forum participant recon would be the minimum screen size and the maximum screen size for pleasurable viewing?

TIA-
-doc
 
My wife makes these major decisions. For a 13' viewing distance 42" would be on the small size, 50" would be about right for 13 feet.
--
Canon A2E, Sony R1 & Panasonic TZ5
 
2nd generation laser printers printed at 600 dpi, 1st generation ones
printed at 300 dpi and you still could not see the individual dots even
then. But, that is not much of a proof of anything because with printers
like lasers and inkjets dot size and dot density both have to be taken into
account when considering their print quality. They are designed so that
the dots overlap to give smooth shapes to letters and graphics.

As for it "being the math", when the math yields results that are at
odds, and significantly so in this case, with observation, then its time
to go back and look at your math again.

--
When a hammer is your only tool, all problems begin to look like nails.
 
The right amount of pixels that is required to print a photo is the result of how beautiful the photo looks on the wall from a normal viewing distance and the number of wows it gets from those who watch it. This is the only correct practical answer, all the rest, excuse me for the term, is pure masturbation.

Cheers
Moti

p.s. what is pixels?
 

Keyboard shortcuts

Back
Top