DPI and PPI with respect to power, speed and output

[James Terry]

I am doing some test learning tonight before I engrave my first plaque with the laser and I have found that when you select a higher DPI in the driver, the output burns more. Perhaps that makes no-brainer sense, but it isnt something I have thought about until now so I was hoping you guys might give me a nutshell brief on how this all inter-relates and works to provide best output.

In this example, the source image is 300dpi and the left output was selected in the driver as 300dpi, 40,100% and the right is 600dpi, 40,100%.

http://www.webglider.com/laser/2010/DSC06651.JPG


Plainly I see more burning at 600dpi, so is the driver interpolating the extra pixel points then?

Did I slow down the output by selecting a higher dpi?

Is there a benefit by selecting a higher dpi and then using a higher burn speed?

When does dpi matter? When does ppi matter and should I just be leaving mine on auto for the time being?

What should I be targeting for dpi of my artwork?

Should the art and the print always match?

In the examples above, I prefer the darker burn on the right but I could do without the overburn/surrounding marking. To correct, should I just speed this up, select a lower dpi, select a lower power...? How does one decide how to load all of the variables to get an optimal output? It seems to me that varying just one of the variables would have similar results as varying just one of the others.

[Niklas Bjornestal]

If you do double the amount of dots you'll get the double amount of power.
It doesn't interpolate the pixels, but uses the closest pixel. It works the same as for example a laser printer, they (usually) use at least 600 dpi when printing.
It will take twice the time to run the jobb, each line will engrave with the same speed, but you'll need to make twice the amount of lines.

The resolution of the art and engraving doesnt have to match (except maybe when using photograv).

PPI is the same as DPI but for cutting (at least on laserpro's machines).

[Richard Rumancik]

If you do double the amount of dots you'll get the double amount of power.

I agree with this statement.


. . .It doesn't interpolate the pixels, but uses the closest pixel. It works the same as for example a laser printer, they (usually) use at least 600 dpi when printing.
It will take twice the time to run the jobb, each line will engrave with the same speed, but you'll need to make twice the amount of lines.

Not so sure what you are saying here, though, Niklas. First you say that it uses the closest pixel as it won't interpolate. But then you say that it will require twice the number of lines. Isn't that interpolation?

If it only used the closes pixel, then plotting a 300 dpi artwork at 600 would result in hitting each pixel twice, which is equivalent to engraving the entire image at 300 dpi twice. I had in my mind that when an image was plotted at twice resolution, that it would be equivalent to resampling the artwork to twice the dpi in Corel. That "fills in" the missing pixels, although it may introduce some jaggies.

Can you clarify?

[Mark Winlund]

This doesn't entirely answer your question, but when viewed under a microscope, I have noticed an interesting thing: quite often, an engraving in wood that looks dark is the result of light getting "lost" in the vertical "canyons" created by the narrow laser beam. When examined very closely under a microscope, the surface color due to charring is actually the same from sample to sample.

Imagine a piece of wood engraved with a single line to a depth of 1/4". When looked at closely, the line looks very black. The light entering this "grand canyon" simply reflects off the walls multiple times and gets absorbed before it can get out again.

To create this effect, the laser must be sharply focused to create the "canyons".

To test this, I have done a series of engravings with rastered letters using different levels of "quality" (Universal's term for the distance between scan lines). In every case, the lower "quality" setting produced a darker looking image. Under the microscope, there was no difference in surface charring. The power and speed settings were the same throughout the tests.

So, a tentative conclusion would be that some of the differences in apparent "darkness" are due to optical effects and not charring of the wood.

Please note that I am not saying that the wood species has no effect, just that there may be other factors at work here. The difference in surface charring between maple and alder is quite obvious, for example.

