Luminous Landscape Forum
Raw & Post Processing, Printing => Printing: Printers, Papers and Inks => Topic started by: ares on December 01, 2007, 03:10:19 pm
-
Hello,
someone may remember that a few weeks ago I asked if someone else was having softness problems when printing from a 7800 on custom page sizes from OSX. Well, after testing 3 different 7800 and 2 9800, and various combinations of PowerPC, Intel, OSX 10.4.x , 10.5.x, Photoshop CS, CS2, CS3, and all the available Epson drivers (including the latest for Leopard), I've concluded that the issue is real and perfectly repeatable. See the attached image. I've also tested Lightroom, but apart from the fact that I'm not able to obtain a good crisp printout event with sharpening set to high, the issue is present too.
Epson is not interested in solving the issue, their only advice so far has been to print from Windows (wich is bug free, prints appear exactly the same on both default and custom paper sizes) or buy a RIP.
Now I'm wondering why nobody else seems to have noticed this problem (or at least I can't find any mention of it on the Internet), and I'm also wondering if the newer x880 printers have this issue too. I was thinking about getting an 11880, but being forced to run Parallels to print from an old version of Photoshop Windows wich I have is not acceptable. Why spend so much money and then get so miserable results, at least printing from OSX? Is printing from OSX to Epson machines definitely out of the question?
Besides, does anyone know if the salmon color bug has finally been fixed? I've noticed that with the latest Leopard driver the default color profile in ColorSync for the x800 printers is no longer Standard but PLPP250, and this makes the trick of changing the Standard profile to Generic RGB impossible.
[attachment=4095:attachment]
-
Hello,
Epson is not interested in solving the issue, their only advice so far has been to print from Windows (wich is bug free, prints appear exactly the same on both default and custom paper sizes) or buy a RIP.
Now I'm wondering why nobody else seems to have noticed this problem (or at least I can't find any mention of it on the Internet),
[a href=\"index.php?act=findpost&pid=157526\"][{POST_SNAPBACK}][/a]
If nobody else noticed the problem, maybe the problem is with something you are doing and not the machine or the operating system. Very many people use Epson printers with Mac operating systems and produce excellent prints. I'm a Windows user, but I know from what I've seen that fine results from Epson professional printers on the Mac OS are certainly attainable. If you're using Epson's current high-end machines and you are in the game for a unit like an 11880 Epson is normally responsive about these things. Are you dealing with the ProGraphics Support Group?
-
This is very odd.
I've been using a 9800 for over 2 years now , with some very critical clients. Every print is on custom paper size , all printed out of OSX. No-one has ever had a problem with sharpness. Maybe no-one was looking for a problem, or doing the comparative tests you describe. I would be very interested if other high end users could come up with similar conclusions. If I can squeeze the time in, I'll do some basis comparisons and see what I come up with. I can't imagine what would cause such a result.
Brian
www.pharoseditions.com.au
-
If nobody else noticed the problem, maybe the problem is with something you are doing and not the machine or the operating system.
I would like this to be the case, but I'm afraid it's not. Or at least it's not so simple. First a little background: I had a 7800 and a 9800, then about two months ago I sold the 9800 to a friend of mine because I need a larger printer (the 11880). A client of mine sent me a very high quality scan of a painter artwork (true 600 dpi, 800 MB TIF) to be printed on canvas. While making test prints I noticed that sometime the printout was softer. It took me many hours (and much ink and paper) to pinpoint the cause to the use of a default paper size vs. a custom one. Then I took Bill Atkinson's test image and used it as a reference, printing on many different papers, including Epson PGPP250. Same result. I downloaded and installed the Windows driver under the copy of Windows XP wich I run under Parallels on my iMac, and this time the results are the ones expected, i.e. no difference whatsoever between default and custom paper size. Note that I used exactly the same image with exactly the same driver settings.
So I went to my friend's studio to repeat the tests with my old 9800, and the results were absolutely consistent with mine (no difference in Windows, softer on custom paper size when printing from OSX).
