.8 to 8mp experiment

Olympus D-500L (.85Mp)
Kodak DX4330 (3.1Mp)
Minota F100 (4Mp)
EOS Rebel (6.3Mp)
Cannon 20D (8.2Mp)

I took the same shot with each of these cameras;
noon, high overcast. The camera setup was done by
its owner. He was asked to set it to the best
quaulity jpg. Each was printed at Walmart's
"one hour" machine at 6x4 (24 cents each - next
size up was over 3X as expensive).

Then I showed these prints, left them in the lunch room
and asked everyone to guess which was which and what they
liked best. Guess what?
Printed at this size, 6x4, there aint much more than
a dimes difference between them (D-500 excepted). The D-500 was
the worst - and everyone saw it. Surprising, but
most liked the DX4330 and F100 shots the best.
I thought the F100 beat it.

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I conclude that chasing pixel count in these lower
regions isn't worth much to a recreational user OR
the Walmart printer really sucks OR the owners of the
cameras can't set them up.

What's a visual decibel? Do we need 10 itimes the pixel count
to make a real difference?

If you want copies of the files, post your address. You
judge.

Hi!

Your questions are right on if I base my judgement on what I read some time
ago.

About 2 months ago, I was looking at photos on Mars taken by NASA
(http://photojournal.jpl.nasa.gov/targetFamily/Mars).
There was an article which I can't find anymore, saying that the camera was
very special. If I remember correctly (I'm not a pro), the camera had a mere
1 million pixels. Where it differed from our cameras was the quality of the
lense as well as the the CCD which is larger. The upshot of it all was
"millions of pixels are not all it takes".

Cheers,

Marcel


"hfs2" wrote in message
om...
Olympus D-500L (.85Mp)
Kodak DX4330 (3.1Mp)
Minota F100 (4Mp)
EOS Rebel (6.3Mp)
Cannon 20D (8.2Mp)

I took the same shot with each of these cameras;
noon, high overcast. The camera setup was done by
its owner. He was asked to set it to the best
quaulity jpg. Each was printed at Walmart's
"one hour" machine at 6x4 (24 cents each - next
size up was over 3X as expensive).

Then I showed these prints, left them in the lunch room
and asked everyone to guess which was which and what they
liked best. Guess what?
Printed at this size, 6x4, there aint much more than
a dimes difference between them (D-500 excepted). The D-500 was
the worst - and everyone saw it. Surprising, but
most liked the DX4330 and F100 shots the best.
I thought the F100 beat it.

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I conclude that chasing pixel count in these lower
regions isn't worth much to a recreational user OR
the Walmart printer really sucks OR the owners of the
cameras can't set them up.

What's a visual decibel? Do we need 10 itimes the pixel count
to make a real difference?

If you want copies of the files, post your address. You
judge.

Although comparing it against film isn't fair. Film sensitometry is such
that consumer films will tend to produce more contrast and colour saturation
than a digital camera. The digital camera is designed to record the scene
as accurately as possible, while a consumer film is more about producing,
sharp, impact images

"Jim Townsend" wrote in message
...
hfs2 wrote:

Cannon 20D (8.2Mp)
Do you honestly believe people are lined up to spend
$1000+ on cameras like the Cannon (sic) 20D when
the images aren't "anywere near the quality" of film ?

Maybe that's his problem -- using a Cannon instead of a Canon ?

(hfs2) wrote in
om:

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I had 8x10 prints made from my 5Mp camera, and I judged them to be much
better than any prints I ever had made from ASA 100 film. My conclusion is
that the photolabs that printed my negatives had crummy equipment (probably
lenses fogged).

But I agree with your general conclusion. If all you make are 6x4 prints,
then the camera doesn't matter a whole lot.

Bob

(hfs2) wrote in
om:

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I had 8x10 prints made from my 5Mp camera, and I judged them to be much
better than any prints I ever had made from ASA 100 film. My conclusion is
that the photolabs that printed my negatives had crummy equipment (probably
lenses fogged).

But I agree with your general conclusion. If all you make are 6x4 prints,
then the camera doesn't matter a whole lot.

Bob

About 2 months ago, I was looking at photos on Mars taken by NASA
(http://photojournal.jpl.nasa.gov/targetFamily/Mars).
There was an article which I can't find anymore, saying that the camera
was
very special. If I remember correctly (I'm not a pro), the camera had a
mere
1 million pixels. Where it differed from our cameras was the quality of
the
lense as well as the the CCD which is larger. The upshot of it all was
"millions of pixels are not all it takes".

Cheers,

Marcel

Yes, but those pictures you see from Mars are actually panoramas made up of
many of those 1 megapixel images. Thus, the one photo you see could
actually be a 20 megapixel photo equivalent.

Thanks for the experiment. It is completely expected from a "theoretical"
point of view. Neither the printing paper nor the normal eye can
resolve more than 300 ppi. In fact if you include a 2 Mp picture (267 ppi),
probably 90% of viewers cannot tell the difference. However, an 8.2 Mp
picture can be cropped to 1/4 size still producing the same decent
image while the 3 or 4 Mp cannot. It is alway nice to verify results
by experiments. Thanks again for the good work.


In article ,
hfs2 wrote:
Olympus D-500L (.85Mp)
Kodak DX4330 (3.1Mp)
Minota F100 (4Mp)
EOS Rebel (6.3Mp)
Cannon 20D (8.2Mp)

I took the same shot with each of these cameras;
noon, high overcast. The camera setup was done by
its owner. He was asked to set it to the best
quaulity jpg. Each was printed at Walmart's
"one hour" machine at 6x4 (24 cents each - next
size up was over 3X as expensive).

Then I showed these prints, left them in the lunch room
and asked everyone to guess which was which and what they
liked best. Guess what?
Printed at this size, 6x4, there aint much more than
a dimes difference between them (D-500 excepted). The D-500 was
the worst - and everyone saw it. Surprising, but
most liked the DX4330 and F100 shots the best.
I thought the F100 beat it.

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I conclude that chasing pixel count in these lower
regions isn't worth much to a recreational user OR
the Walmart printer really sucks OR the owners of the
cameras can't set them up.

What's a visual decibel? Do we need 10 itimes the pixel count
to make a real difference?

If you want copies of the files, post your address. You
judge.

FWIW & IMHO only

I have never had quite as good quality prints from any film from any camera
as I get from my humble Fuji (2MP interpolated to 4MP)

I have had a few prints made at 15" by 10" and the quality is amazing.

While I can't speak for the general public, and have no intention to, the
most common forms (modal forms I suppose in tech speak) are for 2 to 4 MP
cameras with output to 6 x 4. Somehow I think sales would not be quite as
high if digital prints on 6 by 4 were worse than film based prints.

Also, on a 6 by 4 the size of the media limits (IMHO) perception to the main
event happening in the image. Fine details fade at that size whether it be
digital or film based. However at 15 by 10 the intricacies of wider details
start to have presence.

Digital image output tends to be optimised for a particular medium (print or
screen) and the strength (IMHO) of 8MP is the quality of data captured.

Conclusion?
Try the same exercise again this time print onto 15 x 10 OR 18 by 12 OR
both. Then do the coffee table comparison again

If you do that last bit please keep us posted...

Is it important?
Try this:
1 - define a 50% crop on the original image, then a 25% then a 10%. (same
areas on the original image in each case with the percentage based on 6 x 4
print rather than pixel count.

2 - print each of these on to 6 by 4

3 - do the coffee table test again

4 - list results here

Aerticus

It's not surprising that with a small print the lower pixel count could look
sharper, w/o doing anything with the image before printing. (Depending on
what the wal-mart printer/software acually does.)

Printing mostly 8x10 and 11x17 prints on my Canon 9100, I get better results
with my 5MP (resampled 600 ppi Mitchell) digitals than I do with film. (At
least with today's consumer developing and printing - maybe if they were
printed with more care...) And it's certainly orders of magnitude quicker
and less messy...

And I'm not even bothering with RAW, color profiles, or spending inordinate
amounts of time removing noise or sharpening (though I plan to do someof
this in the future) - usually just an auto exposure correct, and a click of
the color-cast wand in the right place.

mike

"Charlie Ih" wrote in message
...
Thanks for the experiment. It is completely expected from a "theoretical"
point of view. Neither the printing paper nor the normal eye can
resolve more than 300 ppi. In fact if you include a 2 Mp picture (267
ppi),
probably 90% of viewers cannot tell the difference. However, an 8.2 Mp
picture can be cropped to 1/4 size still producing the same decent
image while the 3 or 4 Mp cannot. It is alway nice to verify results
by experiments. Thanks again for the good work.


In article ,
hfs2 wrote:
Olympus D-500L (.85Mp)
Kodak DX4330 (3.1Mp)
Minota F100 (4Mp)
EOS Rebel (6.3Mp)
Cannon 20D (8.2Mp)

I took the same shot with each of these cameras;
noon, high overcast. The camera setup was done by
its owner. He was asked to set it to the best
quaulity jpg. Each was printed at Walmart's
"one hour" machine at 6x4 (24 cents each - next
size up was over 3X as expensive).

Then I showed these prints, left them in the lunch room
and asked everyone to guess which was which and what they
liked best. Guess what?
Printed at this size, 6x4, there aint much more than
a dimes difference between them (D-500 excepted). The D-500 was
the worst - and everyone saw it. Surprising, but
most liked the DX4330 and F100 shots the best.
I thought the F100 beat it.

None of these pictures were anywhere near the
quality of a good 6x4 200 asa film shot.

I conclude that chasing pixel count in these lower
regions isn't worth much to a recreational user OR
the Walmart printer really sucks OR the owners of the
cameras can't set them up.

What's a visual decibel? Do we need 10 itimes the pixel count
to make a real difference?

If you want copies of the files, post your address. You
judge.