6MP Nikon D100 - good enough for critical 24" x 42" advertising posters?

X-No-Archive: Yes

I tried asking this question elsewhere but the answers seemed to be in
one direction.

I am one of those old mares who has stuck clear of digital, apart from
a convenience 8MP small-sensor non-SLR P&S which is fine for what I use
it for.

A friend of mine uses a Nikon D100 and a couple of AF-D Nikkor zooms.
He has asked my opinion as to whether that is suitable equipment for an
assignment to photograph some objects for ADVERTISING POSTERS. The
poster sizes will be 24 inches x 42 inches (2 ft x 3.5 ft). The subject
matter will fill almost all of the area of the poster.

The posters will typically be viewed from a distance of a foot, in shop
windows and on display stands.

While I can't tell you what the subject is, pin sharpness of the end
result is critical for the assignment. Sharpness is an attribute of the
very thing being advertised. The subject cannot be photographed in
sections which can then be stitched together.

The D100 is a 6 megapixel camera, making 3008 x 2000 pixel images in
its highest resolution mode. It is certainly better than my 8MP P&S due
to the larger sensor size / better noise, superior lens(es) e.t.c.

But I just can't see how the D100 can cut it to reach 24" x 42". Even
assuming that he fills the frame pretty well with the subject so there
is only a 20% loss through cropping (given the nature and proportions
of the subject, that would be excellent), that would mean he would need
to fill 42" with 80% x 3008 = 2406 pixels. That means the "ppi
resolution" would be no more than 2406/42 ppi = 57 ppi, or just over 2
pixels per mm.

Sharpening or interpolating (creating pseudo-detail) software won't
help much, given the subject.

I think that 57 ppi is far too little, given the likely viewing
distances. The general wisdom is that 300 ppi is best for a hand-held
8" x 10" and 150 ppi may be acceptable for critical work with a viewing
distance of over a foot. But 57 ppi is nowhere close to 150 ppi, and I
think the image will look quite pixellated or dotty and, when
interpolated for printing and the end result viewed as a poster,
unsharp.

Given that the advertising posters are aimed at an audience who will
come up to them and look at them closely (some from closer than a
foot), can 57 ppi be enough?

Oh I know ppi and dpi are not the same.

Please tell me your views. Will the D100 cut it for the assignment? I
think a modern 11+ MP SLR (i.e. Canon EOS-1Ds) just might be OK but it
might be pushing it Certainly if I could stretch to a digital back for
my Hassy it could. My instincts (more reliable than my maths, which
isn't too good) would be to shoot it on Kodachrome 25 (I have a brick
holed away in the fridge) or Velvia 50 (grain being less important than
sharpness for this subject - a little grain may even be flattering) or
similar, and have it drum scanned for good measure. I'd use a 6x6 on a
'pod. Just maybe a 35mm with a good prime lens too.

If I am right, how do I diplomatically tell him this? I value his
friendship and he does not take criticism well. The assignment means a
lot to him, and I don't mind loaning him my analogue gear (a modern
Hassy and/or an almost brand new F6).

Or is my friend right, and my maths wrong as usual, and the D100 will
do the job fine?

 

Given those constraints (viewing distance a foot or less, 57 ppi and
interpolation not being practical), the answer is clearly "no".

A person with normal eyesight will that at the distance of one foot be
able resolve detail less than 0.2 mm (200 microns). The pixels in
your friends photograph will be more than twice, 0.45 mm or 445
microns, so the pixelation in the image will be visible to the
observer.

He might barely pull it off with a Canon 1Ds2 (118 ppi) which will
give him pixels of about 0.2 mm (214 microns to be exact) - but
even that is stretching things. I would say that this is the type
of assigment that calls for at least MF (film or digital back).

That much being said, one foot is not a /normal/ viewing distance
for a poster 24 x 42 inches. People will not able to see all of
it being that close. If people stand further away from the poster,
you can get away with lower resolutions.

