Can You Alter DPI?
Jules Vide
found on Microsoft's Clip Art site for a project. It's horizontal and
vertical resolution is 150 dpi, but the publisher I'm working with
requires 300 dpi.
(Do pixel width and height have anything to do with dpi? I'm sure this
is a stupid question, but I've searched Google and other sites and have
come up empty. The particular pixel width of this photograph I'm
interested in is 600 and the height is an odd 402.)
Thank you for posts or email.
Jim
"Jules Vide" <passep...@yahoo.com> wrote in message
news:1144523088....@j33g2000cwa.googlegroups.com...
publisher requires. Be certain not to resample the image because that may
not help its appearantce.
You can use dpi together with pixel width to determine the print width and
dpi together with pixel height to determine the print height. The
fundamental dimensions of the image are those given by pixel height and
pixel width. Dpi is merely a scaling factor.
Jim
Jem Raid
"Jim" <j...@nospam.com> wrote in message
news:kXTZf.56525$F_3....@newssvr29.news.prodigy.net...
I suspect that you will have to resample it, the size at 300dpi works out to
2" (50mm) x 1.34" (34mm) I wouldn't think a publisher would want an image
this small. There is a resize function in Photoshop, in the Help section
oddly enough, you could give it a try.
Jem
-------------------------------------
Birmingham Independent Photographers
http://bip.wikispaces.com/
Gisle Hannemyr
> I have a HP 4.1 MP digital camera but would like to use a photograph
> I found on Microsoft's Clip Art site for a project. It's horizontal
> and vertical resolution is 150 dpi, but the publisher I'm working
> with requires 300 dpi.
>
> (Do pixel width and height have anything to do with dpi?
No. A pixel has no size.
> I'm sure this is a stupid question,
No, it is not. But it is a frequently asked question, so I've
written a short FAQ to answer it.
> but I've searched Google and other sites and have come up empty.
> The particular pixel width of this photograph I'm interested in is
> 600 and the height is an odd 402.)
Take a look at my FAQ on the subject:
http://hannemyr.com/photo/pixels.html
The specific questions you ask is Q4, Q15 and Q16.
Btw.: 600 x 402 pixels at 300 ppi will be fine, provided you don't
want the print to be bigger than 2 x 1.34 inches. If you want to
print bigger, you need to resample, which is also covered (Q14 and
Q16) in in this particular FAQ.
--
- gisle hannemyr [ gisle{at}hannemyr.no - http://hannemyr.com/photo/ ]
------------------------------------------------------------------------
Sigma SD10, Kodak DCS460, Canon Powershot G5, Olympus 2020Z
------------------------------------------------------------------------
Gisle Hannemyr
> I have a HP 4.1 MP digital camera but would like to use a photograph
> I found on Microsoft's Clip Art site for a project. It's horizontal
> and vertical resolution is 150 dpi, but the publisher I'm working
> with requires 300 dpi.
>
> (Do pixel width and height have anything to do with dpi?
Sort of. A pixel has no intrinsic size, but when printed,
each pixel in the photograph obviously fills some space on
the paper, and pixels per inch (ppi - by some also called
dots per inch - dpi) is a measure of that.
> I'm sure this is a stupid question,
No, it is not. But it is a frequently asked question, so I've
written a short FAQ to answer it.
> but I've searched Google and other sites and have come up empty.
> The particular pixel width of this photograph I'm interested in is
> 600 and the height is an odd 402.)
Take a look at my FAQ on the subject:
kctan
from conventional photography to digital photography. The confusion starts
from the creation of the dots that virtually form the image and the dots
that prints the image. It is very important to note that they are totally
different "animals". The dots that form the image are called pixels, so it
will be more appropriately to describe the resolution of the formed image in
"pixel per inch" (PPI) rather than "dot per inch" (DPI). PPI is square and
DPI can be in any shape but usually in round shape. To be precise, image's
pixels have no shape and no size but just electrical signals that represent
luminance and colors when stored in the form of a digital file. As we need
to see this image visually on the screen, the electrical signals will
transform into squares and join them up in a specific sequence to form a
picture. So the shape is square but how about size? It is the PPI, pixel per
inch. 72 PPI or DPI means 1/72 inch square in every pixel. 150 PPI = 1/150"
per pixel and so on...What I'm trying to emphasize is you can alter the DPI
or PPI but not the total pixels in order to retain the physical quality of
the image theoretically. If you follow me, you will see that (by mental
calculation) when you alter the PPI or DPI, the size of the picture alters
too. This mean that no interpolation or resample of pixel is taking place.
Then someone says I need to have 300 DPI for the print but yours is only
150DPI...well alter it to 300 DPI but hey, the size is too small. This is
where resample come in but that will alter the quality. Yes it does
theoretically but not visually if you do it appropriately. Then why 300 DPI
for print? Because printing DPI has physical shape and size unlike PPI. The
needed image's PPI or DPI is its printing "capability" but not its
availability like 1440 DPI or 2880 DPI. As I said they are different
"animals". It varies from 200 PPI or DPI to 300 PPI or DPI for good print
output in the industry.
It is easier to deal with digital image in pixel rather than cm or inch
because of its behaviour but not easy for beginner but life will be much
better when you learn to do it. I'm sure many have experienced the scrolling
of screen to view large image. Using a program to shrink it too much to fit
on the screen for easy viewing may upset the viewing quality. This is
because the amount of pixels in the image have exceeded the screen assigned
ones (example: 640x480, 800x600, 1024x768... the screen resolutions) by too
much. So create your image to match your screen resolution for quality
viewing if you are not doing printing. You could create the image at the
highest possible resolution available to you and then up or down resample it
to your requirement later with a good resampling program but always keep the
original one untouched by using "Save As" command to save the resampled one.
"Jules Vide" <passep...@yahoo.com> wrote in message
news:1144523088....@j33g2000cwa.googlegroups.com...
Jules Vide
>
> Sort of. A pixel has no intrinsic size, but when printed,
> each pixel in the photograph obviously fills some space on
> the paper, and pixels per inch (ppi - by some also called
> dots per inch - dpi) is a measure of that.
>
> written a short FAQ to answer it.
>
> Take a look at my FAQ on the subject:
> http://hannemyr.com/photo/pixels.html
>
> The specific questions you ask is Q4, Q15 and Q16.
>
> Btw.: 600 x 402 pixels at 300 ppi will be fine, provided you don't
> want the print to be bigger than 2 x 1.34 inches. If you want to
> print bigger, you need to resample, which is also covered (Q14 and
> Q16) in in this particular FAQ.
Gisle (Jem, Jim, and kctan): Thank you. This subject is sufficiently
difficult for laypeople that a tech support rep at Hewlett Packard
admitted to being as perplexed as I am. Gisles, I will definitely
proceed immediately to your site and read the questions.
The HP rep suggested I purchase a program called PhotoPlus. In fact,
this software was available for free download, and so far--as a
layperson--I've found it head's-above all other software I've used thus
far.
