Image enlargement software

On 2014-10-17 03:37:47 +0000, Eric Stevens said:

On Thu, 16 Oct 2014 15:02:27 -0700, Savageduck
wrote:

--- snip ---

Just remember, a dot does not equal a pixel. Depending on the printer a
dot is made up of a matrix, usually of 2-4-16 pixels. So your Epson
printer doesn't print at 360 DPI, it extrapolates 360 ppi resolution
data to the appropriate DPI for the print quality setting you have
chosen.

With my Epson printer, in the driver dialog I have the options of selecting:
Draft
Fine; 720 dpi @ 360 ppi gives me 2 pixels per dot.
Super Fine; 1440 dpi @ 360 ppi gives me 4 pixels per dot.
Photo; 1440 dpi @ 360 ppi gives me 4 pixels per dot.
Super Photo; 5760 dpi @ 360 ppi gives me 16 pixels per dot.

I hate to say this but something is confused somewhere. Pixels are
comprised of dots (that's how you get different shades of color), not
the other way around.

I know. It was a brain fart and I corrected myself in my response to
*Mayayana*, where I wrote:

Yup!
It is the other way around. Easy to confuse sometimes. ;-)

So...
Fine; 720 dpi @ 360 ppi gives me 2 dots per pixel.
Super Fine; 1440 dpi @ 360 ppi gives me 4 dots per pixel.
Photo; 1440 dpi @ 360 ppi gives me 4 dots per pixel.
Super Photo; 5760 dpi @ 360 ppi gives me 16 dots per pixel.

As you know, my printer is an Epson 3800 which is very much the same
as yours except for paper size. The service manual lists the following
print modes for mine with inkjet paper:

Print Quality Print Density Dot Size
(H x V)
Normal (360 DPI) 720 x 360 dpi VSD1
Fine (720 dpi) 720 x 720 dpi VSD1
Superfine (1440 dpi) 1440 x 720 dpi VSD2
Superphoto (2880 dpi) 1880 x 1440 dpi VSD3

Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

Eric Chan (now with Adobe) wrote a series of interesting articles
about the 3800 a few years ago.
http://people.csail.mit.edu/ericchan...tml#native_res
is relevant to the present discussion.


--
Regards,

Savageduck

On Thu, 16 Oct 2014 21:15:19 -0700, Savageduck
wrote:

On 2014-10-17 03:37:47 +0000, Eric Stevens said:

On Thu, 16 Oct 2014 15:02:27 -0700, Savageduck
wrote:

--- snip ---

Just remember, a dot does not equal a pixel. Depending on the printer a
dot is made up of a matrix, usually of 2-4-16 pixels. So your Epson
printer doesn't print at 360 DPI, it extrapolates 360 ppi resolution
data to the appropriate DPI for the print quality setting you have
chosen.

With my Epson printer, in the driver dialog I have the options of selecting:
Draft
Fine; 720 dpi @ 360 ppi gives me 2 pixels per dot.
Super Fine; 1440 dpi @ 360 ppi gives me 4 pixels per dot.
Photo; 1440 dpi @ 360 ppi gives me 4 pixels per dot.
Super Photo; 5760 dpi @ 360 ppi gives me 16 pixels per dot.

I hate to say this but something is confused somewhere. Pixels are
comprised of dots (that's how you get different shades of color), not
the other way around.

I know. It was a brain fart and I corrected myself in my response to
*Mayayana*, where I wrote:

Yup!
It is the other way around. Easy to confuse sometimes. ;-)

So...
Fine; 720 dpi @ 360 ppi gives me 2 dots per pixel.
Super Fine; 1440 dpi @ 360 ppi gives me 4 dots per pixel.
Photo; 1440 dpi @ 360 ppi gives me 4 dots per pixel.
Super Photo; 5760 dpi @ 360 ppi gives me 16 dots per pixel.

As you know, my printer is an Epson 3800 which is very much the same
as yours except for paper size. The service manual lists the following
print modes for mine with inkjet paper:

Print Quality Print Density Dot Size
(H x V)
Normal (360 DPI) 720 x 360 dpi VSD1
Fine (720 dpi) 720 x 720 dpi VSD1
Superfine (1440 dpi) 1440 x 720 dpi VSD2
Superphoto (2880 dpi) 1880 x 1440 dpi VSD3

Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

16 dots per print cell: twice the 2800 dpi of my printer.

Eric Chan (now with Adobe) wrote a series of interesting articles
about the 3800 a few years ago.
http://people.csail.mit.edu/ericchan...tml#native_res
is relevant to the present discussion.
--

Regards,

Eric Stevens

Eric Stevens wrote:
Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

16 dots per print cell: twice the 2800 dpi of my printer.

The R2880 can print at 2880 DPI. Depending on the configuration,
the Epson driver presents the same titles for resolution
options, but they might have different effects.

Here's a blurb from Epson showing that:

http://www.epson.co.uk/gb/en/viewcon.../faq/2733/4011

But the distinction of the 5760 DPI option compared to
the 2880 DPI option is interesting because of the way it
is done.

The print head cannot actually squirt ink out at a rate
higher than 2880 DPI. To accomplish 5760 what they do is
make two passes, the second one offset by one dot width,
and both working at 2880 DPI but with each ink dot covering
only half the size that the head is moved each time.

The first pass is like this, going from right to left:

O O O O

The second pass, which is done without moving the paper,
is from left to right:

0 0 0 0

And the result is this:

O0O0O0O0

Epson nozzles can produce ink dots in three different
sizes in any given mode, and when used in the 5760 DPI
high resolution mode the only size dot that is used is
the smallest. (I don't have specs on the R2880, but the
larger printers typically have three modes, the 3880
model has dot sizes in picoliters of 22.2, 13.8 and 6.6,
in another mode it is 13.2, 5.9 and 3.5. The high
resolution mode uses only 3.5 picoliter dot size.)

If that mode is used to print a photograph the most
significant characteristc is only that it takes two
times as long to make a print! If it is used for text
or line drawings the difference is day and night in
terms of sharpness.

--
Floyd L. Davidson http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)

On 2014-10-17 08:18:42 +0000, Eric Stevens said:

On Thu, 16 Oct 2014 21:15:19 -0700, Savageduck
wrote:

Le Snip

Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

16 dots per print cell: twice the 2800 dpi of my printer.

I guess they want me use ink at a faster rate, out of my small carts.
Anyway, I get some pretty good prints, especially on Red River Paper's
*Polar Pearl Metallic* in combination with the paper/printer color
profiles they provide. Naturally, when using the profiles, Photoshop,
or LR has to manage the print, not the printer.


--
Regards,

Savageduck

On 2014-10-17 09:14:37 +0000, (Floyd L. Davidson) said:

Eric Stevens wrote:
Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

16 dots per print cell: twice the 2800 dpi of my printer.

The R2880 can print at 2880 DPI. Depending on the configuration,
the Epson driver presents the same titles for resolution
options, but they might have different effects.

Here's a blurb from Epson showing that:

http://www.epson.co.uk/gb/en/viewcon.../faq/2733/4011

But

the distinction of the 5760 DPI option compared to
the 2880 DPI option is interesting because of the way it
is done.

The print head cannot actually squirt ink out at a rate
higher than 2880 DPI. To accomplish 5760 what they do is
make two passes, the second one offset by one dot width,
and both working at 2880 DPI but with each ink dot covering
only half the size that the head is moved each time.

The first pass is like this, going from right to left:

O O O O

The second pass, which is done without moving the paper,
is from left to right:

0 0 0 0

And the result is this:

O0O0O0O0

Epson nozzles can produce ink dots in three different
sizes in any given mode, and when used in the 5760 DPI
high resolution mode the only size dot that is used is
the smallest. (I don't have specs on the R2880, but the
larger printers typically have three modes, the 3880
model has dot sizes in picoliters of 22.2, 13.8 and 6.6,
in another mode it is 13.2, 5.9 and 3.5. The high
resolution mode uses only 3.5 picoliter dot size.)

If that mode is used to print a photograph the most
significant characteristc is only that it takes two
times as long to make a print! If it is used for text
or line drawings the difference is day and night in
terms of sharpness.

That makes sense & thanks for that piece of education. However, that
leaves the question; Why is the 2880 dpi option absent from my driver
dialog, only giving me 1440 dpi & 5760 dpi options?

I find that for most prints, particularly B&W on matte, textured, or
Satin paper, 1440 dpi seems more than adequate. With the Red River
*Polar Metallic Chrome* paper there is a visible difference in the
finish when using 5760 dpi. The added cost is time and ink.


--
Regards,

Savageduck

Savageduck wrote:
That makes sense & thanks for that piece of
education. However, that leaves the question; Why is the
2880 dpi option absent from my driver dialog, only
giving me 1440 dpi & 5760 dpi options?

I am not able to determine exactly what makes that
happen, but likely reasons would be using a paper that Epson
deems not better for photographs at 2880 compared to
1440. The 5760 may be offered for two reasons, one
being that for line drawings or text, even with that
paper it is better. Another might be that they just
can't avoid the on/off "high res" option that enables
5760 dpi. Who knows...

Epson's idea of a user interface is one of the most
annoying things around!

I find that for most prints, particularly B&W on matte,
textured, or Satin paper, 1440 dpi seems more than
adequate. With the Red River *Polar Metallic Chrome*
paper there is a visible difference in the finish when
using 5760 dpi. The added cost is time and ink.

I doubt that it actually takes significantly more ink.
Time though, is really what it takes.

--
Floyd L. Davidson http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)

| Is there any reason to assume that
| IV and other programs are not doing the resampling
| themselves before sending it to the printer? I wonder
| if the printer API even has options for resampling. That's
| really not part of its job.
|
| Yes of course, because the printer is taking a pixel image and
| converting it (by and by) into a series of commands to some tiny
| nozzles to squirt out some ink. Is this happening in the pixel domain
| or the mechanism domain? Who knows, it certainly varies per printer
| and possibly depending on the input image size and such even with the
| same printer.

