8 Megapixel camera limited to 8cmx10cm prints?

Tell me what is wrong either with my calculations or my assumptions or
both.


If you have an 8 megapixel camera with a 4x5 ratio area then that
corresponds to 3150 pixels wide x 2520 high = 8 megapixels.

For the 3150 pixels wide the theoretical lp/mm is half that since a
bright dot and a darker dot is required to show a line so that gives a
theoretical maximum of 1575 line pairs.

The "system" resolution (i.e. achievable line pairs with a lens
focussing an image on it) will be about half that or 780 line pairs.

At standard viewing distance of 250mm then with good eyesight a person
can distinguish 8 lp/mm and if the image contains less detail than this
the sharpness will be degraded.

780/8 = 97.5 mm or 100mm for good measure. So those 780 achievable line
pairs will only look sharp at standard viewing distance if it covers no
more than 100mm. That is 10cm.

So the whole 4x5 frame will reach 8cmx10cm and if stretched more than
this it will not look sharp.



Now there just HAS to be things wrong with that calculation. There HAS
to be. An 8 megapixel camera can give great images that cover 8"x10" -
right? Perhaps even 11"x14"? So there must be one or more howling
errors in the calculations and assumptions above.

"RolandRB" wrote in message
oups.com...
Tell me what is wrong either with my calculations or my assumptions or
both.

If you have an 8 megapixel camera with a 4x5 ratio area then that
corresponds to 3150 pixels wide x 2520 high = 8 megapixels.

It is worse than you think. It takes three pixels to make a virtual point.

I've seen all sorts of math-based arguments about the ability of digital
cameras to make "sharp" prints. Some were intended to prove that digital
cameras were great, others were intended to prove that digital cameras
stunk. My preference is to look at the prints themselves. The prints I've
seen in the 13x17 range made by talented photographers with 8mp cameras were
very sharp though maybe math formulas say they couldn't be.

"RolandRB" wrote in message
oups.com...
Tell me what is wrong either with my calculations or my assumptions or
both.


If you have an 8 megapixel camera with a 4x5 ratio area then that
corresponds to 3150 pixels wide x 2520 high = 8 megapixels.

For the 3150 pixels wide the theoretical lp/mm is half that since a
bright dot and a darker dot is required to show a line so that gives a
theoretical maximum of 1575 line pairs.

The "system" resolution (i.e. achievable line pairs with a lens
focussing an image on it) will be about half that or 780 line pairs.

At standard viewing distance of 250mm then with good eyesight a person
can distinguish 8 lp/mm and if the image contains less detail than this
the sharpness will be degraded.

780/8 = 97.5 mm or 100mm for good measure. So those 780 achievable line
pairs will only look sharp at standard viewing distance if it covers no
more than 100mm. That is 10cm.

So the whole 4x5 frame will reach 8cmx10cm and if stretched more than
this it will not look sharp.



Now there just HAS to be things wrong with that calculation. There HAS
to be. An 8 megapixel camera can give great images that cover 8"x10" -
right? Perhaps even 11"x14"? So there must be one or more howling
errors in the calculations and assumptions above.

"Shelley" wrote in message
news:222Pd.19287$uc.7252@trnddc08...
I've seen all sorts of math-based arguments about the ability of digital
cameras to make "sharp" prints. Some were intended to prove that digital
cameras were great, others were intended to prove that digital cameras
stunk. My preference is to look at the prints themselves. The prints I've
seen in the 13x17 range made by talented photographers with 8mp cameras
were
very sharp though maybe math formulas say they couldn't be.

Indeed, and that's the very point. Unless we are doing recon work (and we
are not) it's not all about math, but the subject and perceived sharpness,
or alternately a 'clean' look. Digital images right off the camera (as
opposed to scans) can have a very clean look - no grain artifacts from
scanning (or over-rating the sensor). That said, I've not been a bit happy
with the outcomes of our Olympus 8mp cameras - in any way at all.

On 11 Feb 2005 00:33:11 -0800, "RolandRB"
wrote:


Now there just HAS to be things wrong with that calculation. There HAS
to be. An 8 megapixel camera can give great images that cover 8"x10" -
right? Perhaps even 11"x14"? So there must be one or more howling
errors in the calculations and assumptions above.


Why theorize? It's time for you to beg,
steal, borrow or buy one and find out
for yourself. It's not like digicams
are all that expensive anymore.

Just beware that the pixel count isn't
the final word. Believe it or not, some
pixels are worth more than others.


rafe b.
http://www.terrapinphoto.com

Recently, RolandRB posted:

Tell me what is wrong either with my calculations or my assumptions or
both.

