Sign in with
Sign up | Sign in
Your question

Limits in print sharpness

Last response: in Digital Camera
Share
Anonymous
July 2, 2005 1:50:34 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Right up front I will say that I don't believe we always need the
ultimate in print sharpness for the print to look great, I am however
interesting in what it take to get the sharpest print possible.

I am going to limit the prints to ones that are output on the laser jet
type printers, the printer that output digital files to photographic
paper, it would appear that these print at 300 DPI.

So what do I need to get the sharpest print from one of this printers.
One might assume that there would be no need to take in photo that
would end up printing higher then 300 dpi, since the printer would just
have to down sample the image anyway. For example if I am going to
print out at 4 x 6 inches would I ever get a shaper print using more
then 1200 x 1800 pixels? The answer is more complicated then you might
think, and it all depends on how you got your 1200 x 1800 pixels. If
you got them from a camera that had 1200 x 1800 pixels as it max
resolution then you would find that a photo taken with a good camera
that had more pixels (say 2400 x 3600) would produce a sharper print.
But you could also take your 2400 x 3600 image and down sample it
yourself and get a 1200 x 1600 image that would print just as sharp and
maybe sharper then the original 2400 x 3600 image.

The idea here is very simple, a 1200 x 1800 pixel image from a 2.1 MP
camera will not be as sharp as one taken with a much higher MP camera
and down sampled.

There are a lot of ways of looking at this but the most useful is
thinking of it as the contrast for the smallest detail that will be in
the prints. A image that has not been down sampled will be lower
contrast in the fine pixel level detail then a photo that has been down
sampled.

Here is an example of a photo that has not been down sampled and one
that has, it is as an animated gif file that will blink back and fore
between the two views. Both images were taken with a 20D using the kit
lens, the better looking image was taken by zooming in and then down
sampling the image to match the scale of the first photo.
http://www.sewcon.com/photos/blink.gif

This is a very dramatic view that not all pixels are the same.

When we are dealing with this smallest detail the eye has a hard time
seeing to that level, to make out anything it needs the highest
possible contrast.

So if I really care about getting the very sharpest looking prints from
say Costco I need to down sample my images some to get the pixels
looking their best. This is not to same that the smaller print will
look better then a larger one made without down sampled pixels and the
larger print will have more of the photograph's detail visible in it,
but the sharpness when looking at the photograph will be the best when
I down sample.

So how much do I need to down sample to get to this point? It depends
on the camera and the lens I use. I need to down sample a move for my
F828 then I would for the 20D using a prime lens. You can tell when you
don't need to down sample any more when the pixels stop looking
clearer as you go smaller.

Now again I should say that this level of printing only add a small
amount to the apparent sharpness of a photo and is not normally needed,
but if one is after the best then you need pixels that are as high
contrast in the fine detail as you can get.

So how large of a photo can you print once you have down sampled
enough, on the 20D using a prime lens I judge that I need to down
sample to about 70% of the original to get really good looking pixels.
On the F828 it is more like 50% to 60%. For film, at least the film I
have shoot I need to do a huge amount of down sampling. What I come
up with is that for the 20D and a good lens I might get a print that is
around 6 x 8 inches and have the highest quality pixels possible. For
35mm film it is closer to a 4 x 6 inch print

If you really want to print to this level, and again there is little
need to, then this is what I believe you will find, at great digital
camera like the 1Ds Mark II will allow a near perfect 8 x 12 print. A
good DSLR, like the 20D will allow for a perfect 6 x 8 prints. A 35mm
camera will allow for a very small print in the range of a 4 x 6. What
is film so much smaller, it does not pick up contrast as fast as
digital photos do when down sampling. The MTF curve for film drops off
slower then for digital and to get to the really high contrast part you
have to through away a lot of pixels. If you really want the ultimate
print that is also large then thing about a 4 x 5 camera, if a 35mm
frame can make a 4 x 6 print then the 4 x 5 with the same magnification
would be able to produce a near perfect print at 17 x 21 inches.

Now don't write in all mad saying that you get great prints made much
larger then was I has listed here, so do I. These sizes are for the
cases where the ultimate in print clarity is trying to be achieved,
which is rarely needed but fun to see.

Scott

More about : limits print sharpness

Anonymous
July 2, 2005 11:18:56 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Scott W wrote:

<Very Huge Rambling Snip>

Yawn.

Just use good quality low speed film and the better Zeiss & Leitz lenses
(or others for LF) and you'll have no worries...

Don't worry - Be happy...
July 3, 2005 1:21:23 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Dude, you're using a laser printer.
Snap out of it!
Get a life!
Related resources
Can't find your answer ? Ask !
Anonymous
July 3, 2005 3:16:36 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

frederick wrote:
> Scotty,
>
> http://www.geocities.com/angels2000photos/print.jpg
> The image is from a 6mp dslr with a zoom lens.
> The print is from an inkjet printer at 10x6.7 inches.
>
> Your suggestion to downscale images to increase sharpness before
> printing is just plain daft.
> Within your suggested limit of 300dpi for printing, and a 10/6.7 or 10x8
> print size, then can you show me a scan of a crop from any print using
> any process from an image taken with any camera - that retains the
> detail of my sample (which was printed within 5 minutes of taking the
> snapshot, and with almost no post processing)? No? I thought not.
> It is perhaps possible, but I don't think you have a chance.

Well I have taken a fair bit of heat on this one, and it is not doubt
because I did not explain with any degree of clarity what I was looking
at.

First off I have not suggested that people down sample their photos
when printing.

The limit of 300 dpi is imposed with many of the printer that use laser
to print to photographic paper, just as the Fuji Frontier. This is not
a limit I put on but rather one that anyone who uses one of these
prints is limited to.

The question that I was looking at was, what is the limit for sharpness
when printing on a printer like the Fuji. The criteria is not how much
detail can be seen in a give print, the large prints will always win
on this, but rather how sharp is a the print itself be.

