PPI vs DPI - the difference and effect?

[David Hicken]

It would seem to me that if PPI and DPI are both 1000, you would get one pulse per dot, but if you decreased the DPI to 250, you would get 4 pulses per dot. But the reverse with 1000 DPI and 250 PPI, I'm at a total loss of what it would logically do.
Am I completely off base? How do these work? And are the effects the logical effect of a 1000 PPI 250 DPI having deeper cuts more spaced out vs 1000 PPI 1000 DPI more shallow but closer together? Does anyone ever play with the PPI, and know what is the effect is?

Thanks,
David

[Rodne Gold]

PPI is relevant in vector cutting , DPI is only relevant in engraving.
DPI is the distance the Y axis moves down when engraving , IE if its 100 dpi , the head will scan back and forth 100 x in an inch of Y travel (normally the X axis would be set to fire 100 dots per inch of travel as well , the X axis will match the DPI you set for the Y)

PPI is the amount of pulses the laser will fire when traveling in vector mode , IE how many "holes" it will drill per inch of travel.
The higher the DPI and the PPI , the more the laser pulses will overlap and "burn" part of what its "burnt" before , thus leading to darker or deeper engraving

[Doug Bergstrom]

I am not sure how relevant this is to the laser but in the print industry DPI (Dots per inch) is how everything is being printed through a screen or on press or through any output device. PPI (pixels per inch) is the resolution of the digital file on screen. It takes approx 3 PPI to make a dot (DPI).

[Rodne Gold]

In this case PPI = Pulses per inch.

[Stephen Beckham]

David,

From my reading and understanding of the DPI... The picture shows the overlay of the dots as the laser fires. The 300 will not quite overlay each other and the lower you go, the more speckle you get on the edges. The higher you go in DPI, the more the line looks solid because the amount they overlap. This is also why the higher DPI gives you a darker burn on the wood (lesson I recently learned). Sample picture recreated from what Epilog shows in their documentation.

The first pass in green is to the right, and the return pass in Red to the left. Blue again towards the right and then yellow on the return back for the left...

Rodne hit the PPI on the nose - it is 'constant on' laser during vector cuts - how many times it pulses per inch. It could give you a dashed line if you lower it enough. One other note - on our Epilogs, the PPI is expressed in Frequency for Vector cuts.

It's not enough we have to figure out what we're doing, we have to talk different languages between laser print drivers to help each other... ;)


That's my understanding....

[David Hicken]

Stephen, fun how that works. Someone at some point aught to create a glossary containing terms, what they mean, and aliases for those terms amongst other manufacturers. Something else that would be very nice at some point is a look up table with laser models down the left and material types across the top, the cross sections containing a good starting point for power, speed, and dpi. I'd be happy to create something along these lines once I've had a bit more experience. I appreciate your and others replies and information.

[Mike Null]

Thre are a couple of additions I would make to the descriptions of PPI and DPI.

Going from 300 dpi to 600 dpi quadruples the dots--twice as many vertically and twice as many horizontally. The graphic above is very close.

Increasing the dpi from 300 to 600 increase the file size 4 times as well and may be very slow to get from the pc to the laser.

[Richard Rumancik]

PPI is relevant in vector cutting , DPI is only relevant in engraving.
DPI is the distance the Y axis moves down when engraving , IE if its 100 dpi , the head will scan back and forth 100 x in an inch of Y travel (normally the X axis would be set to fire 100 dots per inch of travel as well , the X axis will match the DPI you set for the Y)

PPI is the amount of pulses the laser will fire when traveling in vector mode , IE how many "holes" it will drill per inch of travel.
The higher the DPI and the PPI , the more the laser pulses will overlap and "burn" part of what its "burnt" before , thus leading to darker or deeper engraving

Rodne, I agree that with the GCC products an image is rastered with constant dpi in both axes. You said "normally" the x and y dpi would match – I didn't think that with the GCC products you could set them to be unequal. Is there some way that you can force unequal dpi in each axis for the GCC products?