I have a much more prosaic question, however. The only reason (or at
least the precipitation reason) I need to learn more about the patois
of digital photography is because I'm creating a cover of a book. If
(as Jem said), I will be able to "blow up" the 300 dpi image to only
such a small dimension, then either I'm interpreting the PhotoPlus
"print preview" of this homemade book cover very incorrectly, or else
the software is so good it actually enlarged the image so that it fits
the 6" width.
If I uploaded it to Photobucket or another site, would anyone be good
enough to visit it and tell me if I'm misinterpreting what it will look
like as a finished product? Of course I'll take it to my local Office
Depot, Staples, etc., and get a mock-up, but I'm having problems with
"blowing up" the 150 pixels that results in the (black) background
layer of the image appearing as a kind of vertical stripe in the cover
gutter.
Apologies for this abstruse and personal post, but I've learned quite a
deal from having made it; and for that I thank you all, Gisle in
particular.
Gene Palmiter
question. If the image is printed at this size you will have this many
pixels per inch.
If you are printing a large number of books you will want to print them with
offset lithography. This process uses halftones and you will want at least 2
picture elements (pixels) for each line of dots.
Really though...your printer or designer should be doing this....lots of
them don't quite get it.
It looks to me like you can't get what you want if what you want is a good
sharp photo on glossy paper.
--
Thanks,
Gene Palmiter
(visit my photo gallery at http://palmiter.dotphoto.com)
freebridge design group
"Jules Vide" <passep...@yahoo.com> wrote in message
news:1144523088....@j33g2000cwa.googlegroups.com...
Gisle Hannemyr
> Gisle Hannemyr wrote:
>> Btw.: 600 x 402 pixels at 300 ppi will be fine, provided you don't
>> want the print to be bigger than 2 x 1.34 inches. If you want to
>> print bigger, you need to resample, which is also covered (Q14 and
>> Q16) in in this particular FAQ.
> I have a much more prosaic question, however. The only reason (or
> at least the precipitation reason) I need to learn more about the
> patois of digital photography is because I'm creating a cover of a
> book. If (as Jem said), I will be able to "blow up" the 300 dpi
> image to only such a small dimension, then either I'm interpreting
> the PhotoPlus "print preview" of this homemade book cover very
> incorrectly, or else the software is so good it actually enlarged
> the image so that it fits the 6" width.
You can not "blow up" a 402 pixel, 300 ppi image to fill a 6 inch book
cover If spread 402 pixels over 6", you're printing at 402/6 = 67 ppi.
It is simple math.
(You can resample it to 300 ppi - amd depending on the picture, that
may be successful or not. Resampling is a subject in itself that I
will not get into here.)
The resolution of computer screens vary, but 72 ppi is a fairly common
figure. That is pretty close to 67 ppi - probably too close for you
to tell the difference by visual inspection of the "print preview"
afforded by this type of software.
In short, because the resolution of the computer screen is much lower
than the resolution your printer has said he requires to make a
quality print (300 ppi), you cannot, at least not without considerable
experience in this area, judge print quality from looking it the
"print preview" image at screen resolution. You really should have a
look at the FAQ. This one is covered too. It is Q11.
(URL: http://hannemyr.com/photo/pixels.html )
> If I uploaded it to Photobucket or another site, would anyone be
> good enough to visit it and tell me if I'm misinterpreting what it
> will look like as a finished product?
Sure, if you upload it, I'll look at it and comment.
And if possible, please upload the original image as well.
ASAAR
> I have a much more prosaic question, however. The only reason (or at
> least the precipitation reason) I need to learn more about the patois
> of digital photography is because I'm creating a cover of a book. If
> (as Jem said), I will be able to "blow up" the 300 dpi image to only
> such a small dimension, then either I'm interpreting the PhotoPlus
> "print preview" of this homemade book cover very incorrectly, or else
> the software is so good it actually enlarged the image so that it fits
> the 6" width.
Any program can "blow up" an image, ie, make a print at whatever
size you want, but the problem you have is that the source file
appears to lack the needed resolution. The 150 dpi figure is really
meaningless as a measure of the resolution of whatever source file
you're dealing with, whether it's clip art or a photographic image.
It's simply a suggestion as to how large to print the file.
Since your goal is to produce a print that's 6 inches wide, at
300dpi that would (if using a camera to take the picture) be best
met using a camera having a sensor having a sensor that has 1800
pixels (6 x 300) in its horizontal dimension. Some cameras have
sensors that have a 3:2 aspect ratio, which exactly matches the
dimensions of common 6" x 4" snapshot prints. That the clip art
measures 600 x 402 pixels instead of exactly 600 x 400 makes me
wonder if that's the reason you're getting the thin black vertical
stripe in the background. If so, you could crop 2 pixels worth from
the top or bottom of the clip art and that may eliminate the black
stripe, assuming that the background uses a 3:2 ratio.
That said, if you had a camera with a 1,800 x 1,200 resolution
sensor, the images it produces really doesn't have a meaningful dpi
value, and whatever it is, it's only stored as a number in the image
file, and is not part of the image at all. Most image editing
programs will let you change the dpi value without making any
changes to the image. It's only used as a suggestion for how large
the picture should be printed. With the camera described here, if
the dpi value is set to 300, the resulting print would be 6" x 4".
Change it to 600 dpi, and the image would print out at 3" x 2". My
own camera produces JPG files that have a dpi value of 1 (yes,
one!). If that value was used for the above image, the expected
print size would be 150 feet x 100 feet, so you should see by now
that the dpi value embedded in the source image file is only a
suggestion.
Your problem is that the clip art is that it is only 600 x 402
pixels, and if printed 6 inches wide, that would produce a print of
only 100 dpi, much less than the publisher's expected 300 dpi. Or,
as Jem and Gisle said, if printed at 300 dpi, the print would be
only 2 inches wide. Image editing programs can resample the clip
art to allow the image to print at 300 dpi, but it would only be
very slightly improved, and any decent publisher would be able to
see that it lacks the sharpness of a true 300 dpi print, and by a
considerable margin. If the subject of the clip art is otherwise
ideal, and you can't find or create a higher resolution version, you
might see if the publisher would accept a resampled 300 dpi version,
but it would be very unwise to submit it without any prior
disclosure. The worst possible outcome for you would be if the
publisher only discovered that the image was unsuitable after
expending time and/or money on the book's cover, or discovering the
problem too late to meet a deadline.
John McWilliams
> "DPI" has been confusing many including myself once on the day I switched
> from conventional photography to digital photography.
Yes, it has. It works for me to remember that dpi doesn't come into play
until the image is sent to the printing machine. Before that, it's all
pixels, ie. ppi, and even pixels can be confusing at first depending on
the software being used, and how an image is displayed.
--
John McWilliams
Marvin
result wouldn't be completely equivalent to a picture taken
with the higher pixel count. But I also figured that the
difference would not be noticable, considering the kind of
paper it would be printed on.
Jules Vide
too much in it I half-understand and am half-confused by. There's a
need for mathematical abilities apparent here, or at least for a
formula, I can't figure out. For example, my HP 4.1 MP allows me to
take pictures with various levels of compression, the best being 4MP
with the "greatest" (or, jeez! lowest?) amount of compression...which
HP identifies only by the use of *star symbols*, NOT by dpi or ppi or
anything professional.