Regards, Mark

[Rodne Gold]

More dpi = more overlapping of dots , the laser ALWAYS works by "dots" , even when cutting.
The reason you get a darker burn is cos by overlapping you dont just vaporise the wood , you create large heat affected zones which gives you the branded look. A high power laser will actually engrave wood a lot cleaner with almost no char as it vaporises more effectively.
You can go a teeny bit out of focus and use a lower DPI (for more speed) and you should get a shallower but darker "burn" Using less air assist when rastering wood could also give a darker "burn" - soft resin free woods (like balsa) wont engrave dark no matter what you do cos its the resins in the wood that burn and give you the darkness.
PPI matters when vector cutting , it sets the amount of pulses or "Dots" per inch of travel - too little and your cut looks like perforations , too much and you load too much energy into the cut resulting in either scorching or bad melting.
Artwok DPI and printing DPI are different things , in fact artwok DPI is actually NOT Dots per inch , but pixels per inch. A pixel is interpreted by the printer (laser)
To give an example , lets say a pixel was interpreted as a 3 x 3 dot (laser dots) element , then a pure black pixel will have all the 9 cells "coloured" in , a light grey might just be one cell coloured in. In reality the laser cannot possibly resolve more than 100-200 pixels per inch in the original picture and it is relatively useless to use more than that. DO NOT confuse the DPI a program like Corel says a picture is with the ACTUAL pixels per inch it contains.
If a picture is 50 pixels by 50 pixels , then a 50 dpi OUTPUT device will display/output it as 1" x 1"
A 200 DPI output device will output it as 1/4" x 1/4" at its native resolution.
I am being simplisitc here to illustrate the point , there is more to it like dither and 1/2 tone screens etc etc etc.

[Richard Rumancik]

. . . In reality the laser cannot possibly resolve more than 100-200 pixels per inch in the original picture and it is relatively useless to use more than that. . . .
Rodne, I'm confused here . . . you seem to be saying that you would not use a setting on the laser higher than 200? If the beam diameter is .003 and you wanted the dots touching, you would need 333 dots per inch of laser travel. But it looks better if the dots overlap. I don't see how you can get a good image at 200. You would have some serious jaggies on text at 200 dpi.

[Richard Rumancik]

Plainly I see more burning at 600dpi, so is the driver interpolating the extra pixel points then? I thought it would, Nikolas says no; let's wait on this one . . .


Did I slow down the output by selecting a higher dpi? Yes, it would have taken twice as long as it is plotting twice the vertical lines. It will not increase the time required for a horizontal line as it is just doubling number of laser pulses in that line.


Is there a benefit by selecting a higher dpi and then using a higher burn speed? In my opinion it will smooth the edges, as there will be more overlapping "dots". Keep in mind that your dots will be likely .003" or larger (which is a limit on the laser technology and lens used) so there is a limit that can be achieved with this.


When does dpi matter? When does ppi matter and should I just be leaving mine on auto for the time being? The required DPI (or dots per inch vertical & horizontal) is subjective. Basically, if the image looks good to your customer, it is good. Some of us tend to be more fussy than our customer. But there is no scientific measure for raster engraving, it is based on visual results. I don't use auto as it hides my settings. I don't think you will learn what the laser is doing by using auto. I'd rather tell it what to do, and if I was wrong at least I know. Some people prefer the "black box" approach - let the laser and software make all the decisions but I don't use that approach.

BTW - Does your laser measure dots in both vertical and horizontal directions? I assume so based on what you asked. Here's the problem - many lasers use the term ppi to mean pulses per inch and it is applied to vector travel only. It defines how many thimes the beam "flashes" along the curve that is being cut.

But some manufacturers like ULS seem to use a different terminology. For raster images, they call the horizontal spacing "ppi" and the vertical spacing "dpi". Newer ULS machines are even more confusing - they replaced the (vertical) dpi setting with a term called "image density". This makes it even harder to speak in general terms, as ULS seems to have their own language. So be careful when reading on this topic.


What should I be targeting for dpi of my artwork? I would say a minimum of 300 dpi for plaques (text) but I often use 500 or 600 dpi for text and/or artwork. Rodne seems to imply this is too high but my eyes tell me otherwise.


Should the art and the print always match? I don't think it has to, but if they are not multiples it will not plot well. ie. you can plot a 300 dpi graphic at 600, or a 600 dpi graphic at 300, but I would not plot a 600 dpi graphic at 500 as it will try to convert on-the-fly and you may end up with a poor image.


In the examples above, I prefer the darker burn on the right but I could do without the overburn/surrounding marking. To correct, should I just speed this up, select a lower dpi, select a lower power...? How does one decide how to load all of the variables to get an optimal output? It seems to me that varying just one of the variables would have similar results as varying just one of the others.

Is all the smoke residue cleaned off? It appears that there are stains. You need to use coated wood so you can wipe the smoke residue off. You could probably reduce power and go a bit lighter. I generally try to raster at 100% speed to reduce job time, and adjust other parameters accordingly. How do you decide? Well, you need to experiment some more and record your settings (on the sample) until you get a feel for things. There are more than one set of parameters that may work for a job.

On some woods you just can't get a dark contrast using power alone. It will look too charred before it darkens. In these cases, you need to use something to darken the image. That's another topic.

[Niklas Bjornestal]

Not so sure what you are saying here, though, Niklas. First you say that it uses the closest pixel as it won't interpolate. But then you say that it will require twice the number of lines. Isn't that interpolation?
It will make 2 identical lines. When interpolating you guess how the missing pixels would look.

For exampel:
original image: "300dpi"
xxxxx
xxx
x

image, 600dpi, not interpolated: (.="added pixel")

x.x.x.x.x.
..........
x.x.x.
......
x.
..

image, 600dpi,interpolated

x.x.x.x.x.
........
x.x.x.
....
x.
.

[Martin Boekers]

Here are a few pages from Epilog, visuals are sometimes easier to work with.

http://www.epiloglaser.com/tl_resolution.htm

http://www.epiloglaser.com/tl_dithering.htm

http://support.knowlton.ohio-state.edu/files/Epilog-manual.pdf
(see pages 72 & 73)


Marty

[James Terry]

Well I did some more testing on some waste veneer panel and decided to go with 600dpi, 50sp, 100po onto this basswood plaque. It came out pretty good. I'm trying to decide if I should seal it before I send it off or just keep it natural which I know she'll love.

http://www.webglider.com/laser/2010/DSC06656.JPG


A friends husband died last week so I decided to engrave his actual wedding proposal into something nice. He was an engineer type if that didnt come through in the design. They are simple and woodsy folk so the bark border really makes this perfect for her.

One thing I learned is that it is time consuming trying to clean up handwriting for engraving, especially when sourced from a low quality jpg.

The laser also plainly points out any stray dots. Even a dot just one number off of white that you dont notice will be lased. So cleaning up the art work goes a long way towards a better end product.

And working with high res art is very important.

[Martin Boekers]

A few tips if you seal it before you laser, then you can clean the "sappy" wood burn easily with a Magic Sponge and some Endust.

If you take the jpg into Photopaint, or Photoshop, convert to greyscale, resample at a higher resolution this may help.

Also after that you can open the levels drop down and work with getting a nice contrast range. I always boost the white balance a bit that takes care of the random dots most of the time.

Use the eye dropper tool and check various areas around the artwork to make sure you don't have a slight background that may not be visible.


Hope this helps.

Marty

[Rodne Gold]

Richard , the laser can resolve at most 300 dots per inch based on beam diameter and thats assuming there is no dot gain , ie the dot "printed" is actually 0.003" without any adjacent heat affected zone. The problem is that a pixel is hardly ever a dot , it is generally a shade of grey and a shade of grey cannot be interpreted by a single dot , it has to be a matrix. I said the laser cant really resolve more than 200 pixels per inch , it can print as many dots per inch as it is allowed , thus photos with more than 200 pixels per inch are a waste of time.
DPI is not the same as PPI - dpi always refers to an output device , its VERY rare to have a single colour printing device able to map a pixel on a 1:1 basis
Often clicking on a picture on screen in corel , it says its 72 or 96 dpi , but it can contain many more pixels per inch which is the only thing that counts in terms of resolution and the size it can be printed.
When scanning , the DPI term is misused , it is actually scanning at PPI , not DPI. It's always better to scan as high as possible as this will determine the max size of the output and will allow you to crop or use a smaller section of the image with no apparent loss of resolution.

[Richard Rumancik]

. . . The problem is that a pixel is hardly ever a dot , it is generally a shade of grey and a shade of grey cannot be interpreted by a single dot , it has to be a matrix. I said the laser cant really resolve more than 200 pixels per inch , it can print as many dots per inch as it is allowed , thus photos with more than 200 pixels per inch are a waste of time. . . .

I understand what you are getting at, but I don't think you can say that a pixel is hardly ever a dot. For photos, yes, but when laser engraving a lot of images are line art and text, a great majority of it is b/w, one-bit, binary, or whatever name you want to give it. In this case a pixel is a laser dot. Since the laser can only "print" one-bit images most files are either created as one-bit, or converted to one-bit, before lasering. The image James showed was basically one-bit line art, so I was not thinking about photos. I guess I was trying to figure out what you were recommending to James in terms of what "dpi" to use for his image. I didn't read your post carefully enough to realize you were only talking about processing photos - wasn't on my mind as he was displaying line art.

If I was scanning the artwork James used, I may have sampled it in color as I find it can be easier to do clean up on color images, but at some point I would have converted it to one-bit before sending it to the laser. I would not have scanned the text as grey scale and then try to let the laser process it into b/w - although it could.

Another question - you say a laser always cuts by making dots. It was my impression that the LaserPro machines automatically change mode to lase continuously at slow speeds for cutting. Do you agree that this is happening? I don't know how to prove or disprove it, but that is the way I read it.

[Martin Boekers]

Actually he was using a photo, he scanned an image making it a bitmap or raster image, that is unless he converted it to vector format.

Not sure about Laser Pro, but Epilog you specify frequency for cutting, I believe it is PPI (pulses per inch). I have "perforated" card stock and can see the spacing so it is not countinuos mode on Epilogs.

I have also seen that with Epilog that it does read grey scale as the denser the grey area the deeper it burns, if it was actually converted to a 1 bit image would that be true?

I may be wrong, so help me understand this a bit(:D) better.


Marty

[Niklas Bjornestal]

It was my impression that the LaserPro machines automatically change mode to lase continuously at slow speeds for cutting. Do you agree that this is happening? I don't know how to prove or disprove it, but that is the way I read it.
At least laserPro mercury II doesnt change mode at slow speeds, you can specify in when printing if it should be continuously or a specific ppi.

[Rodne Gold]

None of the lasers we are talking about *really* fire "continuously" , they all cut by treppaning (IE drilling a series of holes) They mostly pulse to get a peak power
In raster mode engraving , when working with vector files , the laser does work in a 1 bit mode , ie it either fires or doesnt
Any raster (scanning back and forth) engraving of a vector HAS to have the lasers driver convert it to a sort of 1 bit images where one "pixel" will be mapped to one dot

Increasing DPI when engraving is one strategy that compliments both power and speed. DPI , power and speed are sort of interchangeable but are also not.
Too little power will not "spark" the vaporisation process , leading to bad engraving , too much will create large heat affected zones , fast speeds lead to innacuracies and too slow speeds lead to too much heat into the material. High DPI means the laser is overlapping too much , leading to engraving on top of engraving and 1/2 the time you are engraving residue leading to ridges , bad surfaces , bowing and warping etc, too low DPI gives a shallow and "coarser" result.

By far the best way to engrave raster images is to use Photograv which converts them into 1 bit images and also takes into account the material , the materials dot gain , the lasers power and so forth.

In essence one has to balance whats acceptable output with whats acceptable thruput.

If you have adopted a time based model and need to engrave at slower speeds and higher DPI , essentially you are charging the customer for your machines deficiencies. Charging a customer less time and eating the extra thruput time is often not so much of an issue for a jobshop , but are sometimes and issue for production.
We had to do many 1000's of the same flexibrass plate and by juggling speed power and dpi , I managed to get good output from an initial base of 90 seconds down to round 27 seconds and the 1/2 hour of testing etc was WELL worth it in production.
You also have to bear in mind the materials you are working with , some like wood actually need "wrong" parameters to display well (burn not vaporise cleanly) and so you need to take into account or know what the material properties are and how they react to the laser.

[Richard Rumancik]

None of the lasers we are talking about *really* fire "continuously" , they all cut by treppaning (IE drilling a series of holes)

I wanted to clarify what the LaserPro is doing, but I didn't think I would be able to get any useful info from GCC. So I thought of checking the Synrad laser manual. They do say that the Synrad tubes have a CW mode to cause them to fire continuously, so I assume that this is what the LaserPro machines are doing. It makes sense to me, that if the objective is to get the maximum energy out of the tube, you want to fire 100% of the time.


At least laserPro mercury II doesnt change mode at slow speeds, you can specify in when printing if it should be continuously or a specific ppi. I'm glad to hear that they made it user-settable. It always frustrates me when they come up with an algorithm that does not allow the user to set it they way the user wants. In the original Mercury it changes to CW at 2.9 or 3% speed but to me that is quite presumptious of the programmer. CW mode is just a special mode of the ppi.


They mostly pulse to get a peak power What do you mean by that statement? I don't think you can get a higher energy pulse from a normal co2 laser by pulsing the laser.

[Rodne Gold]

You get far more energy at the peak of a pulse than you do with CW which gives you constant energy over time.
The most effective way of cutting is trepanning , ie cutting with drilled holes.
The way a laser works is to hit the object with such a high jolt that it vaporises , not to apply heat to burn or melt - thats why it pulses to get that zap.

[Richard Rumancik]

You get far more energy at the peak of a pulse than you do with CW which gives you constant energy over time.
The most effective way of cutting is trepanning , ie cutting with drilled holes.

That's what I thought you were getting at, but I have not found anything to support this concept. Perhaps you are thinking about YAG lasers, that use a q-switch to generate short-duration high energy pulses at a much higher level than what can be generated on a continuous basis.

There are some manufacturers that have developed lasers that do something similar for co2 lasers - but I don't think they are very common or used in laser engravers. (The GEM and Diamond Coherent products are q-switched. Are these used in engravers?) I can't find anything from Synrad that is. I get the impression that the the q-switched CO2 tubes are used in unique applications that can justify the extra expense - precision machining, medical etc. Some "slab" lasers can deliver higher energy pulses, but again, I don't think these are "engraver" tubes.


The way a laser works is to hit the object with such a high jolt that it vaporises, not to apply heat to burn or melt - thats why it pulses to get that zap.

That is the way acrylic cuts - by vaporization. But many materials won't vaporize and will only laser by melting (which requires air to blow away the molten liquid) or burning (aka thermal degradation). Cutting wood is closer to burning than to vaporization.

There may be good reasons for pulsing, but I don't think it can increase the actual energy of the pulse in the type lasers we use. To me maximum energy is delivered in CW mode (no pulsing), and that is the highest energy output mode. If you have any links that suggest otherwise, I'd be interested.

[Rodne Gold]

http://en.wikipedia.org/wiki/Laser
and
http://en.wikipedia.org/wiki/Laser_engraving

[Richard Rumancik]

Rodne: The first reference contains info about lasers in general. I was not sure which section you were referring me to, as a lot of it is not really relevent to co2 lasers.

However, the sections referring to CW, pulsed operation, and q-switching are consistent with what I said previously.

In the second reference on "laser engraving" I did not find anything on the topic. Not sure what info you were linking to that supports the comment.

[Martin Boekers]

Richard,

I found this in my research, I'm not a aser engineer so some of this is beyond me!

http://www.repairfaq.org/sam/laserco2.htm

The only reference I saw quickly to continous as opposed to pulsed was in the industrial end 500watts and up.

I don't believe that my Epilog can select continoius power for cutting.

I my experience with cutting wood, as I change the frequency (pulsing) to a lower number of pulses I do get quite a bit less charring than with a higher number pulses (closer to continious) so I would say for the appications that most of use use the lower wattage laser for it makes sense that the beam is pulsed.

Peck help us out here!:D

Marty

[Rodne Gold]

Gcc did use coherent tubes in their explorers , in fact I have 3 using em.

Below a certain speed percentage (3%), GCC actually used a different algorithm for cutting in terms of speed relative to what you set - its documented in the Explorer manuals showing relative speeds vs set speeds Enhanced mode also works right up to 100% with different speeds to unenhanced - heres the blurb from its woefully inadequate driver instructions


C-8 Enhanced Vector Mode
Check the enhanced vector mode to increase the cutting capacity.

NOTE:

If you are one of the users had the driver before V.3.43 and upgraded recently, please reset

your parameters when you select Enhanced Vector Mode. The following chart could be your
reference for the relationship of speed with Enhanced Vector and without Enhanced Vector.
a. The speed from 0.1% to 2.9% under enhanced vector mode is linear allocation, such as 0.2%
is 2 times faster than 0.1%.
b. The speed from 3% to 100% under enhanced vector mode is another linear allocation. The
100% is equal to 40ips.
c. The 3% under enhanced vector mode is a turning point of the speed curve.
Speed with Enhanced
Vector Mode
/Speed without Enhanced
Vector Mode
0.10% N/A
0.20% 0.10%
0.30% N/A
0.40% N/A
0.50% N/A
0.60% 0.20%
0.70% N/A
0.80% N/A
0.90% N/A
1.00% 0.30%
1.10% N/A
1.20% N/A
1.30% N/A
1.40% 0.40%
1.50% N/A
1.60% N/A
1.70% N/A
1.80% 0.50%
1.90% N/A
2.00% N/A
2.10% N/A
2.20% 0.60%
2.30% N/A
2.40% N/A
2.50% N/A
2.60% 0.70%
2.70% N/A
2.80% N/A
2.90% N/A
3.00% 0.80%
5.00% 2.50%
10.00% 5.00%
20.00% 10.00%
30.00% 15.00%
40.00% 20.00%
50.00% 25.00%
60.00% 30.00%
70.00% 35.00%
80.00% 40.00%
90.00% 45.00%
100.00% 50.00%

As to what is "continuous" mode cutting , this is what you do , however I am of the opinion that that laser still pulses here and cuts by treppaning, but at its max pulse rate which is so fast (1500 ppi+) that it is effectively "continuous" , it may work in a CW mode in terms of that the beam cannot modulate its pulses with actual speed - IE wont actually take into account slower speeds round corners and curves (ramp mode disabled) and this wont work for some cutting as it will generate lots of extra heat in the material
The "continuous" mode (PPI=X) is set independant of speed , so it appears that this mode is not controlled by speed settings. I think the Enhanced vector mode just allows more fine tuning over speed and allows a much slower speed to be set


PPI (Pen Page)
[DEFAULT SETTING: 400]
PPI (pulses-per-inch) represents the number of times the laser pulses (fires) per linear inch, exclusive for vector cutting. Higher PPI settings will generate deeper, overlapping laser pulses, resulting in cleaner cuts.
Lower PPI settings (lower than 150) will result in the individual laser pulses being spread apart, resulting in a perforated effect (similar to the perforation in the paper between mailing stamps).


If you drag the PPI slider to the furthest right (maximum), the value will change to X. This completely disables the PPI control and continuously fires the laser non-stop, without pulsing. Think of setting PPI to X as being equivalent to turning a water facet on with the water continuously flowing out. This also
disables the power ramp functionality, which automatically controls the PPI depending on the speed of the laser carriage (such as vector cutting around the corner of a square).
TIP
For Vector Engraving jobs, we recommend a PPI setting > 150
For Vector Cutting jobs, we recommend a PPI setting of > 400.


If you dont use PPI or freq when cutting , you run the risk of putting way too much heat into the material over time and thus damaging the piece or getting an edge that is not good.


See also http://en.wikipedia.org/wiki/Laser_ablation in terms of how our lasers "remove" material
and
http://en.wikipedia.org/wiki/Laser_cutting in terms of how they cut.