I also stated that I don't think it is either necessary to even
desirable to push prints to this level.

What I did say is that you can get prints that look just a bit sharper
when doing this.

I have as an example a 4 x 6 print made from a 5 MP camera, here is a
scan of the whole photo.
http://www.sewcon.com/photos/testphoto.jpg

Your scan of the print as I take it was at 2400 then down sampled to
600 dpi.
Here I scanned my print at 1600 and down sampled to 600 dpi, I then put
the two images side by side.
http://www.sewcon.com/photos/compare.jpg

The images are so blown up that it is hard to get a feel for them when
looking at the computer screen, when view from about 8 to 10 feed back
you will get a better feel for how the prints might look.

I believe that my print is a bit sharper, your print will have more
total detail because of its large size but for sharpness mine has a bit
of an edge.

Now you asked a different question, you asked about retaining the
detail of the original photo, here making a larger print will keep
more of the detail of the original photo, but the print will look just
a bit softer. This is the trade off we are faced with when printing, a
print that has all the detail of the original or a print that is as
sharp as it can be. You can come close to having both but you can get
both at the same time.

No matter how many disclaimers I make that I am not trying to tell
people that they need to down size their photos that is somehow the
message that they are hearing. So let me try to be clear, I think you
print looks great, I think that any sharpness gains made by downsizing
would not be worth the smaller print.

All that I was really interested in was how sharp can a print be, when
using the kind of printer that Costco and many others use?

Scott
Anonymous
July 3, 2005 9:16:45 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

frederick wrote:
> You need to think again.
> I don't know how you did your maths, but you have shown me a scan of an
> approximately 1 inch x 1 inch crop from your 6 x 4 image.

Well you are right, something is not adding up. I took your scan of the
eye which is a bit over 600 x 600 pixels.

The text on the scan said 2400dpi downscaled to 4:1, which then show
give us an image that is at 600 dpi.

This should leave us with am image that is just over 1 inch square.

But in looking at the over all photo and taking the size you said the
print was at it would appear that the scan of the eye should be
smaller.

Scott
Anonymous
July 3, 2005 2:28:53 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Don Stauffer wrote:
> I have both a laser printer (monochrome) and a good Canon inkjet. The
> inkjet is far superior to the laser even when printing monochrome
> prints. Undoubtedly the much higher resolution of the Canon inkjet is
> the main factor, but I find the tonal qualities of the inkjet blacks
> nicer than the Brothers laser printer also.

The laser printers we are talking about are the ones that print on
photographic paper, like the Fuji Frontier.

Scott
Anonymous
July 3, 2005 3:46:41 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

I have both a laser printer (monochrome) and a good Canon inkjet. The
inkjet is far superior to the laser even when printing monochrome
prints. Undoubtedly the much higher resolution of the Canon inkjet is
the main factor, but I find the tonal qualities of the inkjet blacks
nicer than the Brothers laser printer also.

Scott W wrote:
> Right up front I will say that I don't believe we always need the
> ultimate in print sharpness for the print to look great, I am however
> interesting in what it take to get the sharpest print possible.
>
> I am going to limit the prints to ones that are output on the laser jet
> type printers, the printer that output digital files to photographic
> paper, it would appear that these print at 300 DPI.
>
> So what do I need to get the sharpest print from one of this printers.
> One might assume that there would be no need to take in photo that
> would end up printing higher then 300 dpi, since the printer would just
> have to down sample the image anyway. For example if I am going to
> print out at 4 x 6 inches would I ever get a shaper print using more
> then 1200 x 1800 pixels? The answer is more complicated then you might
> think, and it all depends on how you got your 1200 x 1800 pixels. If
> you got them from a camera that had 1200 x 1800 pixels as it max
> resolution then you would find that a photo taken with a good camera
> that had more pixels (say 2400 x 3600) would produce a sharper print.
> But you could also take your 2400 x 3600 image and down sample it
> yourself and get a 1200 x 1600 image that would print just as sharp and
> maybe sharper then the original 2400 x 3600 image.
>
> The idea here is very simple, a 1200 x 1800 pixel image from a 2.1 MP
> camera will not be as sharp as one taken with a much higher MP camera
> and down sampled.
>
> There are a lot of ways of looking at this but the most useful is
> thinking of it as the contrast for the smallest detail that will be in
> the prints. A image that has not been down sampled will be lower
> contrast in the fine pixel level detail then a photo that has been down
> sampled.
>
> Here is an example of a photo that has not been down sampled and one
> that has, it is as an animated gif file that will blink back and fore
> between the two views. Both images were taken with a 20D using the kit
> lens, the better looking image was taken by zooming in and then down
> sampling the image to match the scale of the first photo.
> http://www.sewcon.com/photos/blink.gif
>
> This is a very dramatic view that not all pixels are the same.
>
> When we are dealing with this smallest detail the eye has a hard time
> seeing to that level, to make out anything it needs the highest
> possible contrast.
>
> So if I really care about getting the very sharpest looking prints from
> say Costco I need to down sample my images some to get the pixels
> looking their best. This is not to same that the smaller print will
> look better then a larger one made without down sampled pixels and the
> larger print will have more of the photograph's detail visible in it,
> but the sharpness when looking at the photograph will be the best when
> I down sample.
>
> So how much do I need to down sample to get to this point? It depends
> on the camera and the lens I use. I need to down sample a move for my
> F828 then I would for the 20D using a prime lens. You can tell when you
> don't need to down sample any more when the pixels stop looking
> clearer as you go smaller.
>
> Now again I should say that this level of printing only add a small
> amount to the apparent sharpness of a photo and is not normally needed,
> but if one is after the best then you need pixels that are as high
> contrast in the fine detail as you can get.
>
> So how large of a photo can you print once you have down sampled
> enough, on the 20D using a prime lens I judge that I need to down
> sample to about 70% of the original to get really good looking pixels.
> On the F828 it is more like 50% to 60%. For film, at least the film I
> have shoot I need to do a huge amount of down sampling. What I come
> up with is that for the 20D and a good lens I might get a print that is
> around 6 x 8 inches and have the highest quality pixels possible. For
> 35mm film it is closer to a 4 x 6 inch print
>
> If you really want to print to this level, and again there is little
> need to, then this is what I believe you will find, at great digital
> camera like the 1Ds Mark II will allow a near perfect 8 x 12 print. A
> good DSLR, like the 20D will allow for a perfect 6 x 8 prints. A 35mm
> camera will allow for a very small print in the range of a 4 x 6. What
> is film so much smaller, it does not pick up contrast as fast as
> digital photos do when down sampling. The MTF curve for film drops off
> slower then for digital and to get to the really high contrast part you
> have to through away a lot of pixels. If you really want the ultimate
> print that is also large then thing about a 4 x 5 camera, if a 35mm
> frame can make a 4 x 6 print then the 4 x 5 with the same magnification
> would be able to produce a near perfect print at 17 x 21 inches.
>
> Now don't write in all mad saying that you get great prints made much
> larger then was I has listed here, so do I. These sizes are for the
> cases where the ultimate in print clarity is trying to be achieved,
> which is rarely needed but fun to see.
>
> Scott
>
Anonymous
July 3, 2005 4:11:41 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"Scott W" <biphoto@hotmail.com> wrote in message
news:1120371396.339599.34040@g47g2000cwa.googlegroups.com...
[SNIP]

