Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Printer "resolution"  (Read 3716 times)
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2808



WWW
« on: February 07, 2008, 09:00:10 PM »
ReplyReply

A few have asked about how the Epson 11880 compares in speed to the Canon ipf6100.  Being the curious and sometimes bored individual that I am, I decided to see what I could come up with.  I'm sure many don't care, and most feel it really isn't the most important thing to be evaluating (that's how I feel), but I can understand why it is important to some, and as I said ... color me curious on this one.

So I have completed time tests, and have sq. ft/ hour numbers for both printers at various resolutions. Before I post those numbers, there is some information that I feel needs to accompany them to put things in perspective ... that has to do with resolution and the actual number of dots the printer is laying down.

This is an apples to oranges problem - the two printers are quite a bit different in dot density.  To be fair, I believe it is important to know exactly what each printer is doing.  As an example, it is true that at MAXIMUM resolution, the Canon ipf6100 prints faster than the Epson 11880.  However, the Canon is printing at 2400x1200 dpi, or 2,880,000 dots per inch of material versus the Epson at 2880x1440, with is 4,417,200 dots per inch ... substantially more.  

Unfortunately after a couple of hours of searching (I'm not very good with Google), I cannot find information as to what the actual dot densities are for the other settings on either printer, as well as some of the other settings of the Epson driver itself.  Since they can affect speed, I think it is relevant to understand when comparing printer speeds.  So if anyone can point me to some information regarding Canon's dot resolution for its High and Standard quality settings at 600dpi, as well as Epson's dot resolution at 1440 and 720 dpi it would be most helpful.  Regarding the Epson, I'm also curious whether super Microweave and Finest Detail affect the number of dots being printed.

If my guesses are correct on those dot densities, it appears that both printers lay down dots at nearly identical speeds, so it really is a matter of which quality level do you want vs. the speed you need.
Logged

Farmer
Sr. Member
****
Offline Offline

Posts: 1622


WWW
« Reply #1 on: February 07, 2008, 11:18:48 PM »
ReplyReply

Finest detail on the Epson changes the expected input data level - ie ppi from the image from 360 to 720 and affects how the halftoning etc then works.  It doesn't affect the output resolution, but it will require more processing time to raster as it's no longer throwing out unwanted data (assuming you're supply that much ppi).

Microweave will change the speed at which the print head moves and the amount of dots fired as it increases the horizontal resolution in something akin to interpolation that scanners do to achieve a higher resolution than the physical scanning head (except with a printer it's not interpolation, it's just making smaller steps and potentially more dots (I say potentially, because the Epson has variable droplet sizes and depending on the resolution you're printing at will select a set of dot sizes from those available to use - so it may end up being less dots but larger based on the algorithm).

As for the matrix at lower resolutions, the matrix make up is visible in the driver settings for each paper type.  So on luster you have 2880x1440 super microweave, then 1440x720 super microweave, then 1440x720 microweave on, 720x720 super microweave, 720x720 microweave on.

Other papers, matte, drop down to including 360 at the lowest.

What the difference is precisely between the microweave I couldn't say, but again it will be variable because of the variable dot sizes.

Bear in mind that different paper sizes also affect print speed as they have to allow for different drying times (this can be set in custom paper settings).
Logged

jpgentry
Full Member
***
Offline Offline

Posts: 197


« Reply #2 on: February 07, 2008, 11:30:38 PM »
ReplyReply

Wayne!!! You always come through for us (me.)  Thanks alot!  I'm really interested to hear the time differences and more specifics.  

For those who make money printing other peoples images speed affects alot of things like what we charge customers and how much time is spent waiting for the printer.  In my case I print part time and in the evenings and do IT work during the day.  The images are large and not ganged together like smaller photo images can be.  Speed/quality/price breakdowns are major concerns and why I went with the ipf9100.  If the Epson is faster I may have paid more and chosen differently.  

I didn't think it was an accurate statement that the Epson was faster when I read the review of the 11880 here.  

Thanks for plunking through these comparisons.  Many will be more interested when the technology of the Epson filters down the 17, 24 and 44 inch printers.