The difficulty that I have with your assumptions -- not mathematical, but
the general direction of this kind of discussion -- is that they are
devoid of any consideration of the image. If all one intends to shoot are
test charts under ideal conditions, then the math is relevant and can
imply qualitative differences. If one intends to shoot real-world
subjects, then the math goes out the window along with all the "grain
sniffing" comments about film. Real-world images are about perception, and
there won't be a lot of consensus w/r/t what makes an image "good".

Digital cameras produce a particular "look", and are very good when used
to make certain kinds of images; film produces a different "look", and are
good for other kinds of images. But, at this point in time, there is no
"one look suits all" solution to image making, AFAIK.

Regards,

--
Neil Gould
--------------------------------------
Terra Tu AV - www.terratu.com
Technical Graphics & Media

Something which most calculation forget to consider is the printing side
of the equation. Even if you use a 16.7Mpx camera, the digital printer
in a commercial printing center usually does not print better than 300
dpi max. Some high end printers for larger formats print at 360dpi,some
at 400dpi. Many minilabs print at 300-360dpi at 4x6 and therefore their
resolution is less when they print at 8x10inch. This is especially the
case for those minilabs which use the TI DLP technology and no
pixelshift. And even with pixelshift technology, the absolute resolution
is not better, just the perceived one is better as the boundaries of the
pixels are washed out due to the movement of the imaging device between
the exposures made (usually three).
A way out of this dilemma is to use top notch inkjet printing for
digital files.
Oh-before I forget, do not think that laser based printers are the
solution, yes, they are better, but they usually deliver not more than
500dpi....
Hope this did not confuse you too much
rgds George



RolandRB wrote:
Tell me what is wrong either with my calculations or my assumptions or
both.


If you have an 8 megapixel camera with a 4x5 ratio area then that
corresponds to 3150 pixels wide x 2520 high = 8 megapixels.

For the 3150 pixels wide the theoretical lp/mm is half that since a
bright dot and a darker dot is required to show a line so that gives a
theoretical maximum of 1575 line pairs.

The "system" resolution (i.e. achievable line pairs with a lens
focussing an image on it) will be about half that or 780 line pairs.

At standard viewing distance of 250mm then with good eyesight a person
can distinguish 8 lp/mm and if the image contains less detail than this
the sharpness will be degraded.

780/8 = 97.5 mm or 100mm for good measure. So those 780 achievable line
pairs will only look sharp at standard viewing distance if it covers no
more than 100mm. That is 10cm.

So the whole 4x5 frame will reach 8cmx10cm and if stretched more than
this it will not look sharp.



Now there just HAS to be things wrong with that calculation. There HAS
to be. An 8 megapixel camera can give great images that cover 8"x10" -
right? Perhaps even 11"x14"? So there must be one or more howling
errors in the calculations and assumptions above.

In article .com,
RolandRB wrote:

Tell me what is wrong either with my calculations or my assumptions or
both.


If you have an 8 megapixel camera with a 4x5 ratio area then that
corresponds to 3150 pixels wide x 2520 high = 8 megapixels.

For the 3150 pixels wide the theoretical lp/mm is half that since a
bright dot and a darker dot is required to show a line so that gives a
theoretical maximum of 1575 line pairs.

The "system" resolution (i.e. achievable line pairs with a lens
focussing an image on it) will be about half that or 780 line pairs.

At standard viewing distance of 250mm then with good eyesight a person
can distinguish 8 lp/mm and if the image contains less detail than this
the sharpness will be degraded.

780/8 = 97.5 mm or 100mm for good measure. So those 780 achievable line
pairs will only look sharp at standard viewing distance if it covers no
more than 100mm. That is 10cm.

So the whole 4x5 frame will reach 8cmx10cm and if stretched more than
this it will not look sharp.



Now there just HAS to be things wrong with that calculation. There HAS
to be. An 8 megapixel camera can give great images that cover 8"x10" -
right? Perhaps even 11"x14"? So there must be one or more howling
errors in the calculations and assumptions above.

Throw away your calculator and go take pictures.
I have 16x20s that look great from my10D.

"rafe bustin" wrote in message
...

Just beware that the pixel count isn't
the final word. Believe it or not, some
pixels are worth more than others.

The Green channel is my favorite.

"Dr. Georg N.Nyman" wrote in message
...
Something which most calculation forget to consider is the printing side
of the equation. Even if you use a 16.7Mpx camera, the digital printer in
a commercial printing center usually does not print better than 300 dpi
max.

Really? I have a 600ppi laser printer right here.