> This is the trade off we are faced with when printing, a
> print that has all the detail of the original or a print that is as
> sharp as it can be. You can come close to having both but
> you can get both at the same time.

This sounds very like you are mistaking accutance for sharpness.

Detail and sharpness are never mutually incompatible, but detail
(resolution) _is_ diminished by adding artificial accutance - the
'sharpening' that one gets with USM is not actually sharpening, it's
increasing accutance. Your downsampling is increasing accutance, which you
perceive as 'sharpness'.


Peter
July 3, 2005 9:14:55 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Scotty,

http://www.geocities.com/angels2000photos/print.jpg
The image is from a 6mp dslr with a zoom lens.
The print is from an inkjet printer at 10x6.7 inches.

Your suggestion to downscale images to increase sharpness before
printing is just plain daft.
Within your suggested limit of 300dpi for printing, and a 10/6.7 or 10x8
print size, then can you show me a scan of a crop from any print using
any process from an image taken with any camera - that retains the
detail of my sample (which was printed within 5 minutes of taking the
snapshot, and with almost no post processing)? No? I thought not.
It is perhaps possible, but I don't think you have a chance.
July 3, 2005 9:14:56 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"frederick" <nomail@nomail.com> wrote
> Scotty,
> Your suggestion to downscale images to increase sharpness before printing
> is just plain daft.

You aren't going to fool us! The child in question is a subject very well
know for her extraordinarily, high-acutance eyelashes!

Yeah, that's the ticket.
Anonymous
July 3, 2005 9:34:15 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Roger N. Clark (change username to rnclark) wrote:
> Here is my page on print sharpness:
>
> Experiments with Pixels Per Inch (PPI) on Printed Image Sharpness
> http://www.clarkvision.com/imagedetail/printer-ppi
>
> It shows that a 600 ppi inkjet is sharper than a 300 ppi inkjet.
>
> Blind tests in very good light show people can see the difference,
> but as light levels drop, then they can't see a difference.
>
> I'll add to it by getting a lightjet print (305 ppi) and add that
> to the website.
>
> Roger

I did this same test with my ink jet and I could get a bit more when
going feeding it images higher then 300 dpi, but then I am pretty near
sighted and many others could not see the gain.

I have gotten people very confused with what I was trying to show,
which as not that 300 dpi was some ultimate resolution for printing.

What I was interested in was if I am limited to a printer that prints
at 300 (305?) dpi then what does it take to make the clearest print
possible. Here is where we get into the concept that not all pixel
carry the same level of information. Even a good Bayer camera will not
have great pixels, for a camera like my 20D to get the best pixels you
need to down sample by some amount.

Simply put I will not get as sharp a print if I have the prints from my
20D printed at 300 DPI, which would produce a print close to 12 x 8
inches. This is not to say the print would not look very, very sharp
but it would not be as sharp as I could get it if I down sampled my
photos and made a smaller print. This assume I am getting all my
prints done at Costco. The effects of a optical (laser) printer
running at higher resolution is another issue.

No matter how many time I stated that I did not think you needed this
level of sharpness I still had a lot people believing that I was
claiming that you did.

Scott
Anonymous
July 4, 2005 12:29:43 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Bandicoot wrote:
> "Scott W" <biphoto@hotmail.com> wrote in message
> news:1120371396.339599.34040@g47g2000cwa.googlegroups.com...
> [SNIP]
>
>
>>This is the trade off we are faced with when printing, a
>>print that has all the detail of the original or a print that is as
>>sharp as it can be. You can come close to having both but
>>you can get both at the same time.
>
>
> This sounds very like you are mistaking accutance for sharpness.
>
> Detail and sharpness are never mutually incompatible, but detail
> (resolution) _is_ diminished by adding artificial accutance - the
> 'sharpening' that one gets with USM is not actually sharpening, it's
> increasing accutance. Your downsampling is increasing accutance, which you
> perceive as 'sharpness'.
>
>
> Peter
>
>

Peter,
Well said. To increase sharpness one needs more sophisticated
algorithms, like Richardson-Lucey image deconvolution. Even then
it is a combination of sharpness, accutance, and an increase in
noise. An example:
http://www.clarkvision.com/imagedetail/image-restoratio...