Did Farmer fully answer your question or anything else you need to finish the testing?
« Last Edit: February 07, 2008, 11:48:30 PM by jpgentry » Logged
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2808



WWW
« Reply #3 on: February 08, 2008, 12:19:30 AM »
ReplyReply

In the Mac Driver, it gives me one resolution number.  2880, 1440, 720, and 360.  It doesn't state what the other resolution is.

It seemed logical to me that if Superphoto was 2880x1440, then superfine which is labeled 1440 would be 1440x1440, not 1440x720.  Still not completely certain exactly how large each matrix is at each setting.  Are your numbers coming from a PC driver?

As far as drying times, would the main indicator of this being the head hesitating at each pass longer?  Doesn't seem like they would actually slow the head down.

I'm still looking for answers on the Canon side right now ... the exact matrix of various settings there is also a little confusing.

UPDATE:  I seem to have the epson figured out ... tucked away in an 11880 brochure it lists the available matrixs as 2880x1440, 1440x720, 720x720, 720x360, 360x360, 360x180.  So what seems illogical is true.  Not sure why there isn't a 1440x1440 option.  Also, the Mac can only access 5 of the 6, no matter which paper choices, so I assume the mac driver (through photoshop anyway) doesn't print at 360x180.

Also a little in depth exploring solved the super Microweave and Finest detail settings.  Super Microweave is technology to reduce banding.  Finest Detail is used to increase edge sharpness on vector based data such as text, graphics, and line art. It does not affect photographs and also is not recommended for large files.  So it doesn't sound like either affects the total number of dots the printer lays down.

Now just to figure out what the Canon is doing
« Last Edit: February 08, 2008, 01:09:14 AM by Wayne Fox » Logged

Farmer
Sr. Member
****
Offline Offline

Posts: 1622


WWW
« Reply #4 on: February 08, 2008, 07:50:09 AM »
ReplyReply

Yup, the figures are from the Windows driver.

You probably can't access the 360x180 without a RIP - there are certainly no existing paper settings that appear to use it.  I'm not sure if there was an option to use that mode when create custom settings in the Onyx RIP or not.  Can check on Monday if I remember (a weekend can be a long time :-)

The head can move at different speeds, but I agree it's unlikely to vary beyond "printing" and "not printing" so yes, the time taken to make a return pass but it wouldn't be easily monitored by eye.

The reduction in banding from the super microweave is due to the finer movement of the head horizontally (and the varied dot patterns that result).  As with the finest detail, it's not directly changing the number of dots, but how and where and what size.

Of course, comparing just the "highest resolution" isn't necessarily fair.  For example, the Epson, at 1440x720 might produce results equal to the Canon at its highest output levels, or vice versa.  In that case, one or the other might be faster at a given output quality. even though it seemingly is using a lower resolution.  It's quite subjective.
Logged

jpgentry
Full Member
***
Offline Offline

Posts: 197


« Reply #5 on: February 08, 2008, 11:19:19 AM »
ReplyReply

Wayne

I agree that we should figure out the resolution and comparisons between the two printers before making apples to apples comparison but here is a suggestion.  You could print at the three or so major settings from both printers (both in bidirectional) and then you can scan a 1x1 inch portion of the prints at 600 dpi on a flatbed scanner.  Then I can make a little webpage for you that will show the quality of the different settings and the time it took to print them.  This way the quality/speed differences would be aparent for all to see.  I could also do the scanning if you need.

Heck we could have someone do that with the Z also, except the thing that would make this so cool would be that the same profiling package and paper would be used on all three.  For example premium luster and then one matte paper.  We could throw up the gamut plots for the three at highest media setting for comparison.  If you would like to work on this let me know.  If it's too much I understand and am appreciative of what you've done.

-Jonathan


Quote
Yup, the figures are from the Windows driver.

You probably can't access the 360x180 without a RIP - there are certainly no existing paper settings that appear to use it. I'm not sure if there was an option to use that mode when create custom settings in the Onyx RIP or not. Can check on Monday if I remember (a weekend can be a long time :-)

The head can move at different speeds, but I agree it's unlikely to vary beyond "printing" and "not printing" so yes, the time taken to make a return pass but it wouldn't be easily monitored by eye.

