Resolution (Dpi) in print module

[OXiDant (Peter)]

@George

For sure, it’s linear.

i take your word for it!

@George

I still don’t know how the resizing i(s?) done.

Out your link: (about downsizing and upscaling)

Bicubic Sharper (downsizing)

A good method for reducing the size of an image based on Bicubic interpolation with enhanced sharpening. This method maintains the detail in a resampled image. If Bicubic Sharper oversharpens some areas of an image, try using Bicubic.

Bilinear (upscaling)

A method that adds pixels by averaging the color values of surrounding pixels. It produces medium-quality results.

Better methode:

Bicubic

A slower but more precise method based on an examination of the values of surrounding pixels. Using more complex calculations, Bicubic produces smoother tonal gradations than Nearest Neighbor or Bilinear.

you can slow down writing speed on the drum (less lines/sec) so the papermovement has more time to transport between the writing lines. then you lower the y-resolution.

[George]

I refer to resizing when printing.
When taking pictures we deal with a exposure triangle. The exposure is determined by aperture, shutterspeed and iso. The same we have with printing: image dimension in pixels, print resolution in dpi or ppi, and document sizein inch. If we change one of them, at least one of the others has to be changed too. So let’s say I don’t want to resample but still want that specific document size. The only thing I can change is the print resolution. But what happens then, that’s what I don’t know. The printer print resolution can’t be changed, it’s hardware.So my idea is that some kind of resampling is done in the background. I’m not sure of this.

George

[George]

I remember a picture of a grave I took with text on it, shot under 45 degrees. Full size it was ok, just good. But when zooming out the text showed faults. I tried then to see the difference between printing that image when resampling to the right size based on the printers resolution and the other with let the printer do it. The last one was better.

George

[OXiDant (Peter)]

@George

The printer print resolution can’t be changed, it’s hardware

Yes a printer can change it’s resolution.
By placing the dots wider of each other. Less dots is less resolution.
Resampling is deleting some of same color pixels in groups and replace, add more “space of nothing” on that same spot.
Only then is the size the same.

Other way around, a low ppi 100x100ppi causes bloks and zigzaged lines in the image if a 600dpi printer and the full size A4 is. Every 6st dot is a color.
I can show some examples made by a laser printer as in a office 600dpi colorprinter when i make some photo’s in different ppi export.
Printing on A4.
And use a paper loup to show the toner drops.

[George]

I’ve a brother color laserprinter, 600 dpi. I don’t use it for pictures.
And I have a HP Photosmart c309a. Also 600dpi. I can’t change it.

George

[platypus (Canon & Mac (Intel and M1))]

yesyes, already corrected in my post…

[Joanna (Joanna Carter)]

@George

When taking pictures we deal with a exposure triangle. The exposure is determined by aperture, shutterspeed and iso. The same we have with printing: image dimension in pixels, print resolution in dpi or ppi, and document sizein inch.

We never have print resolution in ppi. Printers only know about dpi. They are dot devices, not pixel devices.

@George

If we change one of them, at least one of the others has to be changed too. So let’s say I don’t want to resample but still want that specific document size. The only thing I can change is the print resolution. But what happens then, that’s what I don’t know.

Only if you are talking about images, not printing here. Print scaling is different to resizing an image and can result in very poor results but it will only be printing the pixels in the document, just with more or less dots.

@George

The printer print resolution can’t be changed, it’s hardware.So my idea is that some kind of resampling is done in the background. I’m not sure of this.

Certainly resampling is done but not at a pixel level, just the spacing between the dots.

I repeat - printers are dot devices, not pixel devices. But printers have software in them that manages the output quality, speed and distance between the dots. thus the idea of resampling, but not at a pixel level and changing the dpi of a printer does not change to size of the print.

[OXiDant (Peter)]

To add on your post also for @George ,
The printers i service are 1200dpi which are default set on 600dpi.

It’s “native” resolution is 1200dpi.
About colors, to create a saturation ladder that’s also about space between the tonerflakes the lighter the color the less dots/tonerflakes per mm2 are placed.

The cmy blend is a greyladder build with cmy not Bk black. If this is looked at by a loup you see the three colors spread in a equal way.
When it 's calibration is “off” this build grey is colored.