The printer has to translate, but with interpolation
one is hopefully optimizing. Optimizing would make no
sense without first seeing the result before printing.
So if you send an image to the printer and tell it to
print twice the size I'm guessing it's just going to
enlarge each pixel, which is the data it has to work
with. Even if the printer could do something like bicubic
interpolation, what would be the point? Without
controlling the process and checking the result you'd
have no way to know whether it would be an improvement.
In fact, I wouldn't want a printer to do something like
applying a bicubic interpolation, willy nilly and without
telling me. Wouldn't that defeat the purpose of preparing
the image yourself?

On Fri, 17 Oct 2014 09:45:02 -0400, "Mayayana"
wrote:


| Is there any reason to assume that
| IV and other programs are not doing the resampling
| themselves before sending it to the printer? I wonder
| if the printer API even has options for resampling. That's
| really not part of its job.
|
| Yes of course, because the printer is taking a pixel image and
| converting it (by and by) into a series of commands to some tiny
| nozzles to squirt out some ink. Is this happening in the pixel domain
| or the mechanism domain? Who knows, it certainly varies per printer
| and possibly depending on the input image size and such even with the
| same printer.

The printer has to translate, but with interpolation
one is hopefully optimizing. Optimizing would make no
sense without first seeing the result before printing.
So if you send an image to the printer and tell it to
print twice the size I'm guessing it's just going to
enlarge each pixel, which is the data it has to work
with. Even if the printer could do something like bicubic
interpolation, what would be the point? Without
controlling the process and checking the result you'd
have no way to know whether it would be an improvement.
In fact, I wouldn't want a printer to do something like
applying a bicubic interpolation, willy nilly and without
telling me. Wouldn't that defeat the purpose of preparing
the image yourself?


If you ask the printer to print twice the size it's almost certainly
going to do it by printing twice as many pixels. What it will have to
do is create a new pixel between each of the original ones. If the
values of the originals were A and B, it is almost certain that the
values assigned to the new one will be (A + B)/2. That might be OK but
if A and B deliniated an edge they would be markedly different and the
new intermediate pixel would soften that edge. That almost certainly
is not what you want.

The matter is further complicated by the fact that (except at the
edges) each pixel has 8 neighbours in the first surrounding layer and
16 neighbours in the layer surrounding that. Interpolating the values
of new pixels is both a mathematical nightmare and computationally
heavy. It's better to leave the up-scaling/rescanning of your image to
purpose designed software and only rely on the printer driver to
translate the resulting pixel values to the right pattern of droplets
on the paper.
--

Regards,

Eric Stevens

On Fri, 17 Oct 2014 03:40:57 -0700, Savageduck
wrote:

On 2014-10-17 08:18:42 +0000, Eric Stevens said:

On Thu, 16 Oct 2014 21:15:19 -0700, Savageduck
wrote:

Le Snip

Strangely enough my R2880 doesn't have a 2880 dpi option and Super
Photo is 5760 dpi.

16 dots per print cell: twice the 2800 dpi of my printer.

I guess they want me use ink at a faster rate, out of my small carts.
Anyway, I get some pretty good prints, especially on Red River Paper's
*Polar Pearl Metallic* in combination with the paper/printer color
profiles they provide. Naturally, when using the profiles, Photoshop,
or LR has to manage the print, not the printer.

I don't know about the ink consumption. You are getting more drops but
they are smaller drops. If you get an increased consumption it is
probably due to better coverage in the corners.

Here is an older article which is still relevant to the general topic
in many respects.
http://www.rags-int-inc.com/PhotoTechStuff/Epson2200/
--

Regards,

Eric Stevens

On 2014-10-17 13:45:02 +0000, Mayayana said:

The printer has to translate, but with interpolation
one is hopefully optimizing. Optimizing would make no
sense without first seeing the result before printing.
So if you send an image to the printer and tell it to
print twice the size I'm guessing it's just going to
enlarge each pixel, which is the data it has to work
with.

But is guessing really good enough? The printer or driver (or whatever
is in the imaging chain) might be applying one or more smart transforms
to the image, depending on any number of factors, some of which are
invisible to the user.

Even if the printer could do something like bicubic
interpolation, what would be the point? Without
controlling the process and checking the result you'd
have no way to know whether it would be an improvement.

Precisely my point. Unless you do a number of tests for your
particular printer setup you will never know or be able to control the
process. This might vary depending on file size, driver version, etc.
and there is no way to know unless you test it yourself.

In fact, I wouldn't want a printer to do something like
applying a bicubic interpolation, willy nilly and without
telling me. Wouldn't that defeat the purpose of preparing
the image yourself?

What you want from your printer, and what the people who made it think
would help it sell the best to the greatest number of people, might not
intersect fully.