The reduction in banding from the super microweave is due to the finer movement of the head horizontally (and the varied dot patterns that result). As with the finest detail, it's not directly changing the number of dots, but how and where and what size.

Of course, comparing just the "highest resolution" isn't necessarily fair. For example, the Epson, at 1440x720 might produce results equal to the Canon at its highest output levels, or vice versa. In that case, one or the other might be faster at a given output quality. even though it seemingly is using a lower resolution. It's quite subjective.
[a href=\"index.php?act=findpost&pid=173269\"][{POST_SNAPBACK}][/a]
« Last Edit: February 08, 2008, 11:20:50 AM by jpgentry » Logged
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2808



WWW
« Reply #6 on: February 08, 2008, 11:21:29 AM »
ReplyReply

Quote
Of course, comparing just the "highest resolution" isn't necessarily fair. For example, the Epson, at 1440x720 might produce results equal to the Canon at its highest output levels, or vice versa. In that case, one or the other might be faster at a given output quality. even though it seemingly is using a lower resolution. It's quite subjective.
[a href=\"index.php?act=findpost&pid=173269\"][{POST_SNAPBACK}][/a]

I agree with that, and  believe Canon is doing that to their advantage.  The Canon may appear faster just because its speed at maximum resolution is quite a bit faster than the Epson's is at maximum resolution ... even though the Epson maximum resolution is substantially higher.  So to state that their printer is "60% faster" based on highest quality settings of both printers is misleading.

Thats why this is really an apples to oranges discussion.  Once you elect not to use the highest resolution, as you mentioned, it becomes pretty subjective in determining which resolution on what is comparable image quality to which resolution on the other.  In fact, even using the highest settings it is subjective ... is the Epson better than the Canon with it's higher resolution?

Now if I can just find out exactly what the canon is doing with it's multiple "passes" and what dot matrix it is using for various qualities.
« Last Edit: February 08, 2008, 11:30:55 AM by Wayne Fox » Logged

jpgentry
Full Member
***
Offline Offline

Posts: 197


« Reply #7 on: February 09, 2008, 02:51:06 PM »
ReplyReply

I would think that "passes" is just some Canon speak for how many times the head moves back and forth to complete one inch of vertical paper movement through the printer.  The dpi is the dpi despite the number of passes.  When I crank the Canon up to it's highest mode it starts making very few passes per the number of inches that are spitting out of the machine.  When I put it into it's highest quality mode there are many more passes per inch.  8 passes sounds about right per inch.  I'll have to count as the paper comes out.

I think dpi is what you're after and passes aren't relevent.

Regarding it being subjective, yes I think you are right.  That's why I was suggesting scanning in high resolution so a subjective comparison can be made combined with the factual knowledge of the resolution and the time it took to print.  At some point good enough is good enough and that point for me is when I can't tell the difference (dots be darned.)  When that point is reached now I just want to know which printer will deliver the work in less time.  For example I print my canvas at the fastest Canon setting because ther was very little difference in print quality between the fastest and slowest due to the texture of canvas.  On the other hand I would never use anything but the highest setting for photo papers.

One last point... dpi is not the only measure of quality of an image and I'm not sure that it needs to be a counter-balance to speed.  I am really interested in your timings and the different quality levels and what you found out.

-Jonathan

Quote
I agree with that, and believe Canon is doing that to their advantage. The Canon may appear faster just because its speed at maximum resolution is quite a bit faster than the Epson's is at maximum resolution ... even though the Epson maximum resolution is substantially higher. So to state that their printer is "60% faster" based on highest quality settings of both printers is misleading.

Thats why this is really an apples to oranges discussion. Once you elect not to use the highest resolution, as you mentioned, it becomes pretty subjective in determining which resolution on what is comparable image quality to which resolution on the other. In fact, even using the highest settings it is subjective ... is the Epson better than the Canon with it's higher resolution?