A scan is front rear lines like a electronic shutter.

And the subscan speed does also the subscan resolution. The main scan resolution is software based.

When i scan and reprint from Harddisk, copy is always 600ppi, i can simulate your ppi to dpi conversion. A scan made at 100x100 A4 wil be printed as A4 with a low dpi resolution.
We always say,
100x100 is archiving.
300x300 is default reprinting minimum scan resolution for printing again.
Any thing above is for detailing and when original is A4 wile print needs A3.

Keep in mind that a laserprinted A4 is 600dpi so it become a 300dpi which is computed back to 600dpi.

[Joanna (Joanna Carter)]

Dots per inch

[platypus (Canon & Mac (Intel and M1))]

Let me add something like an analogy regarding dpi and ppi.

Imagine a square 18 Megapixel Bayer sensor that measures 1 by 1 in. If we look at it closely, the sensor has 9M green dots, 4.5M red dots and 4.5 M blue dots (or photosites) that collect light.

Technically, this sensor resolves 3000 dpi for green, 1500 dpi for red and 1500 dpi for blue light.

Demosaicing takes several of the sensor dots into account to create a pixel and there are several ways to do so.

Printing basically does the inverse. It transforms each pixel into a number of dots and there are several ways to do so too.

[George]

There’s a lot of confusion with terms.
A digital image/file doesn’t have a resolution, nor in dpi or ppi. It has a non metric dimension.

Image resolution

Pixel resolution

“The term resolution is often considered equivalent to pixel count in digital imaging, though international standards in the digital camera field specify it should instead be called “Number of Total Pixels” in relation to image sensors, and as “Number of Recorded Pixels” for what is fully captured. Hence, CIPA DCG-001 calls for notation such as “Number of Recorded Pixels 1000 × 1500”.[1][2] According to the same standards, the “Number of Effective Pixels” that an image sensor or digital camera has is the count of pixel sensors that contribute to the final image (including pixels not in said image but nevertheless support the image filtering process), as opposed to the number of total pixels , which includes unused or light-shielded pixels around the edges.”

It’s the screen/monitor or printer that has a resolution and gives the non metric image a metric size.
When printing every pixel is translated in a dot and like a pixel it contains more channels/colors. Or more specific the final dot is created from more dots/colors at the same place.
Though a dot is physical different from a pixel both dpi and ppi can be used to calculate the size of a form/image.
It’s good to learn how printers deal with these settings. I don’t print myself and even when I do I use the predefined sizes, like in the pl print window. But ther must be more.
The size of a dot in the native resoltion of 600dpi = 25.4/600=0.042mm.
When printing on 10dpi the size will be 2.54mm. If using the idea of seperations between the dots, then there should be a gap of 2.498mm. I can’t see that.

George

[Joanna (Joanna Carter)]

@George

Though a dot is physical different from a pixel both dpi and ppi can be used to calculate the size of a form/image.

Here we go again. Dot resolution is a measure of how many dots of ink are laid down on the paper. It has nothing to do with the size of the image, merely the quality of the print. the printer’s internal software will interpolate between the size of an image in pixels and the size of the print in inch/cm.

If you were to use an external printing utility like I do, this becomes obvious. In PL, I export an image to TIFF, stating the finished size in pixels or cm to print on a certain size of paper. I then take the pre-sized image and open it in ColorSync Utility for printing. At this point, I can choose the printing resolution in dpi. Whether it be draft quality at 720dpi or SuperPhoto at 5760, the print will come out at exactly the same size; the difference will be the quality of the print.

@George, you say yourself you do not print and that you don’t know how all this works. But then you make assumptions about how it works, which are are wrong. The fact that some manufacturers confuse dpi and ppi in some of their literature hasn’t helped you in your learning.

Until you get to the printing part of the process, ppi is used to determine the physical size of an image. My monitor has a resolution of around 109 ppi so a 100% scale view of a 36 Mpx image will come out at 67.5 inches wide and I will only be able to see part of it at any one time; which is why, most of the time, the “size to fit” scale, for a landscape image, in PL ends up at around 22% to allow me to see both the image and the rest of PL’s palettes and tools.