If "dpi" is merely a concept, and if the "square footage" of an image
can't exceed 4100 pixels--
Oh, it's too byzantine. Again, I understand much of what you mean, and
you apparently are an educator or tech writer, because as with Gisle's
responses, while reading your explanation, my fair-to-middling brain is
going Yes, Yes, Hmm..., Well, Okay.
But then the amount of factors that contribute to digital camera
resolution become unwieldy. So I'll just finish by asking: was
purchasing a 4.1 just a waste of money? What is the smallest MP size
camera that can take high resolution photographs?
Thank you very much, ASAAR.
Jules Vide
>
> You can not "blow up" a 402 pixel, 300 ppi image to fill a 6 inch book
> cover If spread 402 pixels over 6", you're printing at 402/6 = 67 ppi.
> It is simple math.
Gisle, how do you arrive at the figure 402 for the total pixel amount?
Shouldn't you "multiply" 600 x 402, then divide by 6"?
> The resolution of computer screens vary, but 72 ppi is a fairly common
> figure. That is pretty close to 67 ppi - probably too close for you
> to tell the difference by visual inspection of the "print preview"
> afforded by this type of software.
I had not considered this. Thank you for the head's up.
> In short, because the resolution of the computer screen is much lower
> than the resolution your printer has said he requires to make a
> quality print (300 ppi), you cannot, at least not without considerable
> experience in this area, judge print quality from looking it the
> "print preview" image at screen resolution. You really should have a
> look at the FAQ. This one is covered too. It is Q11.
> (URL: http://hannemyr.com/photo/pixels.html )
I do not have photograph quality cartridges in my inexpensive HP
printer, but I printed a copy of the image. To my amazement, the
printed sample book cover, with photo, was not nearly as diffuse (or
"dithered?" Am I using "dither" correctly?) as I expected.
However, your response made me realize that what is clear on a
primitive system may turn out hideously blurred on a printer
more...sensitive. Thank you again for the head's up. As I asked
ASAAR, was the purchase of a 4.1 MP camera just a waste of money?
Could I *ever* take a picture with a dpi of 300, if the "norm" is 72?
Or will I just be taking teeny tiny pictures I could maybe paste onto
grains of rice and sell at souvenir stands? :)
Jules Vide
> You can not "blow up" a 402 pixel, 300 ppi image to fill a 6 inch book
> cover If spread 402 pixels over 6", you're printing at 402/6 = 67 ppi.
> It is simple math.
>
> (You can resample it to 300 ppi - amd depending on the picture, that
> may be successful or not. Resampling is a subject in itself that I
> will not get into here.)
At the risk of wearing out my welcome, is there a time and place you
*would* get into the subject? Is resampling software extremely costly?
Has it been perfected, or is it still in the experimental stages like,
say, voice recognition software? Am I showing how stupid I am by
asking these questions?
> You really should have a look at the FAQ. This one is covered too. It is Q11.
Yes, I will. Thank you.
Paul J Gans
>The resolution of computer screens vary, but 72 ppi is a fairly common
>figure. That is pretty close to 67 ppi - probably too close for you
>to tell the difference by visual inspection of the "print preview"
>afforded by this type of software.
Nit pic, but I think an important one. One can measure
the ppi of your own screen fairly simply in any photo or
art program that will let you draw a line with a known
number of pixels on it.
For instance, draw a 200 pixel line. Save it as a gif
or some other lossless format. Display it with your
browser and now *measure* the length of the line with
a ruler.
For example if the line measures 3 inches long, then
the screen is
(200 pixels) / (3 inches) = 66.7 pixels per inch
I have a number of various monitors. They seem to go
more toward 95 ppi these days, though I agree that 67 or
so was very common a few years back.
----- Paul J. Gans
Paul J Gans
>Gisle Hannemyr wrote:
>> You can not "blow up" a 402 pixel, 300 ppi image to fill a 6 inch book
>> cover If spread 402 pixels over 6", you're printing at 402/6 = 67 ppi.
>> It is simple math.
>>
>> (You can resample it to 300 ppi - amd depending on the picture, that
>> may be successful or not. Resampling is a subject in itself that I
>> will not get into here.)
>At the risk of wearing out my welcome, is there a time and place you
>*would* get into the subject? Is resampling software extremely costly?
> Has it been perfected, or is it still in the experimental stages like,
>say, voice recognition software? Am I showing how stupid I am by
>asking these questions?
Not at all stupid. Resampling is just that. A program
looks at the image, pixel by pixel, and inserts new pixels
in between the old ones. Each program differs in how it
decides what color to make the new pixels. Some are more
successful in some scenes than others and at other times
it is the other way around.
A free WinXP program called "Irfanview" (Google for it)
has a rather nice upsampling program -- and better downsampling
ones.
The insurmountable problem is this: there is no way for
any program to produce real detail in the newly inserted
pixels. The information about what should be there simply
is not present in the image. You can't guess and neither
can a program.
That's why folks always advise shooting with more pixels
than you need. You can always throw away some pixels.
And programs (Irfanview is one example) are better at
knowing which pixels to delete than they are at creating
new pixels.
------ Paul J. Gans
ASAAR
> But then the amount of factors that contribute to digital camera
> resolution become unwieldy. So I'll just finish by asking: was
> purchasing a 4.1 just a waste of money? What is the smallest MP size
> camera that can take high resolution photographs?
The simple answer is that they all can. To really answer your
question you have to state the largest size prints you'll be making,
as well as the distance they'll be viewed from. Then you can figure
out (roughly) what size sensor you'll need.
Taking your 4mp camera as an example, its sensor should be similar
to the 4mp sensor in my Fuji camera, which at its highest resolution
is 2272 x 1704 pixels. If you multiply these two numbers, you get
the total number of pixels, or 3,871,488. Not quite 4mp, but close
enough.
If the goal is to print at 300 dpi, you could start by dividing
the sensor's dimensions by 300, so 2272/300 x 1704/300 could easily
produce a 300 dpi print that measures 7.6 x 5.7 inches, which is a
bit larger than one of the standard print sizes, 5" x 7". So a 4mp
camera should be capable (if it has a decent lens and is designed
well) of producing extremely sharp pictures at this size. But for
many people, 300 dpi is overkill, because pictures printed at that
resolution, at their normal viewing distance, won't appear any
sharper than prints made at 250 dpi or even 200 dpi. It depends on
the person's eyes, and their normal viewing distance. An
interesting test would be to examine two pictures, one printed at
200 dpi and the other at 300 dpi, and then determine at what viewing
distance the two pictures are indistinguishable.
This distance won't be the same for all people. Without an
extremely high resolution picture next to it, I think most people
would find 200 dpi prints to be fairly sharp. The same picture
described above, if printed at 200 dpi would now print out at 8.5" x
11.4", so you should be able to make very nice 8" x 10" prints
(another standard print size) using your camera. And as larger
prints tend to be viewed from a slightly greater distance, the
effect of reducing the dpi from 300 to 200 is lessened. Then you
have software tools available that can allow you to make even larger
prints that appear to be quite sharp.
You also mentioned compression. You'll get the best results by
using the highest resolution (4mp), and if your HP allows you to
select between two different 4mp setting, choose the one having the
least compression, as when compression is increased, more detail is
lost when the computer loads the image and expands it. The tradeoff
is that higher compression reduces the image file size, which allows
more pictures to fit on your memory cards. If you ever need small
sizes, such as for emailing to friends, or for web images, software
can easily reduce the size of the JPG files. But if you take
pictures at reduced resolution or higher compression, you're stuck
with what you have. The problem with higher compression, btw, is
not that you lose resolution, but that it often produces "artifacts"
and "posterization". These may not always show up in small prints,
but if you examine them closely enough, or make larger prints, the
effects are fairly easy to see. If you're not familiar with these
terms, photo review websites (such as dpreview.com) have glossaries
that define many photographic terms, or google.com can help.
Gisle Hannemyr
> Gisle Hannemyr wrote:
>> You can not "blow up" a 402 pixel, 300 ppi image to fill a
>> 6 inch book cover If spread 402 pixels over 6", you're printing
>> at 402/6 = 67 ppi. It is simple math.
> Gisle, how do you arrive at the figure 402 for the total pixel
> amount? Shouldn't you "multiply" 600 x 402, then divide by 6"?
The 402 is not the total pixel amount. 402 is the number of pixels
across the shortest side of the image. Since it is a book cover, I
assume you want to print in portrait format, so to get the resolution
in pixels per inch (ppi) you just divide the number of pixels along
the side with the number of inches across the book cover.
If you instead print it on the cover in landscope mode, the resolution
is 600 pixels / 6 inches = 100 ppi.
>> In short, because the resolution of the computer screen is much
>> lower than the resolution your printer has said he requires to make
>> a quality print (300 ppi), you cannot, at least not without
>> considerable experience in this area, judge print quality from
>> looking it the "print preview" image at screen resolution. You
>> really should have a look at the FAQ. This one is covered too. It
>> is Q11. (URL: http://hannemyr.com/photo/pixels.html )
> I do not have photograph quality cartridges in my inexpensive HP
> printer, but I printed a copy of the image. To my amazement, the
> printed sample book cover, with photo, was not nearly as diffuse
Well, if the quality is acceptable to you, then it is acceptable to
you.
> (or "dithered?" Am I using "dither" correctly?) as I expected.
No - dithering is what an ink yet printer does to simulate different
shades of colour. The word you are looking for is pixelated - which
is the "look" you get from too much interpolation.
Read more on interpolation here:
http://hannemyr.com/photo/interpolation.html
> However, your response made me realize that what is clear on a
> primitive system may turn out hideously blurred on a printer
> more...sensitive. Thank you again for the head's up. As I asked
> ASAAR, was the purchase of a 4.1 MP camera just a waste of money?
So far, we've been talking about a 600x402 pixel image. That is
241200 pixels, or 0.24 MP. A typical 4.1 MP camera will be something
like 2340 x 1755. If you were to print a photograph taken with
your camera on a 6 inch book cover, the resolution would be
pretty close to 300 ppi, even if you choose portrait format
(1755/6 = 292 ppi).
> Could I *ever* take a picture with a dpi of 300, if the "norm" is
> 72? Or will I just be taking teeny tiny pictures I could maybe
> paste onto grains of rice and sell at souvenir stands? :)
You really should take a look at the FAQ at
http://hannemyr.com/photo/pixels.html
This is answered in Q2. And how to convert from 72 ppi to 300 ppi
without losing quality is covered in Q16.
Gisle Hannemyr
> Gisle Hannemyr wrote:
>> (You can resample it to 300 ppi - amd depending on the picture,
>> that may be successful or not. Resampling is a subject in itself
>> that I will not get into here.)
> At the risk of wearing out my welcome, is there a time and place you
> *would* get into the subject?
Sure. I have a web page on that subject as well:
http://hannemyr.com/photo/interpolation.html
> Is resampling software extremely costly?
Some of it, like Genuine Fractals at $160, is. But there is also
good commercial grade software, like Qimage available for around
$ 50. If you already have PhotpShop, you've got bicubic (and other)
interpolation methods built in.
There is at least one good freeware program out there (IrfanView),
that does an amazing job.
> Has it been perfected, or is it still in the experimental stages
> like, say, voice recognition software?
The software is still improving, but I would say that some of it
is pretty good already. It is certainly usable.
Jules Vide
>
> If the goal is to print at 300 dpi, you could start by dividing
> the sensor's dimensions by 300, so 2272/300 x 1704/300 could easily
> produce a 300 dpi print that measures 7.6 x 5.7 inches, which is a
> bit larger than one of the standard print sizes, 5" x 7". So a 4mp
> camera should be capable (if it has a decent lens and is designed
> well) of producing extremely sharp pictures at this size. But for
> many people, 300 dpi is overkill, because pictures printed at that
> resolution, at their normal viewing distance, won't appear any
> sharper than prints made at 250 dpi or even 200 dpi. It depends on
> the person's eyes, and their normal viewing distance. An
> interesting test would be to examine two pictures, one printed at
> 200 dpi and the other at 300 dpi, and then determine at what viewing
> distance the two pictures are indistinguishable.
Ah, see, when "laypeople" go to buy a digital camera, the
salesmen(women) either are trained not to inquire for what purposes you
*may* need the camera or are ignorant themselves of the limits of a
particular camera's capabilities. How I wish I had the above paragraph
at my disposal before deciding it was finally time to go digital.
> This distance won't be the same for all people. Without an
> extremely high resolution picture next to it, I think most people
> would find 200 dpi prints to be fairly sharp. The same picture
> described above, if printed at 200 dpi would now print out at 8.5" x
> 11.4", so you should be able to make very nice 8" x 10" prints
> (another standard print size) using your camera. And as larger
> prints tend to be viewed from a slightly greater distance, the
> effect of reducing the dpi from 300 to 200 is lessened. Then you
> have software tools available that can allow you to make even larger
> prints that appear to be quite sharp.
Indeed... I took what everyone in my family agrees was the candid
portrait-of-a-lifetime of my mom, who's getting on in years. This
photograph-- You couldn't find a photographer who would have gotten
quite that smile. Then, when I uploaded it to my local print shop, I
got a "cropped" 8 x 10 where her head is the size of a cantalope, the
proportion (and beauty) of the background is nearly obliterated, and
the image is, to put it mildly, lossy.
The explanation was that the dimensions of the image didn't allow for
the 8 x 10 I imagined, which I suppose has something to do with aspect
ratio. Nevertheless, I blamed my camera.
> The problem with higher compression, btw, is
> not that you lose resolution, but that it often produces "artifacts"
> and "posterization". These may not always show up in small prints,
> but if you examine them closely enough, or make larger prints, the
> effects are fairly easy to see.
See above. Hewlett Packard and all digital camera manufacturers should
BOLD FACE the fact that their cameras' default settings are not set at
lowest compression.
Jules Vide
>
> That's why folks always advise shooting with more pixels
> than you need. You can always throw away some pixels.
> And programs (Irfanview is one example) are better at
> knowing which pixels to delete than they are at creating
> new pixels.
When I queried HP on this 300 dpi matter, a tech support rep suggested
software called PhotoPlus. I was prepared to buy the software but was
able to download it for free, and so far, I've found it the most
"dummy-friendly" photo editing software I've used (and believe it or
not, I've tried my hand at whatever photo editing software chooses my
laptop as a dock in a storm!). Anyway, if you or anyone who happens to
read this post is familiar with this software, I'd appreciate hearing
your opinions of its abilities to do what Irfan does.
Irfan's been around for a long time, no? I recall receiving files
created in it from the early "20's."
ASAAR
> Indeed... I took what everyone in my family agrees was the candid
> portrait-of-a-lifetime of my mom, who's getting on in years. This
> photograph-- You couldn't find a photographer who would have gotten
> quite that smile. Then, when I uploaded it to my local print shop, I
> got a "cropped" 8 x 10 where her head is the size of a cantalope, the
> proportion (and beauty) of the background is nearly obliterated, and
> the image is, to put it mildly, lossy.
> The explanation was that the dimensions of the image didn't allow for
> the 8 x 10 I imagined, which I suppose has something to do with aspect
> ratio. Nevertheless, I blamed my camera.
Right. Your camera's sensor doesn't match the 8 x 10 aspect ratio
of the print. What you could do in the future to get the best
result is to crop the picture yourself, using whatever photo editor
is convenient. You'd still lose a bit of the image, but at least
the cropped portion would be what appeared best to you. Your local
print shop may not even have had a person doing the crop manually.
The decisions may have been automated. Another option would be to
have a good print shop print the entire image to a non-custom paper
size. This would be more expensive, and then you'd also have to get
a non-custom sized frame made, be you'd end up a photo not missing
any essential elements.
It would be nice if cameras could overlay a wire frame with the
aspect ratio of your choice in the viewfinder (EVF, not optical) or
LCD display, to match the paper dimensions you expect to use for the
print. The closest you can come to this is with some cameras that
have a slightly lower resolution choice that has a 3:2 aspect ratio.
It wouldn't help for 8" x 10" prints, but it's the same ratio as the
common 6" x 4" snapshots that most people have made. Then, as most
EVFs and LCDs fairly accurately show what will appear in the
captured image, you'd be able to see almost exactly what would
appear in the print, without having to worry about losing something
due to your local print shop's careless cropping.
Paul J Gans
>Paul J Gans wrote:
>>
>> That's why folks always advise shooting with more pixels
>> than you need. You can always throw away some pixels.
>> And programs (Irfanview is one example) are better at
>> knowing which pixels to delete than they are at creating
>> new pixels.
>When I queried HP on this 300 dpi matter, a tech support rep suggested
>software called PhotoPlus. I was prepared to buy the software but was
>able to download it for free, and so far, I've found it the most
>"dummy-friendly" photo editing software I've used (and believe it or
>not, I've tried my hand at whatever photo editing software chooses my
>laptop as a dock in a storm!). Anyway, if you or anyone who happens to
>read this post is familiar with this software, I'd appreciate hearing
>your opinions of its abilities to do what Irfan does.
Sorry. Not familiar with it, but if it does the job for you,
that's great.
>Irfan's been around for a long time, no? I recall receiving files
>created in it from the early "20's."
You've got me confused there.
---- Paul J. Gans
ASAAR
>> Irfan's been around for a long time, no? I recall receiving files
>> created in it from the early "20's."
>
> You've got me confused there.
If a touch typist produced a typo by accidentally using the ring
or middle finger of the wrong hand, "from the early 90's" (or 80's)
may have been what was intended. Or he could be a visitor from the
future, maybe from 2025, where he's been using ver. 14.63. :)
Jules Vide
Jules wrote:
> >Irfan's been around for a long time, no? I recall receiving files
> >created in it from the early "20's."
>
> You've got me confused there.
Actually, it was in the late 90's. One of those annoying friends who
adds your name without permission to mailing lists used to send me
those pluperfectly irritating "Jokes of the Day." These jokes always
contained photos I could not open in Outlook Express, and the file
symbol always attached to them was "Irfan."
Jules Vide
>
> If a touch typist produced a typo by accidentally using the ring
> or middle finger of the wrong hand, "from the early 90's" (or 80's)
> may have been what was intended. Or he could be a visitor from the
> future, maybe from 2025, where he's been using ver. 14.63. :)
I keep my distance from the future; it's too dark and scary :). I've
been in search of an accessible term for this decade ever since it
started. It's so unwieldy saying "the two-thousands." On Meet the
Press, I heard someone call them the "20's."
Jules Vide
> It would be nice if cameras could overlay a wire frame with the
> aspect ratio of your choice in the viewfinder (EVF, not optical) or
> LCD display, to match the paper dimensions you expect to use for the
> print. The closest you can come to this is with some cameras that
> have a slightly lower resolution choice that has a 3:2 aspect ratio.
> It wouldn't help for 8" x 10" prints, but it's the same ratio as the
> common 6" x 4" snapshots that most people have made. Then, as most
> EVFs and LCDs fairly accurately show what will appear in the
> captured image, you'd be able to see almost exactly what would
> appear in the print, without having to worry about losing something
> due to your local print shop's careless cropping.
Are there any rules of thumb you can use on the fly to determine
whether a particular viewfinder image will reproduce as you wanted? I
mean, should you just find out the aspect ratio of your particular
camera, then, say, in a portrait of a person, have approximately
one-third be from the neck down, approximately one-third be their head,
and one-third be background above their head? (I'm using one-third as
an example.)
Jules Vide
> Isn't prints measuring 7.5 x 10 inches an option in the USA?
I've never heard of this anywhere where *inexpensive* prints are made.
Custom is another matter.
Za Phod
If you remember your Pythagoras;
I have a 19" screen with aspect ration of 4:3 (width:height).
Simple scaling of a right angled triangle of sides 3:4:5;
width = 19/5 *4 = 15.2" (of course you can measure this if you have a
ruler to hand).
My graphics card is running at 1600 pixels across giving a measurement
of 1600/15.2 = ~105 pixel/inch (ppi)
Zaphod
ASAAR
> Are there any rules of thumb you can use on the fly to determine
> whether a particular viewfinder image will reproduce as you wanted? I
> mean, should you just find out the aspect ratio of your particular
> camera, then, say, in a portrait of a person, have approximately
> one-third be from the neck down, approximately one-third be their head,
> and one-third be background above their head? (I'm using one-third as
> an example.)
Not too easily, since compared to what the sensor will actually
capture can vary considerably from what the viewfinder shows,
depending on the camera that's used. Optical viewfinders vary the
most, and usually show anything from about 3% to maybe 20% less than
what will end up being captured. The less accurate the viewfinder
though, the easier it will be to print without losing any essential
elements of the subject since you'd have a much more "croppable"
image. EVFs and LCD displays are usually very accurate - at least
the ones I've used have been. Some cameras have the ability to
superimpose a grid of a couple of thin horizontal and vertical lines
on the viewfinder. This may be all that you need to help with your
rule of thumb, but if not . . .
What you can do (if this isn't too much of a chore) is make a
print that includes everything from the image. You'll use that to
superimpose several rectangles on it having the same aspect ratios
that you might expect to use for future prints. When you take the
shot, carefully note any objects that appear in the extreme corners
of the viewfinder. This won't be necessary if you can later easily
return with the camera placed in exactly the same location. But if
you do this, you'd have to be able to return the lens's zoom to
exactly the same position it was in when the shot was taken, so
you'll want to use either its extreme wide or extreme tele position.
Using a ruler, measure and draw several rectangles on it having
whatever print aspect ratios you might use, such as 4x6, 5x7 and
8x10. These should be made as large as will fit on the print. It
might be more convenient to use an image editor's crop tool to
superimpose these rectangles on your monitor's screen. Then you
wouldn't need a ruler, just a straightedge to draw the rectangles on
the print using the screen as a guide. Then, draw another rectangle
on the print, preferably using a different color, that matches the
exact coverage that you see in your viewfinder. That's why I
mentioned recording the objects that would appear in the corners of
the viewfinder. If you didn't do this, you'd have to return the
camera to the exact position it was in when you took the shot. Then
the viewfinder would be able to show where this "viewfinder coverage
rectangle" could be drawn on the print.
The main thing that this print would be useful for is to maximize
the size of the image that you could take, without losing any of it
to cropping. Personally, I think it's a better idea in most cases
to simply include more in the shot that you'll need, relying on your
camera to have sufficient resolution so that there'll be more than
enough remaining after cropping out whatever isn't needed for the
final print. Using the print guide with all of the rectangles, etc.
should only be necessary if you need to produce extremely large
prints, where you can't afford to lose resolution due to cropping.
In the end, if you do make such a print, you'll probably find that
before too long your rule of thumb (for this particular camera only)
will be so ingrained that you won't need to use it anymore.
ASAAR
> I keep my distance from the future; it's too dark and scary :). I've
> been in search of an accessible term for this decade ever since it
> started. It's so unwieldy saying "the two-thousands." On Meet the
> Press, I heard someone call them the "20's."
Wow, that's clumsy, and not very accurate. If the same rule was
applied to events that occured in 1923 or 1978, they'd both be from
the "19's", not very informative for identifying any particular
decade. And if the Press People wanted to distinguish between
events that happened in the 1900's from the 2000's, they could
simply say "last century" or "this century". :)
Paul J Gans
Could be. The answer may be further down this thread...
---- Paul J. Gans
Paul J Gans
Ok. It makes more sense now. Irfanview will convert files
from one format to another. Another of its useful properties.
---- Paul J. Gans
Jules Vide
on this thread interpolated the photograph in question for me. I got a
5 x 7 print today of the 300 DPI (up from 150 DPI) image, and not only
am grateful beyond words for the favor, because of how great a mistake
it prevented me from making (in thinking I could use this photo as the
cover of a book), but am shocked by how, even at 5 x 7 size, the print
quality is so S-t-u-n-n-i-n-g-l-y bad.
First I should say that this picture ain't "resolution rocket science":
there's black silhouette of a summer forest after dusk against a very
vivid purplish sky. So we're not talking a whole lot of palette. But
the photo produced with LITERAL "blocks" of black/gray undifferentiated
pixels. This "blocking" effect of the interpolated photo repeated
somewhat in the reflection of the purplish dusk in a lake; again, there
were "Impressionist-like" whole blocks of color. The tops of the
(black) silhouetted pine trees looked as if someone dropped water on
them.
So for $2.00, I had proof of how "If it ain't there to begin with, it
ain't there to begin with."
Now--
ASAAR, I have profited from *every* thing you've written here on this
thread. But I've read this post several times and am still uncertain
of the purpose of creating what I suppose would be a mock-up (at
different popular dimensions) of a particular scene/portrait.
PLEASE do not be offended if I ask if you'd take the trouble to
re-explain.
Sincere thanks.
ASAAR
> ASAAR, I have profited from *every* thing you've written here on this
> thread. But I've read this post several times and am still uncertain
> of the purpose of creating what I suppose would be a mock-up (at
> different popular dimensions) of a particular scene/portrait.
>
> PLEASE do not be offended if I ask if you'd take the trouble to
> re-explain.
No offense. I hope this will make it a little clearer. If you
had a camera that showed you exactly what would be captured in the
JPG or RAW files it produces, you'd be able to frame the pictures
perfectly if the camera also superimposed gridlines corresponding to
whatever print size you'd want to make. You'd simply make sure that
whatever you'd need to appear in the print would appear inside the
gridline. Unfortunately, cameras have two problems that complicate
getting exactly what you want.
First, many cameras have viewfinders that show a good deal less of
what will appear in the final image. This partially solves the
problem of part of the subject being cropped in the print because
you'd have a shot that includes the subject and more. Usually much
more, and this wastes resolution as the subject is smaller in the
image than it could have been. If the remaining pixels after
cropping are sufficient for the size print you intend to make, then
this wouldn't be a concern, and could be a good reason to get a
camera with more resolution than you think you might need. That
would take the effort out of trying to perfectly frame the pictures
you take.
Another problem would occur with viewfinders that show almost
exactly what will appear in the final image. This is usually the
case when using an electronic viewfinder (EVF) or the LCD display.
In this case, if the subject is framed perfectly (meaning as large
as possible) in the viewfinder, then since the horizontal and
vertical dimensions (or aspect ratio) of the sensor is unlikely to
match the aspect ratio of the print, something is bound to get
chopped off unless unnaturally large borders are used for the print.
With the print I suggested trying to make, the entire picture
would provide an example of what would be included in the JPG, or
image file, and you'd have drawn on it several smaller rectangles,
one of which would correspond to what your particular camera shows
in its viewfinder. With my camera's EVF viewfinder, this rectangle
would be almost exactly the same size as the print, making picture
composition or framing somewhat simpler. The other rectangles would
have the aspect ratios of the prints you're likely to take, and
would be sized as large as possible while still fitting within the
print.
As an example, if you wanted to make an 8" x 10" print, and your
viewfinder shows less than that the camera actually captures.
Assuming you're taking photos using landscape orientation and the
print with the rectangles was made on a 4" x 6" print, the
rectangle corresponding to the 8" x 10" aspect ratio would be 4"
tall and 5" wide, showing that in the final print a good deal of the
image would be truncated from the left and right sides. The
rectangle that shows what would appear in the viewfinder might lie
completely within the 8" x 10" rectangle, or there might be some
overlap. But the point is that with some imagination, you could
take a picture with the camera where the subject doesn't completely
appear in the viewfinder (because you're trying to make it as large
as possible in the final image), and yet not be so large that it's
either too large to fit within the sensor, or too large to fit
within the slightly smaller part of the sensor that would have the
largest possible 8:10 aspect ratio.
This will take some estimating and some trial and error too, and
sometimes you might get the subject too large to avoid getting a
small part of it lopped off from the final print. But before too
long you'd probably get a good idea how to frame your subjects with
a sufficient margin of error so that they'd appear as large as
reasonably possible without risking losing toes or ears. :) It
would be much, much simpler if cameras had accurate viewfinders and
aspect ratio gridlines I mentioned in the first paragraph above.
Fortunately, for most people that couldn't care less about going
to all this trouble, and usually have 4" x 6" prints made, they have
an easy solution. They could get a camera that shows 100% of what
will appear in the image, and also make sure that the camera has at
least one high resolution mode that has a 3:2 aspect ratio.
Unfortunately, all of my 4mp Fuji's available resolutions have a 4:3
aspect ratio. This information isn't often readily available, but
it can be gotten from checking the camera's menu in a photo store if
they let you handle them (one might need to bring a memory card).
Lacking that, you could get the necessary resolution information by
downloading camera manuals from manufacturer websites.
Jules Vide
> First, many cameras have viewfinders that show a good deal less of
> what will appear in the final image. This partially solves the
> problem of part of the subject being cropped in the print because
> you'd have a shot that includes the subject and more. Usually much
> more, and this wastes resolution as the subject is smaller in the
> image than it could have been. If the remaining pixels after
> cropping are sufficient for the size print you intend to make, then
> this wouldn't be a concern, and could be a good reason to get a
> camera with more resolution than you think you might need.
This (and your entire excellent post) raises the question Why? I
remember my Dad's Roloflex, looking at those squares and thinking they
were a pain. If I had known *ANY* of what I've learned on this thread,
I would never have bought my HP 4.1. I would have waited until I was
able to afford a digital camera in the $1000 range.
I badly need a 300 DPI photograph that will not look like something
taken by someone needing cataract surgery, to fill a 6 x 9 inch
paperback cover. I was resigned on foreshortening the photograph for
reasons of economy, time, and annoyance to something 7 x 5 (landscape),
and filling the rest of the cover with a solid black background.
But after reading your (and others') posts, I'm getting indignant about
what I now see was a misleading, VERY misleading, sales pitch by the dp
industry in general: that for a hundred bucks-and-change, I'd be able
to take pictures that would look like pictures taken from that old
Roloflex or Nikon. The 4.1 I have is a *toy*. It should have been
advertised as such.
> Another problem would occur with viewfinders that show almost
> exactly what will appear in the final image. This is usually the
> case when using an electronic viewfinder (EVF) or the LCD display.
> In this case, if the subject is framed perfectly (meaning as large
> as possible) in the viewfinder, then since the horizontal and
> vertical dimensions (or aspect ratio) of the sensor is unlikely to
> match the aspect ratio of the print, something is bound to get
> chopped off unless unnaturally large borders are used for the print.
Again, the question is WHY? Is it to promote the sale/need of photo
editing software? If I'm missing something, please tell me: is it
cheaper to make a camera with 4:3 ratio than 3:2?
>
> Fortunately, for most people that couldn't care less about going
> to all this trouble, and usually have 4" x 6" prints made, they have
> an easy solution. They could get a camera that shows 100% of what
> will appear in the image, and also make sure that the camera has at
> least one high resolution mode that has a 3:2 aspect ratio.
> Unfortunately, all of my 4mp Fuji's available resolutions have a 4:3
> aspect ratio. This information isn't often readily available, but
> it can be gotten from checking the camera's menu in a photo store if
> they let you handle them (one might need to bring a memory card).
> Lacking that, you could get the necessary resolution information by
> downloading camera manuals from manufacturer websites.
All I know is that "most people" is a euphemism (at least I think so).
When the HP rep told me to use PhotoPlus software, for example (which
contains the most well-written user "manual" of any photo editing
software I've come across), I was hit in the face with SO many new
terms you wouldn't find even two years ago--"anti-aliasing";
"bicubic"--for the average consumer who has to work for a living and
who doesn't have the time, energy, or ability to become an
engineering/graphics master, there remains only one feasible option if
you want to be part of the inescapable world of digital photography:
pony up $1000+ for the proper camera.
And I'm still trolling the web for a 300 DPI picture of a
"sinister"-looking dusk...
Thanks again, ASAAR. You make very interesting reading.
Paul Rubin
> This (and your entire excellent post) raises the question Why? I
> remember my Dad's Roloflex, looking at those squares and thinking they
> were a pain. If I had known *ANY* of what I've learned on this thread,
> I would never have bought my HP 4.1. I would have waited until I was
> able to afford a digital camera in the $1000 range.
I haven't followed this thread but you can make a passable 8x10 photo
with a $150 digicam. It sounds like you started with a screen shot or
something like that. Where is the 4.1mp original file that came out
of the camera?
> The 4.1 I have is a *toy*. It should have been advertised as such.
It's just not as flexible as fancier cameras and doesn't handle as
well, and has worse noise in low light. With good lighting and
patience (which you'd have needed with your dad's Rolleiflex) you can
take good shots with the HP.
> Again, the question is WHY? Is it to promote the sale/need of photo
> editing software? If I'm missing something, please tell me: is it
> cheaper to make a camera with 4:3 ratio than 3:2?
4:3 is closer to TV format and is quite normal. Remember that your
dad's Rolleiflex was 1:1. 35mm cameras are 3:2 because they used 35mm
movie film and used two movie frames. The movie frames were 18x24mm
(3:4) so the double frame is 36x24mm (3:2).
> And I'm still trolling the web for a 300 DPI picture of a
> "sinister"-looking dusk...
An expensive camera won't help with that. You need to know what
you're doing no matter what camera you use. Try this though:
http://commons.wikimedia.org/wiki/Image:Twilight3_-_NOAA.jpg
Jules Vide
>
> I haven't followed this thread but you can make a passable 8x10 photo
> with a $150 digicam. It sounds like you started with a screen shot or
> something like that. Where is the 4.1mp original file that came out
> of the camera?
Wow, thanks, Paul, 'cuase I signed on again to ask if the *print* of
the interpolated image (done by a very kind reader of this thread)
could have printed with such great distortion because I uploaded it and
sent it over the 'net rather than go in and use my laptop or a CD to
transfer it from.
> > The 4.1 I have is a *toy*. It should have been advertised as such.
>
> It's just not as flexible as fancier cameras and doesn't handle as
> well, and has worse noise in low light. With good lighting and
> patience (which you'd have needed with your dad's Rolleiflex) you can
> take good shots with the HP.
I'm assuming by "worse noise in low light" you mean low light messes
with the minds of the digital sensors. All I know is that this camera
does work well--excellently well, in fact--in the out of doors but very
poorly indoors. Your explanation seems to fit.
> > Again, the question is WHY? Is it to promote the sale/need of photo
> > editing software? If I'm missing something, please tell me: is it
> > cheaper to make a camera with 4:3 ratio than 3:2?
>
> 4:3 is closer to TV format and is quite normal. Remember that your
> dad's Rolleiflex was 1:1. 35mm cameras are 3:2 because they used 35mm
> movie film and used two movie frames. The movie frames were 18x24mm
> (3:4) so the double frame is 36x24mm (3:2).
Again, thanks. I'm saving and printing out this whole freaking thread,
'cause I've gotten some great minds to contribute!
> > And I'm still trolling the web for a 300 DPI picture of a
> > "sinister"-looking dusk...
>
> An expensive camera won't help with that. You need to know what
> you're doing no matter what camera you use. Try this though:
>
> http://commons.wikimedia.org/wiki/Image:Twilight3_-_NOAA.jpg
I need "sinister" dusk in an isolated summer mountain setting because
the book title is somber. I can't use any of the high resolution shots
(free or per charge) I've come across on the internet because they're
tropical, or too pretty, or too inspirational. Due to the help from
this site, I'm going to adjust the settings on my HP and go out into
the freaking woods myself.
Arthur Small
why you can not have your image printed to size, abet with a bit of a crop.
Send the image to a lab where they will print using a continuous tone rather
than a ink jet. From a 4MP image I have had 24 x 36 inch prints made that
look great.
Jules Vide
>
> Or, you could commission someone to shoot it for you... ;)
Yes, I certainly could, and it may come to that. On another branch of
this thread, someone suggested the impoverished quality of the print I
got from the 300 interpolation is due to the printer. I'm going to
take the file to a specialist and see that sort of effect I get at a
pricier developer's.
ASAAR
> This (and your entire excellent post) raises the question Why? I
> remember my Dad's Roloflex, looking at those squares and thinking they
> were a pain. If I had known *ANY* of what I've learned on this thread,
> I would never have bought my HP 4.1. I would have waited until I was
> able to afford a digital camera in the $1000 range.
>
> I badly need a 300 DPI photograph that will not look like something
> taken by someone needing cataract surgery, to fill a 6 x 9 inch
> paperback cover. I was resigned on foreshortening the photograph for
> reasons of economy, time, and annoyance to something 7 x 5 (landscape),
> and filling the rest of the cover with a solid black background.
>
> But after reading your (and others') posts, I'm getting indignant about
> what I now see was a misleading, VERY misleading, sales pitch by the dp
> industry in general: that for a hundred bucks-and-change, I'd be able
> to take pictures that would look like pictures taken from that old
> Roloflex or Nikon. The 4.1 I have is a *toy*. It should have been
> advertised as such.
No, you're mistaken about this. Your camera is capable of
producing excellent 300dpi prints as long as the sizes aren't very
large. Even 200 dpi prints will look very good. The problem you've
seen is due to trying to get a print that was based on a very small,
limited resolution piece of clip art. Its dimensions were 600 x 402
pixels, and that's what you'd get if your camera's sensor was
limited to 0.25 megapixels. Your camera's sensor has 16 times that
number of pixels. What this means is that if the clip art looks
very good when printed at a size of only about 2 inches, your camera
could produce 8" x 10" prints that look equally good, without the
pixellation or blocks that you saw when the clip-art was enlarged.
Most DSLRs have 6mp or 8mp sensors, and that's not much larger
than your camera's 4mp sensor. I have some of those old Nikons, and
my 4mp Fuji camera produces prints every bit as good as what I used
to get from them, but without the scratches and other problems
caused by not wanting to use very expensive professional photo labs.
>> Another problem would occur with viewfinders that show almost
>> exactly what will appear in the final image. This is usually the
>> case when using an electronic viewfinder (EVF) or the LCD display.
>> In this case, if the subject is framed perfectly (meaning as large
>> as possible) in the viewfinder, then since the horizontal and
>> vertical dimensions (or aspect ratio) of the sensor is unlikely to
>> match the aspect ratio of the print, something is bound to get
>> chopped off unless unnaturally large borders are used for the print.
>
> Again, the question is WHY? Is it to promote the sale/need of photo
> editing software? If I'm missing something, please tell me: is it
> cheaper to make a camera with 4:3 ratio than 3:2?
No, it has nothing to do with trying to drive up sales of
software. The reason for it is to keep from wasting what you can
get out of your lens. Think about it. Your lens is circular, and
the image it produces isn't of uniform quality. The greater the
distance from the center of the image, the more the image suffers.
If you want to limit the degradation to a certain amount, that will
define the size of a circular crop of the lens's image. But
producing circular pictures isn't economical or desirable. So given
a particular size circle, what type of rectangle should be selected
that will fit inside the circle that uses the greatest number of
pixels? Unfortunately, it's a square, which has a 1:1 aspect ratio,
and which most people wouldn't think makes very attractive prints.
The other popular print sizes, (with their aspect ratios) are:
8 x 10 (1.25)
11 x 14 (1.27)
4 x 6 (1.50)
This last one, 4 x 6, is the same aspect ratio you get with
standard 35mm film. Since it has the largest aspect ratio, the
rectangular part of the image uses the smallest area of these three
sizes for any particular circular portion of the lens's image, so it
"wastes" the most of what the lens is able to produce for a given,
or specified image quality. The 4:3 sensors (1.33 aspect ratio) is
a compromise, and minimizes the maximum amount of cropping that
would be necessary. For example, if a 3:2 sensor (same as 35mm
film) was used to make 8" x 10" prints, the amount of crop loss
would be greater than if 4:3 sensor had been used. This is nothing
new with digital cameras. With film, cameras were available that
had aspect ratios ranging from 1:1 to more extreme than 3:2. In
fact, some of the latest digital cameras are starting to provide a
larger range of aspect ratios. The 8mp Panasonic Lumix DMC-LX1 (aka
Leica D-Lux 2) allows you to take pictures with any of 3 aspect
ratios, 3:2, 4:3, and the new HD format, 16:9, which would be good
for landscapes and other situations where very wide shots are
needed. But these will cost much more than your HP, with list
prices of $600 for the Panasonic, and $800 for the Leica version.
> All I know is that "most people" is a euphemism (at least I think so).
> When the HP rep told me to use PhotoPlus software, for example (which
> contains the most well-written user "manual" of any photo editing
> software I've come across), I was hit in the face with SO many new
> terms you wouldn't find even two years ago--"anti-aliasing";
> "bicubic"--for the average consumer who has to work for a living and
> who doesn't have the time, energy, or ability to become an
> engineering/graphics master, there remains only one feasible option if
> you want to be part of the inescapable world of digital photography:
> pony up $1000+ for the proper camera.
I'll have to use that term again. Most people don't have to worry
about anti-aliasing or any of the other special features in photo
software. They're there either for the pros, or for the amateurs or
hobbyists that enjoy learning to take advantage of that sort of
thing. Whether it's an HP rep. or a camera salesman, you shouldn't
really expect them to be experts about what is required by the high
end publishing industry. If they were, they'd probably have told
you that you'd need to spend several thousands of dollars (or more)
on some decent DSLR equipment. Would you have been willing to pay
that much? Probably not. So you asked a limited question (about
the ability to make prints at 300 dpi quality) and got a reasonable
answer. Your HP really can produce 300 dpi prints. But you
shouldn't be indignant about the DP industry tricking you into
getting a "toy" camera. It's not a toy. It can produce what you
need. You might have to spend some time mastering photo editing
software to get the most out of your camera, but you shouldn't
expect the DP industry to go into great detail about what all of the
different cameras can produce. To do that honestly and accurately
would fill a book, and "most people" would rather give up on
photography than have to take an advanced course to learn everything
they need to know to choose an ideal camera.
> And I'm still trolling the web for a 300 DPI picture of a
> "sinister"-looking dusk...
Maybe something with a couple of mounted Nazgul flying through a
partially moonlit evening sky? <g>
> Thanks again, ASAAR. You make very interesting reading.
You're welcome.