Now if I can just find out exactly what the canon is doing with it's multiple "passes" and what dot matrix it is using for various qualities.
[a href=\"index.php?act=findpost&pid=173335\"][{POST_SNAPBACK}][/a]
« Last Edit: February 09, 2008, 03:01:19 PM by jpgentry » Logged
jcote
Newbie
*
Offline Offline

Posts: 48



WWW
« Reply #8 on: February 09, 2008, 05:00:01 PM »
ReplyReply

Quote
This is an apples to oranges problem - the two printers are quite a bit different in dot density.......
[a href=\"index.php?act=findpost&pid=173180\"][{POST_SNAPBACK}][/a]

Wayne,

I love this topic and thank you for your hard work. As an interesting side note to your "Resolution/Dot Density" comments, there is another angle to consider when talking about the Resolution of an InkJet printer than the standard Xdots by Xdots matrix way of determining the resolution of a plot.

Between the Canon and the Epson you want to compare, what I am about to say makes very little difference but I think it is worth pointing because it is relevant to comparisons with other printers.

The other resolution angle you have to think about is the minimum droplet size that a print head can spray. Both the new Canon and the new Epson use variable droplet size heads. The minimum size of the droplet from the Canon Head is 4 picoliters and the new Epson heads produce a minimum droplet size of 3.5 picoliters.

I don't want to get into a discussion about whether .5 picoliters is enough to make much difference but you can see why this might matter. In the past, various printer manufacturers have claimed resolutions of 2880 x 1440 with a minimum droplet size of 8, or even 12 picoliters. While a print head like this may actually be spraying a grid of 2880 x 1440 those big droplets have a lot of overlap at various points in the plot, depending on the density of the print. The plot may be 2880 x 1440 but the resolution, and ultimately, the fine detail is better if the head is able to spray a smaller droplet.

It is really almost impossible to talk about resolution on an inkjet printer using the same terms used for conventional halftone printing or even con-tone pixels. It is even more difficult to use the old grid type terminology with variable droplet head printers. The actual resolution changes depending on how the RIP/Head decides to spray at any given part of your image.
Logged

Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2808



WWW
« Reply #9 on: February 09, 2008, 10:15:40 PM »
ReplyReply

Quote
I would think that "passes" is just some Canon speak for how many times the head moves back and forth to complete one inch of vertical paper movement through the printer.  The dpi is the dpi despite the number of passes.  When I crank the Canon up to it's highest mode it starts making very few passes per the number of inches that are spitting out of the machine.  When I put it into it's highest quality mode there are many more passes per inch.  8 passes sounds about right per inch.  I'll have to count as the paper comes out.


-Jonathan
[a href=\"index.php?act=findpost&pid=173572\"][{POST_SNAPBACK}][/a]

The Canon is a little confusing to me, because it seems to be using 2 different "dpi's".  One seems to be based on the algorithm used as it rasterizes the image ... it seems to do that in blocks ... 600x600, 300x300, and perhaps even lower.

However, that doesn't seem to have much to do with the actual dot matrix it is using.  I'm guessing the number of passes is relevant to those dots ... after all, each pass it is laying down dots.  There seems to be 3 settings here, 16,  12, and 8 pass.  It seems safe to assume that in 16 pass mode is laying more dots down in a denser pattern and thats where you get to the maximum dot matrix the machine is capable of, 2400x1200.  If the printer is laying down the same amount of dots in each pass, that would indicate at 8 passes it is laying down 1200x1200 dots, and at 12 passes it is laying down 1800x1200 dots.  I can't find anywhere to confirm this.

However, I've been working on this for a few days and am concluding as I go on it may be somewhat irrelevant.  Each printer is capable of laying down dots in different ways, and it is quite possible that one printer is capable of higher quality output using far less dots ... the screening algorithms and variable dot sizes of each printer can impact this significantly.

At this point I am printing test targets using different quality settings to see what if any differences become apparent. Personally I don't believe examining high res scans of the images is relevant ... I don't think there is a way to turn this into an objective test.

 I have a few friends with pretty good eyes for this stuff that I have asked to examine the results.  There are two parts to this ... comparing different quality settings of each printer against itself and the other printer.  One caveat to this, I am only printing photographic output, no vector data.  The printers may perform differently with vector type images.

