Hi Collin,
When I put a sample printed on my printer (HP 890 C) under a microscope (yes I did, an I’m getting pro treatment&medicine for that) I can see discreet "dots" of the four (CMYK) guns. So they don’t mix on the paper but rather on my retina or in my brain (or what’s left of it after 30 years in a wet darkroom).
This is also true with gray: looking close it is still the four distinguishable inks.
This is not an answer to your question, but just my experience. Discussion opened AFAIAC.
Rob
The main difference between the way inkjets and offset presses print is that offsets vary the size of evenly spaced dots, and inkjets vary the spacing of identially sized dots. The net effect is the same; to cover some percentage of the paper with ink. Black is added because the three colored inks do not absorb absolutely all the light when mixed together.
Some pressmen (I’m told) refer to cyan as blue, and magenta as red, but magenta is purplish, and cyan is greenish. Pure red and blue (plus pure blue) LIGHT mixed together, as it is from your monitor, give the sensation of white. Cyan, magenta, and yellow PIGMENTS (or dyes) mixed together absorb all the colors. Since no light is reflected back, the result looks black, or nearly so.
You wouldn’t need white ink, since you allow some of the white paper to show through, Since paint completely covers the canvas, white pigment must be added to lighten the tone. If you were printing on colored paper, though, you migh want white ink there.
Colin,
I have the impression that they mix the colours together as the ink is
sprayed onto the paper.<<
No, they don’t.
In conventional halftone printing shades of colour are reproduced by changing the SIZE of the individual ‘cells’ of ink – small areas for light colours, large areas for saturated colours, but printed in a regular matrix or screen. This is also known an Amplitude Modulation as the size (amplitude) is varied.
An inkjet uses stochastic screening or Frequency Modulation. Here the SIZE of the ink dot remains constant but the spacing of the dots (the frequency) changes. Thus, with 4 colour inkjets, to get a very light magenta, large areas will contain no ink whatsoever and this reduces the resolution of the printed output. To solve this, modern inkjets use also a light magenta and light cyan, so that in light areas, they can print the dots closer together to get the same saturation as dots of the dark ink farther apart.
Now the SIZE of the dots from an FM screen are MUCH smaller than the cells of the AM screen, and so it is possible to get much nearer to the continuous tone of a photograph using FM screening than with AM screening.
if inkjets don’t work this way, why not?<<
Well, they DO (effectively) work that way when talking about 100% coverage of say a saturated blue. In fact the dots will still be separate when ‘sprayed’ but there is some spread (dot-gain) when the ink hits the media, so there is some mix of inks. It would be better if ,say, the magenta and cyan inks just touched rather than overlapped. The eye would see a better more saturated blue if they did. The ‘overlap’ tends to darken the image somewhat because dyes are not perfect!
If they WERE perfect, then the magenta would reflect exactly 2/3 rds. of the visible spectrum and absorb completely the other 3rd.(the green light). It doesn’t. Nor does yellow or cyan.
Would it make any sense to have a "RYBKW" colour model<<
No. Let’s assume our ‘perfect’ dyes. Red would absorb 2/3 rds of the visible spectrum and reflect 1/3rd, as would the blue and green dyes. Wherever two overlapped there would be 100% absorption and the result would be black
Colin:
If you look at a grayscale image from an inkjet printer, even one with a light black ink (such as the Epson 2200), you will see that the neutrals are composed of colored dots. Even a 7-color printer like the Epson 2200 will use plenty of cyan, magenta, and yellow as you move into three-quarter tone and shadows and plenty of light cyan, light magenta, and yellow as you move into the quarter tones and highlights. There will be some black and light black used, as well.
For grayscale images, some inkjet printers have a "Black Ink Only" feature. Unfortunately, this does not usually mean Black and Light Black (i.e., what you are calling "Gray"). Black and Light Black would leave you with a duotone image, not a continuous set of gray tones. What "Black Ink Only" does is use a dithering pattern along with the black ink.
As well as a difference in screen methods, the way inks combine in some inkjet printers is different that offset inks. Offset inks tend to be more linear in their combination. This means adding 10% more magenta has a similar visual effect whether you add it to shadows or highlights. Many inkjet inks (such as the Epson Ultrachrome inks) are not even close to linear. Hence, to get the same visual impact from a 10% increase in magenta (using a tool like Color Balance in PS CS), the printer driver might have to add more than 10% magenta to the shadows and less than 10% to the highlights.
One other point that has not been adequately explained is the reason for Black ink in 4, 6, and 7 color inkjet printers. It’s the same reason for the K plate in CMYK. Impurities in the CMY inks, mean they do no perform as they should theoretically. You tend to wind up with a muddy colored black instead of a neutral black. Also, to achieve dark shadows with CMYK, you need to apply a lot of ink with just CMY. Adding black ink lets you reduce the amount of CMY inks.
People tend to think that printers with a Black or a Black and Light Black inks tend to use those inks exclusively for grayscale tones. Wrong! Just go to Photoshop CS and select the default Black foreground patch. RGB values 0,0,0. Go to the Color Palette and switch to CMYK. If you’ve left the CMYK conversion parameters at their default, you’ll see 75%, 68%, 67%, 90%. More cyan than magenta and yellow and a lot of CMY ink all around. (Black is only 30% of the total ink — 90% out of 300%) That’s why printing B&W prints doesn’t drain your Black and Light Black cartridges at a wildly disproportionate rate. Then look at a midtone gray patch (128,128,128 in RGB). 51%, 43%, 43%, 7% for CMYK. Again, more cyan than magenta and yellow (cyan tending to be weaker, therefore requiring more cyan ink relative to magenta and yellow) and much more CMY inks than black. Less than 5% of all the ink is black (7% out of 144%). As you move from modtones into the quarter tones and highlights, the black ink will drop out altogether. A three-quarter tone (196,196,196) will be 23%, 18%, 18%, 0%. No black ink at all. Yet, it is clearly appears as a neutral light gray.
The fact that gray tones are represented by colored drops is why inkjet printers tend to have a problem producing neutral grays and with metamerism. For example, people who use Epson Ultrachrome inks (the Epson 2200, 7600, 9600, etc.) for fine art B&W prints often resort to a RIP to obtain neutral B&W prints and to avoid metamerism in their B&W and color images. The RIP is a substitute for the printer driver. RIPs are more intelligent in how they mix the available colors of ink in order to obtain neutral (or, if you choose, non-neutral) gray tones for B&W images and reduced metamerism in color images.
Cheers,
Mitch
"A three-quarter tone (196,196,196) will be 23%, 18%, 18%, 0%. No black ink at all. Yet, it is clearly appears as a neutral light gray."
Oops! Misspoke. Quarter tone.
😉
Cheers,
Mitch
Thanks for all the input folks. I am certainly educated on the way that inkjets do their job, but I return to my original thought: is there any reason why a printer should not work like an artist?
What I’m thinking is that, with the right inks, they could be mixed instead of juxtaposed. Whether this is done by the printer squirting the right number of tiny drops onto the paper and having mixing occur there or perhaps it could be done on the way to the nozzle …
I am simply making the observation that painters have produced a wide gamut of colours in a particular way for centuries. Could not digital imaging emulate this approach and perhaps gain some benefits? Is it really necessary for inkjets to pretend to be multi-pass offset printing presses, when really they are more like artists?
OK, I’ll give up there. Happy New Year! 🙂
Too complex a system Colin
Mark:
I thought it was simpler. Instad of having to work out what pattern of dots to print, the printer would just need to know how many doplets of each colour to shoot.
My point is that the colour rendering woul dbe more precise. The dots may be bigger, but the effective DPI could well be rather better.
It is a matter of expense, the primary difference is that printers ink is transparent and ment to go down in one pass, artists inks (or actually paint) is generally opaque and goes down in layers (thus the need for white).
For a printer to work the way you would want the amount of ink that was applied to the surface would be substantially more than what printers put on the paper now. You would have lots of problems with drying as well as supersaturation of the paper.
Commercial presses do occasionally use an opaque white ink (usually printing on colored paper) and it is much more difficult to control than the usual inks.
What you are after is not so much a higher DPI but a higher color gamut. Technologies on presses have led us to add other colors to the CMYK mix such as Orange and Green in the Hexachrome model. You can only make so many colors with CMYK even if you mix them before applying them, you need other colors to extend the gamut.
wrote in message
Mark:
I thought it was simpler. Instad of having to work out what pattern of
dots to print, the printer would just need to know how many doplets of each colour to shoot.
My point is that the colour rendering woul dbe more precise. The dots may
be bigger, but the effective DPI could well be rather better.
It is the print representation that worries me. I can understand that, with commercial printing, where there is a matrix of CMYK dots on the page, these merge to give the full gamut just like the RGB spots on a TV. But, ink jet printers don’t work that way. Do they?
Inkjet printers use Stochastic printing dots which are random micro-sized dots that do the same thing as the standard printing ‘rosette’ dots do, except they’re just not laid down in an organized fashion. There are offset printing methods that work this way as well and offer some advantages over the older method. (look up "stochastic printing")
I have the impression that they mix the colours together as the ink is sprayed onto the paper.
All ink printing methods are based on mixing light as the light bouncing off the white paper behind the ink passes through it and this is why we use the ‘light mixing RGB/CMY primary/secondary" colour model for both monitors and printing inks. The dots are each separated and the human eye mixes the coloured light together.
Second, what do artists think about this? They use 3 different primary paint colours: red, yellow, and blue.
These are opaque colours that light reflects off of and does not pass through (it’s paint..not ink). This is why the colour model is different and why it would not work for transparent ink. If you mix paint together then RYB are the 3 colours you need to mix to get black and painting color theory is based on what colours result from physically mixing the paint together. If you used the paint to create small dots next to each other then you would be back to a RGB or CMY world where mixing them optically together returns us back to the world of mixing light and not that of mixing together pigments (in a jar).
Hope this babbling helps make sense of something 🙂
Russell
Thanks for the further explanations Russell.
These are opaque colours that light reflects off of and does not pass through (it’s paint..not ink).
Yes, that’s true, but does it have to be that way? Any fundamental reason why an inkjet printer couldn’t deliver an opaque ink [or paint]?
Any fundamental reason why an inkjet printer couldn’t deliver an opaque ink [or paint]?
No. Except that they would still have to be coloured dots that the eye would blend together from a distance. I think some Epson printers do use opaque pigments and colour laser printers apply opaque toner. But if you wanted to actually mix the colours together before application it would be a very difficult challenge technologically and printing would be incredibly slow (while colours are mixed and applied 1 dab at a time). How would you clean the applicator of the 1st colour before contaminating it with a second or third.. I’m not sure what the advantage would be?
Russell
If you put an opaque dot on top of another, it would be the color of the top dot, not the combination of the two, so you would still have to do the dot placement dance. And it would be even harder, because you wouldn’t be able to overlap dots of ANY color. In inkjet and offset printiong, the inks are transparent, so they do add (subtract, really) together when they overlap.
I think your idea does kind of describe the behavior of dye-sublimation printers, which, as I understand it, vary the density of dots that are uniformly sized AND spaced. But still, these dots of dye are transparent, not opaque.
– John
Russell:
If the ink/paint were "wet" enough, couldn’t mixing occur on the paper [which may have to be particularly non-absorbant]? If all the "dabs" were squirted in quick succession …
Then you’d have to deal with increased drying time. This would be a particular issue with the paper transport mechanism, since you don’t want to get wet ink on the rollers. You could add a dryer fan, but this would add to the size and cost of the printer.
John
These are very good points. I think back to my first ink-jet printer – HP, of course – which needed at least 5 minutes for the output to be dry to the touch. But that was >10 years back and I suppose our expectations have changed.
I think your idea does kind of describe the behavior of dye-sublimation printers
no dots in dye sub. continuous tone.
no dots in dye sub. continuous tone
So, I just re-invented another wheel … 🙂
I was thinking of the pixels themselves as the ‘dots’.
I’m fairly sure Epson uses four dot sizes even in their stochastic.