Roger
July 4, 2005 12:37:05 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Scott W wrote:
> frederick wrote:
>
>>Scotty,
>>
>>http://www.geocities.com/angels2000photos/print.jpg
>>The image is from a 6mp dslr with a zoom lens.
>>The print is from an inkjet printer at 10x6.7 inches.
>>
>>Your suggestion to downscale images to increase sharpness before
>>printing is just plain daft.
>>Within your suggested limit of 300dpi for printing, and a 10/6.7 or 10x8
>>print size, then can you show me a scan of a crop from any print using
>>any process from an image taken with any camera - that retains the
>>detail of my sample (which was printed within 5 minutes of taking the
>>snapshot, and with almost no post processing)? No? I thought not.
>>It is perhaps possible, but I don't think you have a chance.
>
>
> Well I have taken a fair bit of heat on this one, and it is not doubt
> because I did not explain with any degree of clarity what I was looking
> at.
>
> First off I have not suggested that people down sample their photos
> when printing.
>
> The limit of 300 dpi is imposed with many of the printer that use laser
> to print to photographic paper, just as the Fuji Frontier. This is not
> a limit I put on but rather one that anyone who uses one of these
> prints is limited to.
>
> The question that I was looking at was, what is the limit for sharpness
> when printing on a printer like the Fuji. The criteria is not how much
> detail can be seen in a give print, the large prints will always win
> on this, but rather how sharp is a the print itself be.
>
> I also stated that I don't think it is either necessary to even
> desirable to push prints to this level.
>
> What I did say is that you can get prints that look just a bit sharper
> when doing this.
>
> I have as an example a 4 x 6 print made from a 5 MP camera, here is a
> scan of the whole photo.
> http://www.sewcon.com/photos/testphoto.jpg
>
> Your scan of the print as I take it was at 2400 then down sampled to
> 600 dpi.
> Here I scanned my print at 1600 and down sampled to 600 dpi, I then put
> the two images side by side.
> http://www.sewcon.com/photos/compare.jpg
>
> The images are so blown up that it is hard to get a feel for them when
> looking at the computer screen, when view from about 8 to 10 feed back
> you will get a better feel for how the prints might look.
>
> I believe that my print is a bit sharper, your print will have more
> total detail because of its large size but for sharpness mine has a bit
> of an edge.
>
You need to think again.
I don't know how you did your maths, but you have shown me a scan of an
approximately 1 inch x 1 inch crop from your 6 x 4 image.

So if you want to compare, then I'll show you your print compared with a
0.9 inch x 0.9 inch cropped scan from a 10x6.7 print of my image here:
http://www.geocities.com/angels2000photos/crop.jpg

I think your photo and print is not so very sharp - even though your
image has many more "pixels per inch" - to the point that at almost
double the print size that I have used, the enlargement of mine is so
great that pixellation is visible on screen.

To prove a point, I have reprinted mine at 6x4, and then scanned at 0.9
x 0.9 inch crop so all is fair. Same sized print, and I even gave you a
10% advantage - more than needed to compensate for 1mp less.
http://www.geocities.com/angels2000photos/6x4.jpg

That is the print you want to use for a comparison.
Still think yours is sharper?
Is mine sharper by just a "bit of an edge"?
heh.
Anonymous
July 4, 2005 12:37:06 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

frederick wrote:
reat that pixellation is visible on screen.
>
> To prove a point, I have reprinted mine at 6x4, and then scanned at 0.9
> x 0.9 inch crop so all is fair. Same sized print, and I even gave you a
> 10% advantage - more than needed to compensate for 1mp less.
> http://www.geocities.com/angels2000photos/6x4.jpg
>
> That is the print you want to use for a comparison.
> Still think yours is sharper?
> Is mine sharper by just a "bit of an edge"?
> heh.

Somewhere in this thread I've lost the plot.

What is it about?

Douglas
Anonymous
July 4, 2005 12:37:07 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"Ryadia" <just@the.group> wrote in message
news:42c7abd4$1@dnews.tpgi.com.au...
> frederick wrote:
> reat that pixellation is visible on screen.
> >
> > To prove a point, I have reprinted mine at 6x4, and then
> > scanned at 0.9 x 0.9 inch crop so all is fair. Same sized
> > print, and I even gave you a 10% advantage - more than
> > needed to compensate for 1mp less.
> > http://www.geocities.com/angels2000photos/6x4.jpg
> >
> > That is the print you want to use for a comparison.
> > Still think yours is sharper?
> > Is mine sharper by just a "bit of an edge"?
> > heh.
>
> Somewhere in this thread I've lost the plot.
>
> What is it about?
>

Buyer's remorse? Spousal spending guilt? ;-)



P.
July 4, 2005 1:20:48 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Ryadia wrote:
> frederick wrote:
> reat that pixellation is visible on screen.
>
>>
>> To prove a point, I have reprinted mine at 6x4, and then scanned at
>> 0.9 x 0.9 inch crop so all is fair. Same sized print, and I even gave
>> you a 10% advantage - more than needed to compensate for 1mp less.
>> http://www.geocities.com/angels2000photos/6x4.jpg
>>
>> That is the print you want to use for a comparison.
>> Still think yours is sharper?
>> Is mine sharper by just a "bit of an edge"?
>> heh.
>
>
> Somewhere in this thread I've lost the plot.
>
> What is it about?
>
> Douglas
lol - I think I might have too - but to be fair to you and I, I'm not
sure if the OP actually had a plot.
July 4, 2005 4:32:34 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"Scott W" <biphoto@hotmail.com> wrote in message
news:1120437255.105494.150520@g14g2000cwa.googlegroups.com...

> for a camera like my 20D to get the best pixels you
> need to down sample by some amount.

Downsampling is destructive. Perhaps something else is happening when you
downsample, such as a bit of unsharp mask. Do you know? Can you be sure?
July 4, 2005 2:06:11 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Scott W wrote:

>
> frederick wrote:
>
>>You need to think again.
>>I don't know how you did your maths, but you have shown me a scan of an
>>approximately 1 inch x 1 inch crop from your 6 x 4 image.
>
>
> Well you are right, something is not adding up. I took your scan of the
> eye which is a bit over 600 x 600 pixels.
>
> The text on the scan said 2400dpi downscaled to 4:1, which then show
> give us an image that is at 600 dpi.
>
> This should leave us with am image that is just over 1 inch square.
>
> But in looking at the over all photo and taking the size you said the
> print was at it would appear that the scan of the eye should be
> smaller.
>
> Scott
>

Somewhere our maths failed us... Not sure if it was you or me.

If you are interested in getting maximum sharpness from small prints,
then the 1 or 1.5 picolitre droplet printers like Epson R800, Canon
iP8500, (and HP8450?) will resolve down to about 1,000 ppi. You get
diminishing returns of course because the droplets smudge, but they can
resolve down to 10 lines per millimeter in practice, which is much
higher than Frontier prints at 300dpi, which could print 5 lpmm in
theory - if everything was perfect. What I've seen of Fuji Frontier at
300dpi seem a bit less sharp than those printers when you print an image
at 300 pixels per inch.
But you have to look very close to see a difference - a well printed
inkjet print at 300 pixels per inch size, or the same from a good lab
printed at 300dpi is as close to perfect as I think I would ever want.
Anonymous
July 4, 2005 2:06:12 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

frederick wrote:
> Scott W wrote:
>
>>
>> frederick wrote:
>>
>>> You need to think again.
>>> I don't know how you did your maths, but you have shown me a scan of an
>>> approximately 1 inch x 1 inch crop from your 6 x 4 image.
>>
>>
>>
>> Well you are right, something is not adding up. I took your scan of the
>> eye which is a bit over 600 x 600 pixels.
>>
>> The text on the scan said 2400dpi downscaled to 4:1, which then show
>> give us an image that is at 600 dpi.
>>
>> This should leave us with am image that is just over 1 inch square.
>>
>> But in looking at the over all photo and taking the size you said the
>> print was at it would appear that the scan of the eye should be
>> smaller.
>>
>> Scott
>>
>
> Somewhere our maths failed us... Not sure if it was you or me.
>
> If you are interested in getting maximum sharpness from small prints,
> then the 1 or 1.5 picolitre droplet printers like Epson R800, Canon
> iP8500, (and HP8450?) will resolve down to about 1,000 ppi. You get
> diminishing returns of course because the droplets smudge, but they can
> resolve down to 10 lines per millimeter in practice, which is much
> higher than Frontier prints at 300dpi, which could print 5 lpmm in
> theory - if everything was perfect. What I've seen of Fuji Frontier at
> 300dpi seem a bit less sharp than those printers when you print an image
> at 300 pixels per inch.
> But you have to look very close to see a difference - a well printed
> inkjet print at 300 pixels per inch size, or the same from a good lab
> printed at 300dpi is as close to perfect as I think I would ever want.

Here is my page on print sharpness:

Experiments with Pixels Per Inch (PPI) on Printed Image Sharpness
http://www.clarkvision.com/imagedetail/printer-ppi

It shows that a 600 ppi inkjet is sharper than a 300 ppi inkjet.

Blind tests in very good light show people can see the difference,
but as light levels drop, then they can't see a difference.

I'll add to it by getting a lightjet print (305 ppi) and add that
to the website.

Roger
Anonymous
July 4, 2005 4:20:08 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

In message <1120323034.654849.295420@g44g2000cwa.googlegroups.com>,
"Scott W" <biphoto@hotmail.com> wrote:

>So if I really care about getting the very sharpest looking prints from
>say Costco I need to down sample my images some to get the pixels
>looking their best. This is not to same that the smaller print will
>look better then a larger one made without down sampled pixels and the
>larger print will have more of the photograph's detail visible in it,
>but the sharpness when looking at the photograph will be the best when
>I down sample.

I don't think so. What you are calling sharpness, I would call
pixelation, and while not perceived as softness per se, I would call it
"fuzz".

It fools some people, but not others.

My recent experience has been with Canon printers (since I have been
shooting digital), and I always get the sharpest prints, by my
definition, with all of the original pixels, no matter how soft they may
be.

Some people have suggested that printers use something like the nearest
neighbor algorithm if you try to print above a certain PPI, but I don't
belive that (not for my printer, anyway), and would try it and see,
instead. I use error-diffusion and just give it all the pixels I have.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
Anonymous
July 4, 2005 4:26:33 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

In message <1120382379.360451@ftpsrv1>,
frederick <nomail@nomail.com> wrote:

>lol - I think I might have too - but to be fair to you and I, I'm not
>sure if the OP actually had a plot.

I think what Scott is trying to say is that if the printer firmware or
driver is going to resample an image, you might be better off resampling
it yourself.

That may be a good idea if the printer does that. Whether or not it
does is another question. I know for a fact that it happenhs with
dye-subs, but for inkjets/bubblejets I am not so sure, especially if
error diffusion is used.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
July 4, 2005 4:56:04 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

JPS@no.komm wrote:

> In message <1120382379.360451@ftpsrv1>,
> frederick <nomail@nomail.com> wrote:
>
>>lol - I think I might have too - but to be fair to you and I, I'm not
>>sure if the OP actually had a plot.
>
> I think what Scott is trying to say is that if the printer firmware or
> driver is going to resample an image, you might be better off resampling
> it yourself.
>


If he was, no one would have argued. It seems he was trying to say we should
downsample ~20% then upsample to print size which seems very wrong to me.
Like throw out non-perfect real information, then create false information
to replace it with?

--

Stacey
Anonymous
July 4, 2005 4:56:13 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Does anybody here have a clue about what they are talking about?
From all the banter, I doubt very much any of you actually do.
A 10D image printed at 20"x30" using my enlargement process is sharper
than an 8x11 print made from an ISO 100, 35mm negative.

Although I post as Ryadia, the name of my studio, I am Douglas MacDonald
and my company: Technology Australia Pty Ltd ...has pioneered
Interpolation enlargement of digital images with stunning detail and
extreme sharpness for several years.

Alan Browne and Gordon Moate - to name drop just two regulars from this
group who have received 24"x 36" photographs and the original 10D files
to compare the claims I make as part of my promotional publicity last year.

http://www.technoaussie.com/images/examples.jpg shows some large posters
I printed from Nikon files of 2.8 mega pixels, shot at ISO 400. Provide
me with a full size image file and I can produce 6 feet wide photographs
from them: http://www.technoaussie.com/big-prints.htm for more information.

On the subject of sharpness. Continuous tone printers such as 'Lambda'
laser printers can produce clearer, sharper images at lower resolution
than so called "digital" imaging printers like Fuji's Fronterier. In
fact Dye Sublimation printers can produce sharper images at 125 DPI than
Digital machines at 300 DPI produce.

The fallacy about print resolution and and image size needs to be laid
to rest. All domestic, Inkjet printers which use the Windows interface
for their printer drivers (Epson, Canon, HP, Lexmark etc)interpolate the
printing data either up or down in resolution during the print process
so it matches the printer's native resolution and produces the "most
effective print".

It's not until you get into raster image processing of mathematically
correct print data such as Postscript files that you gain any real
control over a printer and need to be aware of resolution dependencies.
Until then, you have no control over print resolution and sending your
Epson (or any other brand) printer 700 dpi images or 150 dpi image will
not materially alter the print you make.

My small HP photo plotter runs happily at 180 DPI. The Sherpa needs 300
dpi to produce the same clarity prints and the laser takes photos at any
resolution from 75 DPI to 1400 DPI and prints them all at 300 DPI,
regardless of the size of the photograph!

Douglas
Anonymous
July 4, 2005 4:56:14 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"doug" <nospam@this.com> wrote

> Does anybody here have a clue about what they are talking about?

No. If we did we wouldn't be wasting our time talking about it.

--
Nicholas O. Lindan, Cleveland, Ohio
Consulting Engineer: Electronics; Informatics; Photonics.
To reply, remove spaces: n o lindan at ix . netcom . com
Fstop timer - http://www.nolindan.com/da/fstop/index.htm
July 4, 2005 4:56:14 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"doug" <nospam@this.com> wrote in message
news:42c8a55a$1@dnews.tpgi.com.au...
> Does anybody here have a clue about what they are talking about?
> From all the banter, I doubt very much any of you actually do.
> A 10D image printed at 20"x30" using my enlargement process is sharper
> than an 8x11 print made from an ISO 100, 35mm negative.

Physics must be different in your part of the universe.
Anonymous
July 4, 2005 4:56:15 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

johnboy wrote:
> "doug" <nospam@this.com> wrote in message
> news:42c8a55a$1@dnews.tpgi.com.au...
>
>>Does anybody here have a clue about what they are talking about?
>>From all the banter, I doubt very much any of you actually do.
>>A 10D image printed at 20"x30" using my enlargement process is sharper
>>than an 8x11 print made from an ISO 100, 35mm negative.
>
>
> Physics must be different in your part of the universe.
>
>

He didn't say what camera he used for the 35mm print. The camera has as
much (or more) effect than the film. A cheap enough 35 mm camera won't
make very good 8 x 10 prints :-)
July 4, 2005 5:40:34 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Roger N. Clark (change username to rnclark) wrote:

>
> Here is my page on print sharpness:
>
> Experiments with Pixels Per Inch (PPI) on Printed Image Sharpness
> http://www.clarkvision.com/imagedetail/printer-ppi
>
> It shows that a 600 ppi inkjet is sharper than a 300 ppi inkjet.
>
> Blind tests in very good light show people can see the difference,
> but as light levels drop, then they can't see a difference.
>
> I'll add to it by getting a lightjet print (305 ppi) and add that
> to the website.
>
> Roger

Thanks - that's interesting. It would be interesting to see a lightjet
print resolving your test chart at 150lpi.

I did a quick test at 300dpi and 600dpi. I can see the 1 pixel lines at
300dpi very clearly, but not at 600dpi. My (not leading edge) scanner
couldn't resolve it either - no better than my old eyes could.
But taking a macro photo of the print, I can see at 600dpi (300lpi) that
the horizontal lines are quite well resolved at 300lpi - but the
vertical lines not so clearly and losing contrast (not as much as my
photo indicates). I didn't try at 600 lpi, but doubt it would resolve
the 1 pixel lines at all in either direction. That was with an R1800
on premium semi-gloss paper using all default settings. Premium gloss
paper might be better (I don't have any - because I don't like it).
http://www.geocities.com/angels2000photos/print300.jpg
http://www.geocities.com/angels2000photos/print600.jpg
Anonymous
July 4, 2005 6:53:49 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"johnboy" <okaynow@nospam.no> wrote:
> "Scott W" <biphoto@hotmail.com> wrote:
>
>> for a camera like my 20D to get the best pixels you
>> need to down sample by some amount.

Well, not by much. Both Canon and Nikon noticed how badly the 100D, with its
adequate AA filter, went over, and have been using less than adequate AA
filters ever since. That means that there's significant detail quite close
to the Nyquist (when there isn't Moire or Bayer demosaicing disasters). So
you can't downsample by much. Downsampling can create aliased images that
are problematic.

A far more reasonable thing is to simply print them at higher dpi settings
on current inkjets, which can render detail well over 300 dpi. The 11MP
Canon 1Ds images look real nice at 8x12, about 335 dpi. A lot nicer than 6MP
images.

> Downsampling is destructive.

Downsampling is not necessarily destructive. If the image has no information
above a certain frequency, then you can downsample to a size such that said
frequency is just below the Nyquist frequency of the smaller image with no
loss of information.

Downsampling is also not necessarily significantly destructive. If all the
information above a certain frequency in the image is of too low contrast to
significantly contribute to the image, then you can downsample to a size
such that said frequency is just below the Nyquist frequency of the smaller
image with no significant loss of information.

ScottW had a good point in an earlier post. Take your favorite 4000 dpi (or
5400 dpi) scan and downsample it (bicubic in Photoshop) by a factor of two
and then upsample the result. Compare that to the original. You'll note that
the two are essentially indistinguishable. (This points out that scans are
grossly oversampled.)

David J. Littleboy
Tokyo, Japan
Anonymous
July 4, 2005 7:18:27 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

<JPS@no.komm> wrote in message
news:thaic1hfv8kgqhh5vrgvjkaq253oi58cm0@4ax.com...
SNIP
> I think what Scott is trying to say is that if the printer firmware
> or driver is going to resample an image, you might be better
> off resampling it yourself.

Yes, and sharpen after that to compensate for the resampling losses.

> That may be a good idea if the printer does that. Whether or
> not it does is another question. I know for a fact that it
> happenhs with dye-subs, but for inkjets/bubblejets I am not so
> sure, especially if error diffusion is used.

Yes they do resample to a predefined resolution (based on paper choice
and quality settings). The printer driver can feed-back that 'native'
resolution to the printing program, which could be used to optimize
the print data (resample and re-sharpen) before sending it to the
printer.

In fact Qimage (http://www.ddisoftware.com/qimage/) does just that,
all based on a single original file, it resamples (usually to 600 or
720 ppi, depending on IJ manufacturer) and sharpens on-the-fly. That
eliminates the need for different versions of the original for
different sized output. My output has never looked better with so
little effort (and for a modest price).

Bart
July 4, 2005 9:39:13 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Scott W wrote:
> frederick wrote:
>
>>Scotty,
>>
>>http://www.geocities.com/angels2000photos/print.jpg
>>The image is from a 6mp dslr with a zoom lens.
>>The print is from an inkjet printer at 10x6.7 inches.
>>
>>Your suggestion to downscale images to increase sharpness before
>>printing is just plain daft.
>>Within your suggested limit of 300dpi for printing, and a 10/6.7 or 10x8
>>print size, then can you show me a scan of a crop from any print using
>>any process from an image taken with any camera - that retains the
>>detail of my sample (which was printed within 5 minutes of taking the
>>snapshot, and with almost no post processing)? No? I thought not.
>>It is perhaps possible, but I don't think you have a chance.
>
>
> Well I have taken a fair bit of heat on this one, and it is not doubt
> because I did not explain with any degree of clarity what I was looking
> at.
>
> First off I have not suggested that people down sample their photos
> when printing.
>
> The limit of 300 dpi is imposed with many of the printer that use laser
> to print to photographic paper, just as the Fuji Frontier. This is not
> a limit I put on but rather one that anyone who uses one of these
> prints is limited to.
>
> The question that I was looking at was, what is the limit for sharpness
> when printing on a printer like the Fuji. The criteria is not how much
> detail can be seen in a give print, the large prints will always win
> on this, but rather how sharp is a the print itself be.
>
> I also stated that I don't think it is either necessary to even
> desirable to push prints to this level.
>
> What I did say is that you can get prints that look just a bit sharper
> when doing this.
>
> I have as an example a 4 x 6 print made from a 5 MP camera, here is a
> scan of the whole photo.
> http://www.sewcon.com/photos/testphoto.jpg
>
> Your scan of the print as I take it was at 2400 then down sampled to
> 600 dpi.
> Here I scanned my print at 1600 and down sampled to 600 dpi, I then put
> the two images side by side.
> http://www.sewcon.com/photos/compare.jpg
>
> The images are so blown up that it is hard to get a feel for them when
> looking at the computer screen, when view from about 8 to 10 feed back
> you will get a better feel for how the prints might look.
>
> I believe that my print is a bit sharper, your print will have more
> total detail because of its large size but for sharpness mine has a bit
> of an edge.
>
> Now you asked a different question, you asked about retaining the
> detail of the original photo, here making a larger print will keep
> more of the detail of the original photo, but the print will look just
> a bit softer. This is the trade off we are faced with when printing, a
> print that has all the detail of the original or a print that is as
> sharp as it can be. You can come close to having both but you can get
> both at the same time.
>
> No matter how many disclaimers I make that I am not trying to tell
> people that they need to down size their photos that is somehow the
> message that they are hearing. So let me try to be clear, I think you
> print looks great, I think that any sharpness gains made by downsizing
> would not be worth the smaller print.
>
> All that I was really interested in was how sharp can a print be, when
> using the kind of printer that Costco and many others use?
>
> Scott
>
There is a limit to the resolution that is possible on paper, even the best photo paper.
It is possible, of course, to do worse than that limit.
Anonymous
July 6, 2005 11:54:24 AM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

doug wrote:
> Does anybody here have a clue about what they are talking about?

Inimitable!

[That'd be what *he is* talking about.]

> From all the banter, I doubt very much any of you actually do.

That's why we let you post here!

> It's not until you get into raster image processing of mathematically
> correct print data such as Postscript files that you gain any real
> control over a printer and need to be aware of resolution dependencies.
> Until then, you have no control over print resolution and sending your
> Epson (or any other brand) printer 700 dpi images or 150 dpi image will
> not materially alter the print you make.

IAE, in the above example, there has to be an optimum size to send to
the Epson. Note I am not claiming a material or immaterial improvement
in the quality of the print. Maybe it just prints sooner or faster.
Maybe it's all moot, but still there has to be an optimum in what the
printer receives.

--
John McWilliams

Coach: "Are you just ignorant, or merely apathetic?"
Player: "Coach, I don't know, and I don't care."
Anonymous
July 6, 2005 8:28:03 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

On Wed, 06 Jul 2005 07:54:24 -0700, John McWilliams
<jpmcw@comcast.net> wrote:

>IAE, in the above example, there has to be an optimum size to send to
>the Epson. Note I am not claiming a material or immaterial improvement
>in the quality of the print. Maybe it just prints sooner or faster.
>Maybe it's all moot, but still there has to be an optimum in what the
>printer receives.

Epson printer drivers communicate the image to the printer at 720ppi,
Canon drivers work at 600ppi. There is little point exceeding these
resolutions when sending a job to the printer, but I doubt there is
much harm either.

Anyway, you shouldn't expect to see all the detail that might be
present in the 720ppi image when it gets printed, especially if colors
or shades of gray are concerned (and they *usually* are).

What's the minimum ppi you can get away with and still get a decent
print? Well, that depends on viewing distance and how picky you are,
it *could* be as low as 50ppi.

--
Owamanga!
http://www.pbase.com/owamanga
Anonymous
July 6, 2005 8:28:04 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

Owamanga wrote:
> On Wed, 06 Jul 2005 07:54:24 -0700, John McWilliams
> <jpmcw@comcast.net> wrote:
>
>
>>IAE, in the above example, there has to be an optimum size to send to
>>the Epson. Note I am not claiming a material or immaterial improvement
>>in the quality of the print. Maybe it just prints sooner or faster.
>>Maybe it's all moot, but still there has to be an optimum in what the
>>printer receives.
>
>
> Epson printer drivers communicate the image to the printer at 720ppi,
> Canon drivers work at 600ppi. There is little point exceeding these
> resolutions when sending a job to the printer, but I doubt there is
> much harm either.

But does it follow that either value is the correct way to send out of
PS to those respective printer drivers?

Does the driver convert my 300 ppi image to 720 ppi before converting it
to 360, 720 or 1440 dpi?
>
> Anyway, you shouldn't expect to see all the detail that might be
> present in the 720ppi image when it gets printed, especially if colors
> or shades of gray are concerned (and they *usually* are).
>
> What's the minimum ppi you can get away with and still get a decent
> print? Well, that depends on viewing distance and how picky you are,
> it *could* be as low as 50ppi.

Oh, heck yeah and all. A 10 x 14 *foot* print looks fine at about 50
Yards or meters maybe even rods.

> --
you lack a space after the --

--
John McWilliams
~A Serenity Prayer~
"God, grant me the senility to forget the people I never liked anyway,
the good fortune to run into the ones I do, and the eyesight to tell the
difference."
Anonymous
July 6, 2005 9:34:55 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

On Wed, 06 Jul 2005 09:52:59 -0700, John McWilliams
<jpmcw@comcast.net> wrote:

>Owamanga wrote:
>> On Wed, 06 Jul 2005 07:54:24 -0700, John McWilliams
>> <jpmcw@comcast.net> wrote:
>>
>>
>>>IAE, in the above example, there has to be an optimum size to send to
>>>the Epson. Note I am not claiming a material or immaterial improvement
>>>in the quality of the print. Maybe it just prints sooner or faster.
>>>Maybe it's all moot, but still there has to be an optimum in what the
>>>printer receives.
>>
>> Epson printer drivers communicate the image to the printer at 720ppi,
>> Canon drivers work at 600ppi. There is little point exceeding these
>> resolutions when sending a job to the printer, but I doubt there is
>> much harm either.
>
>But does it follow that either value is the correct way to send out of
>PS to those respective printer drivers?
>
>Does the driver convert my 300 ppi image to 720 ppi before converting it
>to 360, 720 or 1440 dpi?

Where the logic lies (eg in the printer, or in the driver) to convert
the pixel into whatever dots of whatever size and color are required,
I don't know. It may differ for each printer model or brand
(certainly, with postscript printers, the task of dithering is done in
the printer). I'd guess it's a bit of both for inkjets.

I believe the application must provide the data to the printer driver
in whatever native resolution it reports itself as being (eg 720ppi in
Epson's case), regardless of the chosen dpi you've set on the printer
driver. So behind the scenes, Photoshop is upscaling/downscaling your
image to provide the printer driver with 720ppi data. The dpi settings
in the driver just tell the logic (wherever that may be) how to handle
the dithering + print head movements.

>you lack a space after the --

Ooops...

--
Owamanga!
http://www.pbase.com/owamanga
Anonymous
July 6, 2005 9:34:56 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

"Owamanga" <owamanga-not-this-bit@hotmail.com> wrote in message
news:p 75oc1dp8elr989i2ci42ntr81b7bnudmg@4ax.com...

> I believe the application must provide the data to the printer driver
> in whatever native resolution it reports itself as being (eg 720ppi in
> Epson's case), regardless of the chosen dpi you've set on the printer
> driver. So behind the scenes, Photoshop is upscaling/downscaling your
> image to provide the printer driver with 720ppi data. The dpi settings
> in the driver just tell the logic (wherever that may be) how to handle
> the dithering + print head movements.

Nope. Don't thik so. It is up to the driver to resample.
Anonymous
July 6, 2005 10:08:39 PM

Archived from groups: rec.photo.equipment.35mm,rec.photo.digital,rec.photo.equipment.large-format (More info?)

On Wed, 6 Jul 2005 12:45:19 -0500, "Lorem Ipsum" <Lorem@ipsum.xxx>
wrote:

>
>"Owamanga" <owamanga-not-this-bit@hotmail.com> wrote in message
>news:p 75oc1dp8elr989i2ci42ntr81b7bnudmg@4ax.com...
>
>> I believe the application must provide the data to the printer driver
>> in whatever native resolution it reports itself as being (eg 720ppi in
>> Epson's case), regardless of the chosen dpi you've set on the printer
>> driver. So behind the scenes, Photoshop is upscaling/downscaling your
>> image to provide the printer driver with 720ppi data. The dpi settings
>> in the driver just tell the logic (wherever that may be) how to handle
>> the dithering + print head movements.
>
>Nope. Don't thik so. It is up to the driver to resample.

A quick check with google, and yes, this is correct, the print driver
will do the resample to the internal resolution of 720ppi, not the
host application. Ken Rockwell suggests Photoshop can do a better job
of this than the Epson driver, and recommends resampling to either 360
or 720ppi in Photoshop prior to printing.

http://www.kenrockwell.com/tech/gf.htm

Here is a large article on quality with inkjet prints:

http://www.ddisoftware.com/qimage/quality/

--
Owamanga!
http://www.pbase.com/owamanga
!