I have the tests done on the Epson, but won't be able to print them on the Canon till next week (it's at my office).  By Wednesday I should have a brief write up and thoughts.
Logged

Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2804


« Reply #10 on: February 10, 2008, 06:31:04 AM »
ReplyReply

Quote
The Canon is a little confusing to me, because it seems to be using 2 different "dpi's".  One seems to be based on the algorithm used as it rasterizes the image ... it seems to do that in blocks ... 600x600, 300x300, and perhaps even lower.

However, that doesn't seem to have much to do with the actual dot matrix it is using.  I'm guessing the number of passes is relevant to those dots ... after all, each pass it is laying down dots.  There seems to be 3 settings here, 16,  12, and 8 pass.  It seems safe to assume that in 16 pass mode is laying more dots down in a denser pattern and thats where you get to the maximum dot matrix the machine is capable of, 2400x1200.  If the printer is laying down the same amount of dots in each pass, that would indicate at 8 passes it is laying down 1200x1200 dots, and at 12 passes it is laying down 1800x1200 dots.  I can't find anywhere to confirm this.

However, I've been working on this for a few days and am concluding as I go on it may be somewhat irrelevant.  Each printer is capable of laying down dots in different ways, and it is quite possible that one printer is capable of higher quality output using far less dots ... the screening algorithms and variable dot sizes of each printer can impact this significantly.

At this point I am printing test targets using different quality settings to see what if any differences become apparent. Personally I don't believe examining high res scans of the images is relevant ... I don't think there is a way to turn this into an objective test.

 I have a few friends with pretty good eyes for this stuff that I have asked to examine the results.  There are two parts to this ... comparing different quality settings of each printer against itself and the other printer.  One caveat to this, I am only printing photographic output, no vector data.  The printers may perform differently with vector type images.

I have the tests done on the Epson, but won't be able to print them on the Canon till next week (it's at my office).  By Wednesday I should have a brief write up and thoughts.
[{POST_SNAPBACK}][/a]

Speed, image quality and resolution have always been flexible terms in printer manufacturer's advertising and documents. Image quality on a given paper is the result of a lot of parameters and at the same time speed can not be qualified separate from image quality. So you need a good target to measure image quality and a compatible paper for your printer, that could mean two different papers on two machines. If you want to compare two printers against one another that have different native resolutions you need two targets each with the right resolution to suit the printer's native resolution.
Print the samples at different output resolutions, bidirectional and unidirectional and you should get prints that can be equal in image  quality. Only after you get that comparable image quality you can decide which one is the fastest on printing speed alone. After that you can take one image file and using the settings that deliver the equal print results you process that image through two  drivers and clock the process before printing starts, that's another time factor. The results may differ again however as the two resampling routines in the drivers may not be a match for one another. That couldn't be verified in the first test that delivers the targets at native resolution to the driver. To avoid the resampling routines of the drivers you could use Qimage that will do the resampling to the different native resolutions with the same algorithms and should deliver almost the same quality at 300/360 or 600/720 or 1200 PPI input resolution to suit the different native resolutions of printers and driver settings.

If there's a desire to go through all that testing then this site has the targets and the advice to do the job:

[a href=\"http://www.ddisoftware.com/qimage/quality/]http://www.ddisoftware.com/qimage/quality/[/url]

Although based on one vector design the targets are raster images and are a more objective tool than photo images.

For the Canon's: the iPF9100 got an extra higher number pass mode compared to the iPF9000 if I remember it correctly. At the Photokina 2006 the general consensus was that the image quality of the 3: C, HP, E, was equal but the Canon had a slightly coarser dot (iPF9000, Z3100, 3800). The two things mentioned here can be related and ratio speed/IQ changed. I do not have all the printers at hand and can not make objective measurements. What I do know is that Canon's docs contain very aggressive texts on the qualities of other manufacturer's printers at the edge of deceiving potential customers. Something I have not seen in this industry for some time. That still doesn't say a Canon can not be faster.

Like with pixel peeping it is not always relevant either to test all this. A shop or an individual is satisfied with a quality as a result of his workflow and what customers expect and either accepts the time it takes to produce or not. Extra speed is usually the result of a new generation of heads, all 3 competitors are more or less in that phase now. Other than textural aspects like gloss there's in my opinion not that much to gain in image quality in next generations.


Ernst Dinkla

try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/
Logged
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad