Can someone explain DPI?

[DigiPrint]

What is the difference between 2400x2400 dpi and 1200x1200x8? What is better? I am trying to understand all this, but it is confusing me. Canon and Konica are both trying to get in more door telling me that their IQ is higher, Canon touting 1200x1200x4 on the C6000.

Can someone explain this? What is the difference between the bit depth vs true resolution. Another sales rep told me bit depth was kind of like analog zoom vs digital zoom on a digital camera.

Thanks.

 

[gordo]

What is the difference between 2400x2400 dpi and 1200x1200x8? What is better? I am trying to understand all this, but it is confusing me. Canon and Konica are both trying to get in more door telling me that their IQ is higher, Canon touting 1200x1200x4 on the C6000.

Can someone explain this? What is the difference between the bit depth vs true resolution. Another sales rep told me bit depth was kind of like analog zoom vs digital zoom on a digital camera.

Thanks.

First off - "DPI" is not resolution. It is addressability. I.e. it does not tell you the size of the mark the device makes. Instead it tells you the spacing of the grid that guides the droplets, toner, or laser energy.
This graphic shows how it works for laser exposure in CtP - the same principle applies to printers:

Note that, in this case, the addressability grid is 2,400 dpi - that is that they all can hit the target location (address) with their beam of energy - however the exposing spots of energy (could also be splats of ink or toner) are all different sizes, in this example ranging from about 2 microns on the left to about 30 microns on the right.
So:
Addressability = The number of discrete locations along a line that a printer can place dots
Resolution = The smallest feature that can be imaged (Often confused with addressability)
Increasing addressability without decreasing ink drop/toner particle size can improve lines and curves by smoothing out the roughness caused by the discrete raster locations, but it does not allow finer lines or features to be resolved and does not improve the true resolution. Even with perfectly placed dots, a printer with a 60-micron dot cannot image a line finer than 60 micron regardless of how high the addressability is. In reality, drop/toner particle placement errors increase the actual width of the finest line to well above that of the actual minimum dot size.

Bit depth is used to increase the levels of grey, or tones, the device can deliver. It is similar to the use of light cyan, light magenta, and light black inks in inkjet printers. It's especially useful if the splats of toner or ink are large since making light tones with large splats would make for a grainy appearance.

Bit depth is desirable in case of low addressability but is not required if using small drops/toner particles with high addressability. For contone images or photos, being able to use bit depth is like doubling or tripling the printing resolution with little or no impact on throughput.*

The best thing would be to do a test using an image with gradients, very small type, etc. to test the quality yourself.

best, gordon p

 

[killer1loop]

Provided that every sales rep is willing to boost his own product and technology, basically everyone is correct. You should make up your mind on your specific needs.

At your perception, nowadays digital prints all look good and if not explained maybe you can't focus on any difference.

This is also true talking about end customers. Most of them could not really fancy about 2400 dpi 1 bit or 1200 dpi 4 bit or 600 dpi 8 bit prints. This is because from the point of view of eye perception results are generally the same.

A higher res low color depth print is generally better for small texts, details, lines under 1 point and general artwork definition. Counterpart is that photos, gradients, spot backgrounds uniformity is generally better on higher color dept prints.

Talking about marketing claims, it is easier that 8 bit dept machines are marketed for the photo finishing industry and glamour/fashion high quality print. High res printers are often compared to traditional offset presses because prepress machinery work at 2540 dpi...

Konica Minolta digital presses are very good in quality and bit depth, even if they are 600x1800 dpi. Also Xerox iGen 3 and 4 have the same res and bit depth.

Hope you did not get confused by numbers, if you need more info feel free to ask!

 

[PineyBob]

If I were trying to decide on either Canon or K-M I'd do the following.

If you're running an EFI front end calibrate to the appropriate specs that you plan on using

Run the GATF test charts on both units on a variety of stock.

take some customer files and create samples, finished if need be.

Make your determination as to which device most meets your quality standards.

Take the customer samples to the customer and do a blind comparison.

Bluntly put, who cares about the DPI wars, can you sell the output to your desired customer base. To me "sellable output" in the differentiator here not bit depth or anything else.

 

[Digital Press Guy]

Printability is about more than DPI and bit depth. The overall look and feel of the imaging. How well does it handle different substrates? Your personal tastes or preferences or that of your customer's matter as much or more.

At the end of the day can your engine produce salable product?

 

[werner61]

Exactly, Repeatability is key, do the colors shift at 100, 500,1000? Your customer could careless about DPI they just want a product that looks good time after time and matches every time.

 

[David Ingram]

Its Important to know for an offset plate, coming from a digital imager such as a platesetter or film imagesetter, that 1200 DPI resolution is Insufficient for Line Screen Frequency over 120 LPI
and may also show jaggies in very fine type details as well, that may take a trained eye to see.

Take a capital Letter T in GOUDY for example, at 1200 DPI and say 144 point for the letter, you CAN see a difference between the lower and higher dpi on proof, on and on final output. Especially across the top of the letter. Some digital presses like the Heidelberg DI have a mode where they can do 150 LPI a a 1200 DPI resolution. (It images the plates FASTER this way) but for 133 LPI + quality, you really MUST go to 2400 DPI to avoid seeing those type jaggies mentioned earlier.
if your LINE work is really critical I say go w 2400 - why lose a client over jaggy output?

You may need a LOW LPI for your printing conditions, IE Web press, printing on porous paper, high dot gain etc. However, there is no reason you should not do your imaging for PRESS at at least 2400 DPI. Whether your frequency is 55- 85 or 175.

 

[werner61]

I'm not sure but neither Konica nor Cannon print Line screen dots on their toner based systems. But what is more interesting is what does the x4, or x8 stand for? Are they using 4/c vs. 8/c toner systems? Or is that a resolution or a type of interpolation? Maybe Gordo can help me out here(hey Gordon) you are/were the king of dots...

 

[gordo]

I'm not sure but neither Konica nor Cannon print Line screen dots on their toner based systems. But what is more interesting is what does the x4, or x8 stand for? Are they using 4/c vs. 8/c toner systems? Or is that a resolution or a type of interpolation?

I can't find any reference to X4 (or X8) on Canon's web site nor in their product specifications nor product literature. So I don't know what X4 means. Can you post a document where they say 1200 x 1200 x4?

best, gordon p

 

[gordo]

Its Important to know for an offset plate, coming from a digital imager such as a platesetter or film imagesetter, that 1200 DPI resolution is Insufficient for Line Screen Frequency over 120 LPI and may also show jaggies in very fine type details as well, that may take a trained eye to see.
[SNIP]
You may need a LOW LPI for your printing conditions, IE Web press, printing on porous paper, high dot gain etc. However, there is no reason you should not do your imaging for PRESS at at least 2400 DPI. Whether your frequency is 55- 85 or 175.

I would disagree.
One can certainly do 120 + lpi on a platesetter at 1200 dpi. The halftone dots will appear a bit "chunkier" but that will not affect final print quality.
High dot gain and/or printing on porous paper (e.g. newsprint) does not force the printer to use low lpi screening. For example, about 70% of all newspaper inserts and 80% of all telephone directory printing in the U.S. is printed with FM screening at an approximate screen ruling equivalent to 325 lpi on pretty nasty paper.

best, gordon p

 

[werner61]

Maybe it has something to do with cost...;-) Or the salesperson just made it up.

 

[UberTech]

This really needs to be a sticky.

This really needs to be a sticky.

The x1, x4, x8 is the bit-depth. This is how many grey shades can be created per dot. x1 is 1, x4 is 128 and x8 is 256. The 8000 is different again as it can address 3 sections of each dot so in theory is 3600DPI. This is all irrelevant as all files are processed at 600DPI regardless.

 

[X33]

Digiprint, I bet you are more confused than before. I am after all these explanations.

Here's the point, no matter what the DPI is, the important thing is if the output is acceptable by you and your clients. Is it?

There are many ways that every vendor will tell you how mathematically and theoretically one DPI is better than the other.

I would take a couple of files and print the same file on all the machines that you are looking at. Probably select something with shadows, skin tones, line art, etc.

 

[UberTech]

I would take a couple of files and print the same file on all the machines that you are looking at. Probably select something with shadows, skin tones, line art, etc.

...and lots of prints with big solid dark areas, like a dark dark blue. Then some lighter ones with reverse text.

 

[werner61]

Actually, I thought about it and I think I have figured it out... It's all the ex-Creo reps who are now at Canon and Konica and they have just revamped the old Film to CTP jargon they used back in the day. LOL! Sorry Gordo I couldn't resist;-) I cannot believe I said Back in the Day though...

 

[Dan Bergen]

It is very simple and the easiest way to explain it is a "bit" is used to describe each "dot" of color. The
more bits, the better the description. A bit depth of one can only be "on" or "off" which is the
equivalent of white and black. Example: In a 2-bit image, there are four possible combinations
of 00, 01, 10 and 11. In this manner, 00 would represent black and 11 would represent white, while 01
would equal dark gray and 10 would equal light gray. Even though the bit depth is two, a maximum
of four tones can be generated. Let's put the 2-bit example into a four-color printer using CMYK. With
each color, a 2-bit color printer could print a range four tones for each color or a total of sixteen
tones that is the low end for color printing. Even with a high print resolution the human eyes would be
able to determine that this is not an optimum color print. Nevertheless, several vendors use
2-bit color printing. If you are following, you will notice how everything doubles as we go up the
bit-depth ladder. We started with one-bit that gave us two tones. Then we went to 2-bit which
gave us four tones. Now we're going to 4-bit color depth and that will give us how many tones?
And the answer is--16 tones. With 16 tones for each color, pictures begin to look like pictures instead of bad Matisse paintings. In the four-color process, 4-bit printing now makes 16 tones available for each color or a total of 64 colors and 65,536 color possibilities! There are a couple of vendors who use this bit depth with their color printers. Combined with high-resolution (600dpi or more) these prints look good. each color, a 2-bit color printer could print a range four tones for each color or a total of sixteen tones that is the low end for color printing. Even with a high print resolution the human eyes would be able to determine that this is not an optimum color print. Nevertheless, several competitive printers use 2-bit color printing. If you are following, you will notice how everything doubles
as we go up the bit-depth ladder. We started with one-bit that gave us two tones. Then we
went to 2-bit which gave us four tones. Now we're going to 4-bit color depth
and that will give us how many tones? And the answer is--16 tones. With 16 tones for each color,
pictures begin to look like pictures instead of bad Matisse paintings. In the four-color process, 4-bit printing now makes 16 tones available for each color or a total of 64 colors and 65,536 color possibilities! There are a couple of competitors who use this bit depth with their color printers.
Combined with high-resolution (600dpi or more) these prints look good. When the bit-depth is low, details are lost. Solid areas become featureless and fine details disappear. The mark of low bit depth is the loss of details in the shadows and highlights. Good color rendition shows details into the darkest
shadow as well as brightest highlight. When working in the world of computer files you will run into instances of 16-bit color, 24 bit color and even 32-bit color. Don't let thee terms confuse you; they are simply additives of the number of tones used in the color process. Example: 24-bit color is simply
8-bits for each additive color channel or RGB, while 32-bit color is 8-bits for each subtractive color channel C, M, Y and K. These terms are used for color file generation--not printing!

 

[Eyal Gragir]

to make it simple:
calculate the amount of combinations of pixel values . the higher the better.
the equasion is:
Xdpi * Y dpi * gray level#. thats it !
but,
this formula assumes that the printer vendor makes all combinations usable. if the dot or spot is too big when you print 50% you might get a full covergae and max density, then all above 50% is useless. this is known to all vendors and they usualy adjust the printers accordingly to get the best performance.
example :
1200x1200x4 = 2400x2400. it is that simple.

 

[gordo]

to make it simple:
calculate the amount of combinations of pixel values . the higher the better.
the equasion is:
Xdpi * Y dpi * gray level#. thats it !
but, this formula assumes that the printer vendor makes all combinations usable. if the dot or spot is too big when you print 50% you might get a full covergae and max density, then all above 50% is useless. this is known to all vendors and they usualy adjust the printers accordingly to get the best performance.
example :
1200x1200x4 = 2400x2400. it is that simple.

Hmmm. Keeping in mind that, as far as I can find, there is no reference to X4 (or X8) on Canon's or Konica's web site nor in their product specifications nor product literature. Nothing where they say, for example, 1200 x 1200 x4. So it is all speculation as to what is meant 'cause, so far, we have no reference.
That being said, they probably mean "greyscale" capability. Grayscale = Variable optical density per pixel.
There are several ways to achieve greyscale capability - none of which change addressability - which is what 1200, or 800, or 2400 dpi means. Grayscale is desirable in case of low addressability but not required if using small drops with high addressability. For contone images or photos, although grayscaling does not change the actual resolution it is like doubling or tripling the printing resolution with little or no impact on throughput.

The basic grayscale methods. The grid is the addressability grid - what is incorrectly referred to as "dpi"

From the left:
• Creating drops of variable size. For example, in one device, the drop size can be either 5pL, 11pL, or 27pL (pL = picoLiter).
•Printing more than one drop per pixel by causing drops to merge in flight to create larger drops.
• Using lighter inks, e.g., use of CMYKLcLm

For text and line work, grayscale by using multiple drop sizes also provides gives the ability to smooth curves without having to increase addressability (incorrectly referred to as "dpi").

Grayscale capability is desirable in case of low addressability but, provided that the drop size is small, is not required on devices with high addressability.

Again "DPI" does not indicate resolution or the minimum size of a mark that can be made.

best, gordon p

 

[Eyal Gragir]

Gordon,
thanks for the very learned explanation.in my opinion the phrase " keep it simple" is worth sacrifying the 100% accuracy . indeed it is only 80% accurate but works for most cases and VERY VERY VERYSIMPLE.
regards
EG

 

[gordo]

Gordon,
thanks for the very learned explanation.in my opinion the phrase " keep it simple" is worth sacrifying the 100% accuracy.

I disagree.

If information is inaccurate then, whether it is simple or not, it is no longer information - it is disinformation and will likely have people making the wrong decisions, being mislead by marketing hype, or unable to ask intelligent questions of their suppliers.

I have tried to keep the explanation simple enough so that the average printer would understand it while still maintaining accuracy. If I have failed, please let me know and I will try and explain the concept a different way.

best, gordon p