 

A recent article in Outdoor Photographer discussed making a 19x28 print
from a D100 image. I have no idea if the process followed would work
to a larger size, but it did involve considerable interpolation.

http://www.outdoorphotographer.com/content/2004/sept/poster.html

Paul B.

 

He's dreaming, and I doubt the agency will be too impressed with his
approach. For poster work you can get away with 150-180ppi, but any lower
and you've really got to be confident about the application and the printing
technique. It sounds like your mate has spent too long on digital - looking
at the back of his camera after every shot rather than trusting his exposure
decisions.

For that size work, actually, for any kind of pro poster work, you'd have to
be shooting medium format, unless you've coughed for some insanely expensive
digital gear.

Give him a good slapping, lend him your camera, a light meter and a Polaroid
back.

 

I doubt if people are going to be viewing a poster at 1 foot. But if
they tried, even forgetting the digital aspect, you would have a hard
time getting enough resolution for a critically sharp image that size.
The image area of the sensor array is just under 16 x 24 mm. A 24 x 42
inch image is just under 610 x 1067 mm. That requires an enlargment
factor of about 38. If you assume that the human eye can resolve about
5 lp/mm at 1 foot---which is a low end estimate---then at the image
plane, you would have to be resolving about 190 lp/mm. It seems
doubtful that the lenses used with the D100 can do that well. Of
course, taking the digital aspect inot account, the sensor array
certainly can't. You get about 125 pixels per mm, and you can resolve
at best half that or 62.5 lp/mm.

 

In message <>,

Why doesn't he just take a 57 PPI crop of a sharp image he has and print
it on his printer at highest quality and see what it looks like? You
don't have to waste paper and ink printing an entire image to get an
idea of quality from a certain distance.
--

 

Interpolation can indeed work well with certain subjects - but
unfortunately not all.

The tree frog image that is discussed in this article, consisting
of large areas of strong colour, looks like a subject that will
work well with interpolation.

The OP stated that interpolation was not an option. I assume that
this means that the subject has a lot of high-frequency detail.

There is one serious error in the article in Outdoor Photographer
you link to. The author states: "Interpolation is a process where
the computer software adds information to and rebuilds the original
image..."

But computers can not create information out of thin air.
Interpolation is a mathematical technique to construct new
intermediate data points from an existing set of known data points.
No new information is added to the image, just intermediate pixels
based on the pixels already present.

For those interested in this topic, I've set up a page where I
compare some popular interpolation methods:
http://folk.uio.no/gisle/photo/interpolation.html

 

A James Russell shot from a Fuji S2 once sat on one of those huge billboards
in Times Square. I've had a Canon 10D image used for 60x40 inch billboards
on eye level display in malls (and it had been cropped by approx 20% too).
Like your friends potential commision these posters could be viewed very
closely by shoppers. If the subject requires great detail - something like
an architectural showpiece - then 6MP's not even close. From what you have
described this may well be the case. He could hire a Nikon D2X (or even a
1DS mkII and a lens). But for many subjects and uses, including most that
feature people, 6MP will really be fine at 24"x42" on a poster if processed
accordingly. If he does try it with a D100 then advise that he applies no
sharpening at all. Sharpening before resizing is what will create the
aliasing & pixelization you're concerned about. Capture in raw and process
in ACR if possible. In-cam JPEGs and some raw converters sharpen even when
set to minimum or no sharpening. Resizing from a clean unsharpened 6MP
capture can be taken to extreme sizes if necessary. The 'laws' of acceptable
inkjet resolutions as well as shaprening thechniques really do not apply for
posters

 

Quite true.

The bit of the article that you quote doesn't suggest that the information
comes from nowhere. It does not mislead. It certainly does add information -
information based on that in surrounding data points.

You may think that you can explain the matter more clearly or in a different
way - but that is different from the original being an error.

 

Asking in r.p.e.l-f? The answer is -no-. For 24x42 output 4x5"
is the _minimum_ negative size required to do a professional advertising
job. Most photographers would (should) use an 8x10 or 11x14.