You said ppi (pulses per inch) is relevant only for vector cutting, not engraving. Again, I would say "true" for GCC products. But ULS seems to use a system where dpi is the measurement in the "Y" axis and ppi is the measurement of "dot density" in the x axis. At least, this is what I have gathered when I read an old ULS manual I have around from a 1999 M-series machine. As well, what I see on the ULS and Rowmark sites seem to support this. The ULS website often refers to ppi in raster engraving articles. I do not own a ULS so perhaps some ULS owners can add some clarification if this is incorrect.

So I don't think we can say that ppi is irrelevant for engraving because the different laser manufacturers use different terminology.

I can see some value in permitting different "dot density" in the x and y axis if one were rastering some text or line art such as on a plaque. The reason being is that if I increase the dot density in only the x axis, I could achieve a darker/deeper burn for "no cost" i.e. no increased job time. So the ULS system where the x and y can be made independent seems to have some theoretical advantage. (However, with photos, I would think you would need to have identical values in the x and y axes.)

Mike, you said that doubling dpi increases file size 4x. I don't think that the time that it takes the file to get to the laser is the biggest issue – it is job time. There is no time penalty associated with higher dot density in the x axis, as the laser beam has to pass over each microinch of surface whether it fires or not. But in the y axis there is a time penalty, as double the dpi in the y axis would double the machine time.

I don't know which category the Epilog falls into – is it a ppi/dpi machine or a dpi/dpi machine? Stephen's sketches from Epilog suggest that it is a dpi/dpi machine, as the dot density increases at the same rate in both axes. What about the Accuris that David has? (Is it a GCC machine?) What system does the Trotec use?

David, you have the right to be confused by this subject – hope this does not make it any worse . . .

[Rodne Gold]

I can see no advantage whatsoever in setting X dpi/PPI independantly of Y and in fact can see some major disadvantages such as banding issues and the improper engraving of 1/2tone images.
You can "force" unequal DPI in either axis using the scaling command (sort of) but this would stretch the graphic either way.
Perhaps with other systems the PPI is settable as tho it were DPI in engraving mode but only really applies to vector cutting?

Here's instructions I took off ULS's site and I see in engraving they dont mention anything about DPI but only PPI , so perhaps that is then equivalent to our DPI?

We were able to cut and engrave the acrylic satisfactorily using a 50-watt laser system and a 1.5 lens with the following settings:
Vector Cutting Settings
•100 percent power
•1.5 percent speed
•1,000 PPI (pulses per inch)
•Image density 5

Raster Engraving Settings
•100 percent power
•30 percent speed
•500 PPI (pulses per inch)
•Image density 5


I think its all semantics ...........

[Mike Null]

Richard

You are correct. Increasing the dpi increases the engraving time and increasing it to 1000 dpi increases it a lot. But it can be troublesome as far as increasing file size that much in terms of sending the file to the laser.

[Rob Bosworth]

In the ULS systems driver, they use the term image density as the equivelent of the DPI (lines per inch). They use a slider bar with settings of 1 through 6 for DPI. Then you have a seperate control to set PPI.

In the Epilog driver, you can set your DPI, which is shown as equal (300 X 300 or 600 X 600) but you can also set the "rate" to help shape the pulse.

I believe pulse shaping is what most people are trying to do when changing the PPI or rate setting on their machines. (Oh boy, I'm not sure I have time to get too far into pulse shaping.) But basically, using a 25 watt CO2 laser, you will not get more average power than what the laser will output. So if you turn your laser on and it stays on for a second, it will only output 25 watts. Now, if you turn your PPI down, the shape of the pulse energy will start out very high, say 150 watts, but will drop off very quickly towards the end of the pulse. Over the duration of that pulse, you will still only have 25 watts of power. And if you turn you PPI up, the energy will be very even duing the duration of that pulse.

So if you are having trouble getting through a certain material, turning the PPI number down, might allow the laser beam to pierce the material with the initial blast of energy, and allow the material to cool or simmer as the energy level drops down towards the end of the pulse. As long as the PPI allows the beam to overlap, you should not see excessive striations.

On the other hand, you are cutting through something and you are getting a "raggedy" edge, turning up the PPI rate might smooth out the edge of the cut material.

(Sorry. I usually confuse myself with all this deep thinking. I am even worse on the golf course. :eek: )

[Richard Rumancik]

I can see no advantage whatsoever in setting X dpi/PPI independantly of Y and in fact can see some major disadvantages such as banding issues and the improper engraving of 1/2tone images.
You can "force" unequal DPI in either axis using the scaling command (sort of) but this would stretch the graphic either way.
Perhaps with other systems the PPI is settable as tho it were DPI in engraving mode but only really applies to vector cutting?

Here's instructions I took off ULS's site and I see in engraving they dont mention anything about DPI but only PPI , so perhaps that is then equivalent to our DPI?

(snip)

I think its all semantics ...........

Hi Rodne

You anticipate disadvantages with unequal "dot density" (as I call it for lack of common terminology) . . .

Why do you think banding would be worse? If a text graphic had more dots-per-inch in the x axis as the y-axis it is not obvious to me that banding problems would increase. I thought that "theoretically" the ULS could have advantages for some graphics, (as I have never tried a ULS machine), but we have lots of ULS users on this forum that should be able to offer some real-world experience. I'd like to know if there is an advantage (or disadvantage.)

Suppose I burned graphic at 100% power and 40% speed. If I want to go darker (or deeper), I can reduce speed, since I am maxed out on power. Reduced speed = more time. The other option is to increase dpi (dot density) in both axes (as they are not independent). Again, this increases time. (Twice as much is you double the dpi.) With the ULS, it would seem to me that I could double ppi (dot density) in the x direction (only) and presumably improve the darkness/depth without the time penalty. If this is true, why is ULS not selling this as an advantage to their system?

What I was saying is that for ULS, they use the term ppi for x-axis dot density and dpi (or resolution) for y axis dot density. This part is "semantics", but the independence between the two settings is a real functional difference which should be investigated. (I would do it if I had a ULS machine.)

I did note that for photos (halftones, PhotoGrav images, etc) that I would expect dpi and ppi to always be equal (unless you had some unusual image processing software.) The odd thing is that in the ULS manual that I have, they don't come right out and say "make sure dpi = ppi for photos." Maybe there is an automatic override for bitmaps?

Those with ULS machines have to read comments about ppi and dpi very carefully, due to the inconsistencies between manufacturers.

We can't say ULS's ppi = other manufacturer's dpi, as ULS uses both terms for engraving. From the ULS M series manual 1999:


"When engraving, it is advisable to use a PPI setting of 500 or higher." "For cutting, a PPI of 150 and above can be used".
"Resolution: This setting is the image resolution of the engraving in DPI or dots per inch. There are six DPI settings to choose from: 1000, 500, 333, 250, and DRAFT. The resolution can also be referred to as the vertical lines per inch or fill spacing. This is how many raster strokes per inch of travel that the motion system steps down to produce the engraving."

I don't have a more current manual available, but from what I see on various websites I don't think the ULS temoinology has changed.

[Mike Null]

I had a ULS and could not set dpi independently nor do I think that can be done now.

[Rodne Gold]

If you lets say halved the y axis with repect to the X , you would essentially be skipping a line of dots , giving rise to a ploughed field effect rather than a level plain. This would exhibit as some banding , the image would be composed of a lot of stripes.
To add to what Rob said :There are no free lunches in lasering or power , if you are maxed out in power , and want to go deeper/darker , you have to go slower. Power , like 1 watt is 1 joule of energy per second. If a lasers output is 25 watts , it is capable of outputting 25 joules per second , to equal the energy of lets say 50 watts , that would be needed for the deeper/darker burn , you would have to need 2 seconds over the same area.
You have now got to look at the laser traversing an inch in lets say 1 second. The lasers output can only be 25 joules in that second , so no matter how many times you switch the laser on or off , when at full power you wont be able to put any more depth or darkness in that inch.
The issues of depth or darkness are actually also related to the materials , spot size etc. Darkness is normally due to heat affected zones cos a laser does not burn , it vaporises , but the area around the vaporisation is heat affected and thus chars.
Depth would be more related to power as a higher thermal shock would lead to better vaporisation.
You might do better in terms of time and lets say darkness of wood engraving , by using a lens with a bigger spot size and reducing your DPI.
An analogy to that is burning a hole in paper , if you use a bic lighter with a far bigger flame than a pencil blowtorch but the same energy in the flame , you would get a much bigger char than doing it with the pencil torch where you will get a neat clean hole.

[Scott Shepherd]

Let me start by saying that my knowledge is so limited that I can't speak technically on the matter, but I will share my real world experience.

I do a lot of reverse engraved material. It's probably been 80-90% of our business. On the Epilog with DPI, we tried every possible combination and continued to get very excessive banding problems. I was constantly told that it was the material. I took the same file, same material to a ULS rep and first pass had banding, second pass with changing the Image Density (Y-Axis) and PPI (X-Axis) settings, the banding all but disappeared.

I find that being able to adjust the Imagine Density (Y-Axis) and PPI (X-Axis) independent of each other allows me to fine tune things that really help me. I have kept Image Density on 5 and bumped the PPI down to 150-200 and I've managed to get some very clear burns, where as before, with a constant, the result was cloudy.

I don't know how it works or why it works, but I do know I use those two settings to fine tune things that really help my output from quality to time. I do know it makes a big difference on the work I do.

[Richard Rumancik]

Scott:
Thanks for your comments on this. I was hoping to get some input from some actual ULS users and would like to hear more comments from the field . . . we now have at least one person who has observed an advantage with independent horizontal and vertical dot density. If it is a good idea then I'd like to give them credit - if it is not, I'd like to know why it isn't.

However, I am a bit confused as to how the input from the "image density" slider is mapped to an actual dpi. Since with most lasers you normally have discrete choices of vertical resolution (dpi) I wonder why ULS implemented this slider control. If the lowest is say 100 dpi and the highest is 1000 or 1200, it is not clear to me why they would "hide" the dpi parameter by making a slider control. Sometimes attempts to make an interface more "user friendly" tend to cloud what is actually happening, and it also makes exchanging infomation more difficult.

If you are engraving a photo on a ULS machine I would think you would want to match the "PPI" to the photo. I don't quite understand how this aspect is handled in the ULS machine.

[Mike Null]

As far as the slider control is concerned it didn't exist when I bought my ULS machine in '97-'98. But in subsequent upgrades to the drivers they incorporated the slider concept though I don't recall that it was called image density.

My guess is that ULS was trying to minimize confusion owners had between laser dpi and photographic dpi.

[Scott Shepherd]

Richard, I do believe there is a conversion chart somewhere. I don't have the manual in front of me, so I can't give specifics, but it was something like the 1 was equal to "X" DPI in the Y-Axis, and 6 (being the top of the scale) was something like 1200 DPI. 5 was something like 600 DPI. I'll post the actual once I can find it, assuming I did see it in the manual rather than someone telling me that in person.

I can't comment on photographs, as I don't do them except to play around and I've done so few on various materials, that I really can't comment.

I do mess with those two settings all the time. I can't explain why it works, but I do know that I am able to adjust speed and quality related settings by manipulating them. I like keeping the Y-Axis number up (on 5) and then going down on the DPI for the X, which appears to introduce a lot less heat into the object. I do that for anything that has a larger burn area.

Let me be clear, I'm a hack, so anything I'm experiencing may be for other reasons.

[Alan Young]

For my two cent here is what is in the manual on a GCC Mercury 30. I most alway use Auto. I have seen some banding when doing large area engraving, say for inlaying

Alan


P.P.I.
P.P.I. stands for the laser firing pulses per inch.



Purpose: By changing the laser firing pulses per inch, the distribution of

energy would be changed.




Description: The adjusted range is 30 to 1500 PPI or you can choose auto

mode. When choosing auto mode, the system will come up with the proper

PPI value automatically. In order to maintain the quality of raster graph, the
PPI value will at least equal to DPI value automatically even you set the lower
PPI value than DPI value. In 100W machine, the PPI value of raster function always is auto.

[Jerry Allen]

According to GCC techs when asked, PPI will affect the raster engraving although that was not the original intent, which was for vector only. I got no specifics about how it affects it. I took it to mean that since you could raise the PPI over DPI (but not lower it), that you could increase the pulses across X, while the line spacing in Y would reflect the DPI setting. I have not tried to validate that theory yet. I respect Rodney's experience and knowledge. The the GCC manuals and their derivatives have contradictory or poorly written statements. Add to that the direct statement from GCC and you end up with a lot of confusion.

[Jerry Allen]

Richard,
I've seen that statement as well as the one quoted above by Alan.
Seems to be a discrepancy between the two. Why would one set DPI for engraving and then even change the PPI if it's for vector only?
I just set the DPI and use Auto for PPI which seems to do what I need regardless of whether I'm cutting or engraving. Someday I'll test it, but if anyone can answer your question I'd sure like to hear it.

(somehow this message is showing up before Richard's which it shouldn't be)

[Richard Rumancik]

Jerry: I agree that the GCC manuals are not the best. When I bought my Mercury 6 years ago the manual was poorly written; at that time GCC was just getting started in North America and I gave them the benefit of the doubt. It is a bit disappointing that they have not put a bit more effort into the manual. Some companies don't realize how a poorly translated document reflects on their company and equipment.

The official statement in the Mercury manual reads:

"PPI – Pulses Per Inch. This setting should not be confused with DPI. PPI determines the gross amount of laser pulses there will be per linear inch. PPI is exclusively for the vector setting. A PPI setting of 500 results in the laser firing every .002" (500 times per inch). If the standard lens is producing a vector laser focal point of .007", then higher PPI settings will result in deeper, overlapping laser pulses. PPI settings lower than 150 will result in the individual laser pulses being spread far apart, so they will not touch each other. Low PPI settings are a good example of how the LaserPro Mercury can perforate paper. "


I don't know what to make of the statement the GCC tech made regarding ppi affecting raster. I can't say that I have actually tested it, but have always assumed that with the GCC there was no effect, based on the above paragraph. As a result I have never experimented with the setting.
Maybe they liked what ULS did and implemented it without properly documenting it?

If anybody else can support or refute the GCC comment I would like to hear about it.

[Mike Null]

With Trotec PPI is a raster setting as well as dpi and they are independent. On vector the there is a hertz setting and PPI as well dpi are not operable.

[Alan Young]

Richard

What is the date of your manual. Mine is dated Sep 11 2000. It is interesting the differences. I will agree the manuals are a little to be desired.

Alan

[Richard Rumancik]

Alan, when I wrote that post I did not have my manual handy so I took it out of a copy that Jorlink posted on their website. The Jorlink Mercury manual does not have a date on it. You can see it at

http://www.jorlink.com/page.aspx?page_id=49

It is 83 pages so takes a few minutes to download.

There is info there for other GCC products as well.

[Richard Rumancik]

If you lets say halved the y axis with repect to the X , you would essentially be skipping a line of dots , giving rise to a ploughed field effect rather than a level plain. This would exhibit as some banding , the image would be composed of a lot of stripes.

Rodne, you are looking at it the opposite way that I am. I will not halve the density in the y axis and skip any dots. I will set the y axis to the number of dots I need for the application. Then I will DOUBLE the number of dots in the x axis (by setting the ppi to 2x the dpi). So I have MORE dots in the y axis than I would have otherwise. If 300y and 300x generates an acceptable image, then I would suggest that 300y and 600x has the potential to look even better. I brought up "deeper and darker" as possible outcomes to higher dot density in the y axis (which you don't seem to agree with), but forgot to include the subjective issue of "image quality".

This "trick" to improving image quality was implemented a long time ago on early dot matrix printers. The printer could be set into a higher quality mode where it would offset the dot placement by half-a-dot and thereby improve the text quality. I had a printer which printed the offset dots on the reverse pass (right to left motion) when set to highest quality. Very similar to overlapping dots in a laser engraver. On the dot matrix printer the dpi in the x did not need to equal the dpi in the y.

Can't comment on the energy analysis at the moment. . . maybe later . . .