Following rules for minimum pixel density for a given viewing distance, I know that using 240 ppi to create an output file that will be printed at 100% will give me a quality where a person viewing the print at arm’s length will not be able to discern the pixels. No matter what printer resolution I choose, much less than 240 ppi will degrade the print quality, even if I use the highest possible printer resolution.

However, I print at 1440 dpi, which will give a quality of image, on photographic paper, that is perfectly acceptable, when viewed at arm’s length, provided the image resolution isn’t much less than 240 ppi.

So, I send a pre-sized image at 240 ppi to the printer and the printer driver works out how many pixels it takes to represent each pixel at 1440 dpi - I make it 6 dots per pixel. Since a pixel can only ever have one CMYK value, that means all 36 dots for that pixel on the print should have that same value, except that there will have to be a mixture of coloured dots to include however many different coloured inks it takes to give the illusion of that specific colour.

Think of it like this - ppi is used for input/display/sizing, dpi is used for output to a printer.

From what you have said about your workflow, you tend to let the printer dialog take care of scaling the image to fit on a sheet of paper, so you won’t have any idea of the eventual ppi resolution of the image. All you need to know is that the printer driver will translate each pixel of a 240 ppi image into a 2.5 x 2.5 dot block on your 600 dpi printer, except you will not know what the image resolution was because you are letting the print dialog work it out for you. Suffice to say that, for most 24 Mpx cameras, you should be able to print an uncropped image up to A3+ without too much worry about print quality.

[George]

One try more.
There’s something like a print triangle: image size in pixels, print resolution in dpi and document size.

Info window of Irfanview. A printresolution of 300x300 gives a document of 15x10cm.

Just changing the resolution in 100x100. A document size of 46x31cm.

And for fun a resolution of 100x500. Document size 46x6cm.

What you call a resize of 240 is nothing else as a question to the printer: please print me on 240dpi so my document size is…

You own an Epson. From Epson. When they talk about image resolution they mean image size in pixels.
https://www.epson.nl/viewcon/corporatesite/kb/index/2091.

All these formula’s can be done only if there is a one to one conversion between pixel and dot.

George

[Joanna (Joanna Carter)]

There is no such “triangle” and document size refers to the image file, not the print. The only physical size an image file knows about is the size of the sensor that captured the image.

Since most full frame sensors are 1" x 1.5", we know that the original resolution of the sensor image must be the number of pixels on the shorter side, expressed as ppi. In the case of my Nikon D810, this is 4912.

However, when printing, you need to take into account what distance the print will be viewed from and this often leads to people using 300 ppi when resizing a file. In reality, the human eye cannot distinguish much more detail at 240 ppi, which is what I use.

Why some people get confused over the difference between ppi and dpi is that a lot of printer manufacturers define their dot resolutions in multiples of 300 and the assumption is that there is some kind of relationship between the the two 300s. There is not, it’s just coincidence.

If you printed an image at 300 dpi, you would get a very poor quality print where the viewer would be able to see the individual dots (a quality similar to fax documents); which is why we tend to print photographs at a dot resolution in the thousands.

It seems like IrfanView is one of those companies that wrongly use dpi when it is talking about ppi. As to the idea of being able to have a different horizontal and vertical resolution, I have but one question - why???

The dialog you show is titled Image Properties, not Print. The author is not a native English speaker and this may be the reason for the mistake of using dpi instead of ppi, just as Epson are conflating the idea of dots and pixels on some of their literature, especially on the subject of scanning, where it gets even more confusing.

I (as others on this forum) have been doing this stuff for many years and can assure you that printer ink dot size has nothing whatsoever to do with image resolution in pixels; it is the job of the driver to map between the two. What is confusing the issue for you is the conflation of dpi and ppi by some companies.

I never use dpi when resizing an image because it is something that only the printer driver needs to know about when it translates pixels into dots. You simply don’t need to know anything about the relationship between the number of pixels in an image and the number of dots on a print; it’s all taken care of by the printer driver. The only thing that affects print size is the pixel resolution of the image.

But if you wish to believe otherwise, I can only wish you luck in trying to work out something that is simpler than you are making it.

[George]

@Joanna

There is no such “triangle” and document size refers to the image file, not the print. The only physical size an image file knows about is the size of the sensor that captured the image.

I did write “print triangle”. Three variables where each variable is dependent of the other two: image size in pixels, resolution and document size.
Inch=pixels/(pixesl/inch)

@Joanna

Since most full frame sensors are 1" x 1.5", we know that the original resolution of the sensor image must be the number of pixels on the shorter side, expressed as ppi. In the case of my Nikon D810, this is 4912.

Think about what you write. PPI = PIXELS PER INCH. I never heard of this naming but even then. Short side is 24mm. That resolution would be 4912/24=204.66 pixel per mm or 8.05ppi.

@Joanna

However, when printing, you need to take into account what distance the print will be viewed from and this often leads to people using 300 ppi when resizing a file. In reality, the human eye cannot distinguish much more detail that 240 ppi, which is what I use.

A resolution of 240ppi means 240 pixels per inch. Every pixel has a size of 25.4/240=0.1mm.
Consider the maximal accepted circle of confusion of a ff camera is 0.030 and that picture is printed on A4, meaning a enlargement of let’s say 8 then that coc on the sensor will be 0.24, 2.4 times bigger. You could do it with 100ppi. The human eye doesn’t see ppi.

@Joanna

If you printed an image at 300 dpi, you would get a very poor quality print where the viewer would be able to see the individual dots (a quality similar to fax documents); which is why we tend to print photographs at a dot resolution in the thousands

I don’t understand. 240 is less the 300.

@Joanna

It seems like IrfanView is one of those companies that wrongly use dpi when it is talking about ppi. As to the idea of being able to have a different horizontal and vertical resolution, I have but one question - why???

Every companie uses the same formula. Different resolution? As I wrote: for fun and to show you what happens. Have a look at ps or any other editing program. You can lock or unluck the relation between the 2 sides,

@Joanna

I never use dpi when resizing an image because it is something that only the printer driver needs to know about when it translates pixels into dots. You simply don’t need to know anything about the relationship between the number of pixels in an image and the number of dots on a print; it’s all taken care of by the printer driver. The only thing that affects print size is the pixel resolution of the image.

As you can see from the formulas, the numbers of pixels equals the number of dots.

George

[OXiDant (Peter)]

@George

Info window of Irfanview. A printresolution of 300x300 gives a document of 15x10cm.

This is if you are using 1:1 conversion, every pixel becomes one dot CMYK

@George

Just changing the resolution in 100x100. A document size of 46x31cm.

You changing the the output resolution conversion so the file can be plotted/printed in a larger plane/ on larger paper with the same source resolution.
If you lower the DPI the source can be spread over a larger field. The image resolution doesn’t change the “dots get bigger” not really but there are more dots the same color creating a image.
This image was printed by kids on a 600dpi laserprinter.:

Look quite nice on your computerscreen right? even at 100% view.

@Joanna

The only thing that affects print size is the pixel resolution of the image.

proof to the pudding:
Wel this was it’s A4 print view:

Look at it at 100% ( i took a photo with my phone) download and open on your screen.
Horrible. (remember the print is stil 600dpi!)
Now some shots from test sheets directly out the printer so native resolution pcl default.
The gray line:

See the colors thrue a paperloup shot by my phone.
Then a RGB green color made by CMY:

See it’s not a plane full color it’s playing with your eye’s.
A fish lour image:

Finally the Magenta ladder: full saturated:

somewhere in the middle:

i made a scan of the test page at 200% to get a "bad"sample and converted this PDF to jpeg:

Al this is a non photoprinter. a general office printer with a modest 600dpi resolution using Toner particals which alow to go to1200DPI output.
I use it to print "cheap"A3 fullcolor prints and they look good enough.
(if i use silk paper it get smoother because the surface is big factor on the endproduct. this was cheap paper on a school.)