24x42" from a 4x5" negative is equivalent to 11x14" from 35mm -
quite a bit farther than one can go and claim commercial quality.

When one is squashed in a subway station one foot is getting pretty far away
from the poster ... If you are -selling- something it has
to look good at all distances.

 

NO!

 

It's similar to how error correction works in DVD and CD players where
there is a "read ahead" function that fills in gaps with similar
"notes" to what is around them. Most people won't notice the
difference. But some of the makers of software designed for enlarging
are to blame for a perception that they magically "add" new data,
which as you pointed out they do not. One other thing; Sometimes,
digital zooming (in non-DSLRs) does a slightly better job of enlarging
an image than does bicubic enlargement. I know this sounds like
heresy, but people should experiment with many enlargement methods,
since most programs offer different modes built-in.
-Rich

 

This only really matters when someone is trying to portray the most
excellent quality possible. You see it primarily in magazines where
images are obviously much smaller than 24x42. Another thing; Take a
look at the huge posters produced for theatrical display. If it's a
monster 12 feet long, you walk up to it and you'll see that those
"benzene ring" colour elements used to produce the picture can be
almost 1/4" across!!!
-Rich

 

What is added is not information, it is data.

We disagree on semantics.

The word "information" have a specific meaning (grounded in
mathematics and in the extraordinary work by Claude E. Shannon
which is known today as "information theory"). I hate seeing it
misused as a synononym for "data", because such usages lead to
misunderstandings.

It may well be that such a rigorous definition of the word
"information" is pedantic and old-fashioned. But we are living
in an age where most things branded as "information" isn't.
I therefore think the defintion of "information" is important,
and try to point out what I view as misuses.

 

It does not add 'information' - it fills out pixel values but that is
is not the same as adding information. Your problem stems from not
knowing what the word 'information' actually means (hint: its certainly
not the same as data).

If you have three boxes with two kinds of vegetable and fill the third
with a ground-up mix of the two, do you claim that there are three kinds
of vegetable in three boxes, the third of which is a interpolated
vegetable? You still have just two vegetables. Its the same with
picels and their information content.

 

I'll accept that criticism. I've not seen the quibble outside information
theory texts so generally follow the 'general purpose dictionary' route.

 

Interesting analogy.

A similar argument could be advanced about the combination of signals
from the single colour pixels of Bayer pattern sensors to interpolate
an image with full colour at each final pixel position.

 

A 24 x 42 inch poster that will be viewed from a foot? Are they going to
have some kind of ladder and scaffold available so people can eventually see
the whole poster?

 

X-No-Archive: Yes

Hi Bob,

Could you have got "feet" and "inches" mixed up, Bob? I am feeling
better about my maths already.

A 24 inch x 42 inch poster is not very big - it is just 2 ft x 3.5 ft.,
not untypical for shop windows.

If you need a ladder and scaffold to see the whole of it close-up, then
you must be very, very, short, and probably not representative of the
target audience of potential customers, most of whom are going to be at
least 3 feet high.

 

No, the field of vision at one foot will be about 8" x 11", give or take.
You cannot see a whole 24 x 42 inch poster from one foot. You cannot even
scan your eyes over that large an area from 1 foot and maintain focus. Try
it. Hold an 8 x 10 picture 1 foot in front of your face. Move the image 1
foot to the right, or left without moving your head and see if you can focus
on it. You will have to move your body, or head from side to side and up and
down to eventually see the whole poster, something people will not do. If
it's for advertising, of any kind, people will not work to read it, and they
will have to work at one foot.

While you may not need a ladder and scaffold, I really doubt that folks will
be viewing from one foot, and if they do they will give up quick trying to
figure out what it is they are looking at. Unless of course the sidewalk in
front of the store is only 18" wide.

Try another test. Print a sample poster at 24 x 42. Put it in a window and
see how many people walk up to 1 foot from the window to view it. Most will
stand back at about 4 feet, maybe more, unless they are restrained from
doing so.