A inkjet printer uses “blending” of blobs so the paper is fully covered by ink.
But wile the resolution fysicly doesn’t change as in wider nozzles but it’s dots of one color does.
If you send that 500x500 image to print on a A4 it looks just the same as the Laserprinter version.
And that color ladder would be done the same way: spreading dots more om a mm2 so you think it’s a lighter color. but stil it’s one CMY. no white ink to blend with Magenta for a lighter color.

So infact a printer does have different DPI in 1 image wile the “resolution” of the printed image doesn’t changed. (Saturation changing means DPI changing physicly )
in digital image mapping it’s a "white"dot between the inkdots. non writing by laser or non spitting by inkjet.)

So the only thing you need to remember is that the source image has more pixels in xxx by xxx then the printed resolution so it’s always “downsizing” and not upscaling by adding “color” pixels.
last thing can maybe help to understand the propertie which uses dpi :
300ppi or 72ppi.docx (437,2 KB)

[platypus (Canon & Mac (Intel and M1))]

@RobEW

I’m not clear whether I am supposed to select a (dpi) value which accurately describes the quality of the image I am using, or whether this is somehow instructing the printer to print at a particular resolution. Is there any advice as to what value to select to get the best quality?

In view of the number and contents of the replies, I propose that you ran a few test prints at different settings. They will show you the setting that will give you what you want.

[OXiDant (Peter)]

i think the best way to deal with ppi/dpi story is just say “I want a print of 15cm by 22cm” and let the driver deal with the conversion.
The printer wil be always as close as it can render to “native resolution”.
So the printed image has “unique” dots and as less “sibling” as possible.
my doc above states that the jpeg properties “dpi setting” doesn’t matter. it’s just for sizing in "paperlayout software like “word” and does no “physical” resolution change of the image.

(i hope i didn’t overtechy the post above )
When people state a printer has a fixed resolution in dots per inch is the same as stating a camera has a fixed jpeg ppi output.
i can lower the pixel resolution of a jpeg in camera or use Electronic Zoom and “crop” both are the same end resolution but different angle of view.
One is deleting pixel info to skrink the overal image in 100% view and the other is cutting a part out the original image.

Same does the printerdriver in “printing”: it’s a latent image rendered/and scaled down or up to a point that it fit inside the Physical resolution of the native resolution. in my mfp 1200dpi.
then the internal software skrink this image to 600dpi for printing speed. (it’s default resolution)
(rendertime and papertransportspeed).
you can see this back in auto document scanning (ADF with a slidglas (standing scannermirror and moving paper.)
Scan a A4r in to saving a A3 it slows down in order to write more sublines to enlarge the image.
scan a A3 to A4R and it speeds up to skrink the subscan. (mainscan is renderend by software)
Change scanresolution in 100x100 and it’s fast change it in 600x600 and it slows down even with the same A4. (because the physical resolution of the CCD doesn’t change only that of the image in pixels, the clocking rows of ccd input is the same so the objectspeed(A4) is the key for the file document format(100% preview in A3 or A4rsize) and thus filesize.)
A printerdriver does the same: low resolution print? it skips pixels when you keep the papersize the same or it spread the image over a larger piece of paper if you keep the pixelcount the same.
highest resolution? (the native resolution of the hardware)
the image skrinks to a point it’s one on one converted on a part of the paper or if you want to keep the same papersize (full covered) it wil add pixels to the image in order to get the wanted printsize.

So the “preview modes” in a printer driver can help out to see if a print wil be too large for the actual jpeg resolution. just see when 1:1 is smaller then the papersize fully filled.

(geesh sorry, job enthousiastic behaviour… )
i shut up now.

[George]

There is the question again. Is the document format changed by resampling the image or by changing the print resolution?

@OXiDant

When people state a printer has a fixed resolution in dots per inch is the same as stating a camera has a fixed jpeg ppi output.

Comparing it with the properties of the sensor is more logic.

George

[OXiDant (Peter)]

@George

Is the document format changed by resampling the image or by changing the print resolution?

What do you mean by Document format? Dimensions?
X and y size? Or type of file? like tiff of jpeg or raw or.
or native resolution, the maximal resolution it can give?
A hard copy doesn’t change.
A sensor’s rawfile resolution doesn’t change.
if i copy it and trow the original away it changes, same in if i make a jpeg and trow the rawfile away it changes.