After that I went to a certain number of print shops with an Epson x800 machine and OSX, and asked them to print my target image two times, one on a default paper size and the other on a custom one. Again, same result, softer on custom. Note that at this point the test had been made on various combinations of OSX (10.4.9,10.4.10,10.5.0,10.5.1), processor types (PowerPC and Intel) and Epson drivers.
I'm in Europe, but just to be sure I downloaded firmware and drivers from the Epson USA site and tested them with my own 7800, but the results were, again, the same.
Meanwhile, I've contacted Epson, and their response was vague, with something like "maybe you are making mistakes in Photoshop", "maybe you have clogged nozzles", "print from Windows or buy a RIP".
Yesterday I made another discovery: the degree of softness depends on the size of the margins you choose when creating a custom paper size. If you click on "
Manage custom sizes...", then "+", you end up with an Untitled size with margins fields already filled with 0,63cm for top, left and right, and 1,45cm for bottom. If you use these values, changing only the size of the paper, the printout comes out sharper. Not as sharp as a print made with a default paper size, but in the middle between default paper size and custom paper size with any other values for margins (Photoshop CS3 10.0.1, OSX 10.5.1, Epson driver 6.11 downloaded from UK site).
At this point I don't know what else I can do.
-
interested if other high end users could come up with similar conclusions. If I can squeeze the time in, I'll do some basis comparisons and see what I come up with. I can't imagine what would cause such a result.
I'd be extremely happy if you can repeat on your system the test I've already done. If you prefer I can send you the image I'm using.
-
Ares,
OK - the amount of testing and research you've done is impressive and far more extensive than I understood from your previous posts. One clue PERHAPS differentiating your comment from Brian Gilkes' is that you mention you are using driver version 6.11. I wonder whether Brian is using the previous driver version. IF SO, that would point in the direction of an issue with the Epson driver, because the control of paper sizes and the ink laydown and dithering done by the printhead are all intimately related to the driver. So it could be helpful if you and Brian compared driver versions.
I've not had this experience - don't use a Mac so I'm not the one who can say "been there, seen that and here's what to do" - that would be nice, however I'm confined to helping with pointers about where to look - but you clearly know that already.
What surprises me more than the technical problem is the kind of lacklustre customer support response you appear to be getting from Epson - not only because of what they should be doing for you as a high-end customer, but also because just in case your experience is the tip of an underlying iceberg, they need to know about it, take it seriously and fix it for everyones' benefit. One presumes that drivers are tested for custom paper sizes, but one never knows. It's not impossible that there are gremlins in the nooks and crannies waiting to be discovered and dealt with!
-
understood from your previous posts. One clue PERHAPS differentiating your comment from Brian Gilkes' is that you mention you are using driver version 6.11. I wonder whether Brian is using the previous driver version.
I've tested, on my own 7800, driver version 3.6d (from Epson UK) and 3.5aE (from Epson USA) under OSX 10.4.10 and 10.4.11 on a PowerMac G5 dual core and on a new iMac 24" Core2 Duo. I've also tested both these drivers under OSX 10.5.0 (Leopard), because they work, albeit without the new enhanced user interface, both on the PPC and the Intel Mac. I've also installed different printer firmwares from both Epson UK and Epson USA sites.
Then I've tested the new Leopard drivers, version 6.11 from Epson UK and the beta version from Epson USA (they are both 6.11 but are slightly different) on the iMac and then on the PPC. I've tested both Leopard 10.5.0 and 10.5.1.
I've also installed and old version of Photoshop CS to be sure it is not a problem with Photoshop CS3.
In all these tests, the printout obtained using default paper size (I've used many sizes, both sheet and roll) is absolutely the same, both for sharpness and color. Instead, the results obtained with custom paper sizes vary slightly in sharpness (now I believe it is due to different margins), but they are always inferior to the ones printed using default paper sizes. The color is the same.
Printing from Windows is bug free, i.e. I obtain exactly the same results on both default and custom sizes.
BTW, I've tested 1440, 1440 HS, 2880 and 2880 HS, with both photo black and matte black inks, and while the resulting sharpness is obviously different, the relationship between default and custom paper sizes is always the same (the former is sharper, the latter is softer).
As for Epson support, I don't know what to think.
-
Wow - you must have devoted a huge amount of time and effort to this. The sheer size of the test matrix is becoming hard to keep straight in one's head. This is intriguing (for me - for you I can see it being a royal pain). Now let me see if I understand your results correctly:
On Windows there are no problems period. Whether you use old drivers, new drivers, custom sizes or standard sizes the sharpness is the same and satisfactory.
On Mac, whether you use old drivers or new drivers, UK drivers or US drivers, OSX4.xx or Leopard, sharpness is fine as long as you stick with standard sizes. However, under all these conditions, as long as you make a custom size, sharpness diminshes.
Have a I got it right?
One last question - I suppose you do this, but just to be sure. When you create a custom size, do you save it as a preset and then recall this preset before trying to use the custom dimensions? I have found this necessary in the context of the whole print-centering business we went through with CS3, and I'm wondering whether this could be a factor contributing to your problem. If you are saving the custom size as a Preset and then recalling it for use, then if no-one else with more experience reading this thread has better ideas, I think this is an issue which you really need to escalate to the Epson ProGraphics people in the USA, regardless of where you live.
-
.
-
Wow - you must have devoted a huge amount of time and effort to this. The sheer size of the test matrix is becoming hard to keep straight in one's head. This is intriguing (for me - for you I can see it being a royal pain). Now let me see if I understand your results correctly:
I've printed almost 50 tests.
On Windows there are no problems period. Whether you use old drivers, new drivers, custom sizes or standard sizes the sharpness is the same and satisfactory.
Correct.
On Mac, whether you use old drivers or new drivers, UK drivers or US drivers, OSX4.xx or Leopard, sharpness is fine as long as you stick with standard sizes. However, under all these conditions, as long as you make a custom size, sharpness diminshes.
Correct.
One last question - I suppose you do this, but just to be sure. When you create a custom size, do you save it as a preset and then recall this preset before trying to use the custom dimensions?
With Mac drivers there is only one way to create a custom paper size: you select "Manage custom sizes..." at the bottom of the OSX "Paper Size" drop down list (in the "Page Setup" dialog), then a new window opens. You enter your values for paper size and margins, then you close the window (after naming the new paper size something else than Untitled, but this is not mandatory) and at this point your new paper size is saved. After that, its name will show at the end of the Paper Size list, and you can select it like you do with any of the default ones. You cannot use a custom paper size before saving it. It's all managed by OSX itself, but every printer driver fills the Paper Size list differently, according to the printer capabilities.
In fact, there should be no difference at all between a custom paper size and a default one. The only difference, at least for Epson large format printers drivers, is that there are multiple choices for every single paper size, one for sheet paper and the others for roll. For example, you have "A4", "A4 roll paper-banner" and "A4 sheet", and, depending on wich one you select, the driver automatically chooses the correct paper source for you (sheet or roll). Instead, when using custom paper sizes, you have to select the correct paper source by yourself, from a drop down list, in the "Basic" tab of the "Print Settings" section of "Print" dialog.
OSX printing dialogs and workflow are totally different from the Windows counterparts.
are saving the custom size as a Preset and then recalling it for use, then if no-one else with more experience reading this thread has better ideas, I think this is an issue which you really need to escalate to the Epson ProGraphics people in the USA, regardless of where you live.
I would like to hear from someone else with a x800 printer and OSX. I've done may tests on different machines, but maybe I'm still doing something wrong. But I'll try to contact Epson USA as well.
-
I will get some tests through as soon as I can. I have a very big run with a deadline at present, using Somerset White Book paper and utilising Auto Expand to print on deckles, which would throw some other variables into the equation.I will contact a couple of colleagues who are very critical with their 9800s to see what they think and report.
Your tests are comprehensive , but not qualitative . The next step, assuming similar qualitative results from elsewhere , would be to devise some more exacting tests involving measurements.
I'll think about that one.
Cheers
Brian
www.pharoseditions.com.au
-
Your tests are comprehensive , but not qualitative . The next step, assuming similar qualitative results from elsewhere , would be to devise some more exacting tests involving measurements.
Well, that could be interesting, but is it relevant? What I have found is a difference in output quality where there should be none. In fact printing from Windows there is absolutely no difference between an image printed on a standard paper size (e.g. A4) and the same image, printed with the same driver settings on the same paper, but using a custom paper size. No difference at all, you cannot tell one print from the other.
Instead, I've found that, doing exactly this with the OSX driver, you'll always end up with a difference in sharpness. Why? Is it my fault? I've tested more than one printer, but maybe all these printers are defective. I don't think so, because they work perfectly (apart form this problem), and they work perfectly at all when driven from Windows. This is why I've asked if someone else could try to replicate my results, simply printing the same image twice, and then comparing them. If they are different, it's Epson's problem, not ours. If not, I'll see to have Epson send someone to check every printer and every Mac I've tested to determine exactly where the problem is, because I don't know what else I could do.
-
I didn't mean to be picky, and was not criticising your methodology; it's just the old science degree rumbling in the reptilian brain.
I have emailed the associates that I mentioned, and subsequently spoken by 'phone to one. Like me he is currently printing a major project , but we will resolve this one as soon as possible. At this end, I need to get uncoated matte off the printer and get in something like Harman gloss. That will pick it up.
I would hope a few other LLers , who try to squeeze out the last drop, have a go too. If we all concur , Epson and/or Apple and or Adobe, will have some explaining to do.
Cheers,
Brian
Pharos Editions
-
I didn't mean to be picky, and was not criticising your methodology; it's just the old science degree rumbling in the reptilian brain.
Brian, obviously I appreciate very much your help, and I didn't mean to attack you. It's only that, since we have paid good money for our printers, Mac and software, I think that if the issue is confirmed it's not our duty to investigate it and find a solution.
Not only that: even if you or your associates are not affected, I surely am, I don't know what else I can do and I am not receiving a good support from Epson. On top of that, it's now clear that many prints I've made on custom paper sizes came out softer than they should've been, and you know perfectly well what this mean.
Just to be sure, I redid my usual print, then scanned it at 2400 dpi. On the left is the one on default paper size (A4) on the right is the one printed on a custom size. You can see that many details (look at the reflections in the eyes) are simply not there.
[attachment=4121:attachment]
-
Curious as to what paper sizes you used. I assume there was no chance the driver was having to resize one.
Curiousity got the best of me. I used the Lab Image from Bill atkinson's site, which includes the Photodisc image as one of about 15 images as well as some color gradients and solid color patches.
I "resized" without resampling to a 13x16 inch image, which yielded a 424 ppi value.
I printed one print on Lustre paper (didn't have glossy on hand) using the build in A3+ setting, and another creating a custom paper size of 13x19".
I really couldn't find any apparent differences using a 10x loupe. I then scanned the images with an Epson 4270 scanner at 1200 dpi. Attached are 3 very small segments from each scan. The file is saved at maximum quality to minimize jpeg artifacting, so it's pretty big ...
... 1.4 mb ... just fair warning if you have a slow connection.
Anyway ... curious if anyone sees that either the top or bottom row is "better". I can't. I can find areas on each row that appear sharper than the same area on the other row, but it goes both ways, and I can't see that one row is obviously sharper.
I'll try this on my 11880 tomorrow on glossy paper.
-
Curious as to what paper sizes you used. I assume there was no chance the driver was having to resize one.
No, I've always used paper sizes much bigger than the image.
I "resized" without resampling to a 13x16 inch image, which yielded a 424 ppi value.
But you should send a 360 ppi image to the printer, otherwise the printer driver will scale it.
I'll try this on my 11880 tomorrow on glossy paper.
I've attached a JPG version of the original image I'm using for my tests. It's already sharpened (0,5-160). Could you be so kind to try with this one?
[attachment=4123:attachment]
-
I printed a Kodak Color Evaluation Target on Harman Gloss today.
Power Mac G5 Quad, MacOS 10.4.9 and Epson Driver 3.8c for Ep11880.
Photoshop CS2 with my own made RGB profile.
First "default size A3" and then I made a customized A3.
Printed both. Beautiful as usual
And even with my best Pentax 8x30 loupe I can't see ANY difference at all.
(I've been printing inkjet since 1997)
/Sven
-
Well, it seems finally I have found a solution, or at least a workaround.
When creating a custom paper size, you have to set all four margins to 0. This of course makes no sense, because the printer cannot print borderless on all four sides, but if you do, the sharpness is the same as the one obtained when printing on default paper sizes.
I need some more testing, but so far the results look promising.
-
But you should send a 360 ppi image to the printer, otherwise the printer driver will scale it.
[a href=\"index.php?act=findpost&pid=158106\"][{POST_SNAPBACK}][/a]
True, but the scaling should be identical in either case so it shouldn't result in a difference.
As far as image, what you included was a section of the image I am using.
Regarding the border size, when creating a "border" with your custom setup, did you do it by selecting your printer which creates the correct default border for a printer/paper combination, or did you use your own numbers and create a "border".
I'll try this tonight on my 11880.
-
Regarding the border size, when creating a "border" with your custom setup, did you do it by selecting your printer which creates the correct default border for a printer/paper combination, or did you use your own numbers and create a "border".
I'm not sure I understand what you mean. I create a cutom paper size first by selecting "Manage Custom Sizes..." (picture 1), then clicking on "+". At this point, I obtain what is shown in picture 2.
If I leave the values for margins alone, I end up with a soft print. Instead, I've discovered today that if I set all four margins to 0, the resulting sharpness is good. But this should not be the case, because margins should not have anything to do with printing sharpness (unless, maybe, you use auto expand or something like this and the driver resizes your picture). Besides, using the Windows driver I can set whatever margins I like and the output is always as expected.
-
If I leave the values for margins alone, I end up with a soft print. Instead, I've discovered today that if I set all four margins to 0, the resulting sharpness is good. But this should not be the case, because margins should not have anything to do with printing sharpness (unless, maybe, you use auto expand or something like this and the driver resizes your picture). Besides, using the Windows driver I can set whatever margins I like and the output is always as expected.
I never would have thought of that one.
To solve a problem I had with my Epson 4000, I was advised by the Epson Techo to set margins at zero. That has been my pracice since out of habit and stll using the 4000 for some jobs.
That could explain why no one has noticed a sharpness problem. I'll test that too asap.
Nothing further to add at the moment.
Brian
Pharos Editions
-
That could explain why no one has noticed a sharpness problem. I'll test that too asap.
If you have always printed with zero margins then you should be ok. Lucky you. Anyway, I've made other prints, and while they seem identical to the naked eye, examining them under a 32x microscope I think I'm able to spot a very slight difference in sharpness, always in favor of prints made on a default paper size. But I'm not sure, the difference is so small that I probably would need a proper double blind test to draw a definitive conclusion.
-
At this point, I obtain what is shown in picture 2.
[a href=\"index.php?act=findpost&pid=158208\"][{POST_SNAPBACK}][/a]
The pop-up menu in that picture allows you to select a printer, and it will set the default margins to the maximum the printer can handle for that size of paper, based on sheet feed. Helpful especially on the bottom where some printers require some area for gripping the paper.
Since all of my custom setups are for roll paper, I usually leave them at 0.
-
For what it's worth, tonight I resized to 360 ppi without interpolation the test page that has been used, and printed 3 versions on my 11880 printer on 24" roll kodak Glossy paper at 2880 dpi.
Version one was using the standard 24x30 roll option.
Version two was with a custom paper size with all margins set to 0 (which is also the default for the 11880 printer)
Version 3 was a custom paper size with manual margins of .5 inches on sides and 1 inch top/bottom.
I marked the back, and cut the prints to the same size, cutting off the identifying text from the front. Mixed them up, and spent about 10 minutes with a 10x loupe - not knowing which one I was looking at ... just trying to see if one was slightly better or worse than the others.
While there are some subtle differences in dot patterns, none equate to a difference in sharpness.
-
If you have always printed with zero margins then you should be ok. Lucky you. Anyway, I've made other prints, and while they seem identical to the naked eye, examining them under a 32x microscope I think I'm able to spot a very slight difference in sharpness, always in favor of prints made on a default paper size. But I'm not sure, the difference is so small that I probably would need a proper double blind test to draw a definitive conclusion.
[a href=\"index.php?act=findpost&pid=158254\"][{POST_SNAPBACK}][/a]
But who is looking at a print under a microscope
/Sven
-
But who is looking at a print under a microscope
/Sven
[a href=\"index.php?act=findpost&pid=158354\"][{POST_SNAPBACK}][/a]
The photographer, of course!
-
If truly microscopic differences are all this thread is about, one wonders what's the fuss, but I thought the OP was about differences of sharpness you see with the naked without having to look too hard.
Wayne - really great you are taking the trouble to do those tests and with such a state of the art printer. The fact that you can't replicate the problem would seem to suggest that the issue is not generic but case-specific, and that would suggest "ares" really needs some high quality concentrated tech support from Epson to sleuth what's going in his set-up and workflow to cause this.
Wayne - also - not clear that the resampling to 360 is done exactly the same way (mathematically) in the Epson driver versus in Photoshop. There's been much inconclusive discussion about this in the past, for earlier versions of Photoshop and Epson drivers. It's an interesting question that perhaps should be revisited - again making a test or two for a SUBSTANTIAL resampling to 360 both ways, and seeing whether there is a visible (naked-eye) difference of image quality (sharpness, tonal gradations) between the two approaches. Time permitting I'm thinking of doing it, but please beat me to it.
-
Wayne - also - not clear that the resampling to 360 is done exactly the same way (mathematically) in the Epson driver versus in Photoshop. There's been much inconclusive discussion about this in the past, for earlier versions of Photoshop and Epson drivers. It's an interesting question that perhaps should be revisited - again making a test or two for a SUBSTANTIAL resampling to 360 both ways, and seeing whether there is a visible (naked-eye) difference of image quality (sharpness, tonal gradations) between the two approaches. Time permitting I'm thinking of doing it, but please beat me to it.
[a href=\"index.php?act=findpost&pid=158727\"][{POST_SNAPBACK}][/a]
In this case I don't think any resampling took place ... I resized the print without resampling to 360ppi, then printed it at the resulting size.
I think your talking about when printing to a specific size is it better to resample with photoshop to 360 dpi, or is it better (or as good) to let the driver resample, which is the recommended workflow presented by Jeff Schewe in the LL Camera to Print video.
There is some logic in my mind that the driver may be able to screen and resize the image in one step better than resizing in photoshop and screening in the driver. But then I'm clueless into the actual math of it all, so my logic may be out in left field
I have recently been using Jeff's approach, sending the master file at native resolution to the printer, and letting the printer driver handle resizing, which outstanding results. The only exception is if I go below about 220ppi, I then uprez the unsharpened version 200%, sharpen that, and again send that to the printer and let the driver size to the final print size.
I have not tested to see if there is any perceptible difference in image quality. The resulting prints are very very good, and it saves so much time in the workflow with great results I haven't really thought about "testing" it. I suppose upon micro examination there may be a difference, but Jeff is pretty clear on the video that normal viewing distances and the ability of the human eye are an important reason why he recommends this workflow.
Since I'm having a lot of fun playing with my new printer, maybe I'll try some tests.
-
Wayne, good if you could run a test or two. I'm in the thickets of testing my 1DsIII just now and have an issue or two I really need to get comfortable about in a hurry.
I agree - Jeff did/does espouse that workflow. However, I know of at least one other technically expert practitioner (un-named to protect the guilty) who resamples to 360 first - but mainly to get an exact match to PK Sharpener Output Sharpening, rather than a specific concern over resampling differences between Photoshop and the Epson driver. I've tried it both ways on different images at different times and I find both totally acceptable. But I haven't tested it on the same image at the same time recently.
-
Hello,
just to clarify: when I was talking about the microscope I was referring to the difference between a print made on default paper size and one made on custom paper size with all four margins set to zero. In that case there is no visible difference, but I wanted to be sure that really there is no difference, as it should be.
The difference between a print made on default paper size and one made on custom paper size with margins other than zero is readily apparent to the naked eye.
So, to summarize my findings:
Under Windows:
print on default paper size -> sharp
print on custom paper size with margins set to zero -> sharp
print on custom paper size with margins other than zero -> sharp
Under OSX:
print on default paper size -> sharp
print on custom paper size with margins set to zero -> sharp
print on custom paper size with margins other than zero -> soft (visible to the naked eye)
I've now tested this on two 7800s and one 9800 with consistent results.
Maybe Epson has fixed the issue with the drivers for the new x880 printers. Maybe it has something to do with the new dithering algorithm, I don't know.
Anyway, thank you all for your help, it's been really appreciated.
-
Glad you found a solution/workaround. That's some extensive testing you did!
-
Now I am somewhat puzzled. I still have my 9800, so curiosity got the better of me.
I followed the exact same procedure as I did with my 11880, printing 3 versions. What puzzles me is the one printed with a custom paper size with 0 as margins is the soft one.
Yes, it's really hard to see with the naked eye, very easy to see with a loupe or scanned image.
I have no idea why the printer driver would be handling the data differently based on the border settings/custom paper settings.
-
Wait a minute - Wayne - you are getting the OPPOSITE problem from Ares on the 9800? He said the soft image is the one WITHOUT the 0 borders, and you're saying the soft one is the one WITH the zero borders? Both of you are using OSX? Are both of you using the same version of the Epson driver? Same photo? Same settings? Something is REAL WIERD!
-
Wait a minute - Wayne - you are getting the OPPOSITE problem from Ares on the 9800? He said the soft image is the one WITHOUT the 0 borders, and you're saying the soft one is the one WITH the zero borders? Both of you are using OSX? Are both of you using the same version of the Epson driver? Same photo? Same settings? Something is REAL WIERD!
[a href=\"index.php?act=findpost&pid=159123\"][{POST_SNAPBACK}][/a]
yes ... that's what has me puzzled. I am using OS X 10.5 and the new leopard beta drivers for all the tests I have printed.
Going to retest tonight just to make sure I didn't mess up the labeling on the 3 prints.
BTW, checked this with a Canon ipf6100, and couldn't see a difference. So at this point, my 3800, my 11880, and my ipf6100 all print the same regardless of paper size/border size choices, but my 9800 seems to be exhibiting the problem discussed here in the thread.
-
Just an update to this, I retested this on the 9800.
I was out of 24" paper so I used 16" paper. The custom paper size with 1" margins required the image to be clipped, so this was a difference between the two tests.
This time there was no visual difference between any of the 3 prints.
I have no conclusions at this point other than there is a problem, it isn't exactly repeatable, and no idea how to insure a custom paper size will deliver a sharp print other than testing each one that is created.
I have a distrust for custom paper sizes at this point however ... I may do some more testing on my 3800 and 11880 just to make sure this problem hasn't migrated to those printers.
-
I have no conclusions at this point other than there is a problem, it isn't exactly repeatable, and no idea how to insure a custom paper size will deliver a sharp print other than testing each one that is created.
[a href=\"index.php?act=findpost&pid=159898\"][{POST_SNAPBACK}][/a]
The one thing I have _NOT_ seen mentioned is what starting page size was selected before going into the custom page size...if one started on a standard page size that represented a borderless printing and made a custom page size from that, I could see where the driver would be doing on the fly resize because that's what borderless page sizes do.
-
has anyone tried to print some digital technical test prints, like a resolution test? surely the best way to test for this, not scanning in images.
I've got my hands for for a few day now but if someone can quickly summarize the settings i'll try some tests on my Epson's.
-
The one thing I have _NOT_ seen mentioned is what starting page size was selected before going into the custom page size...if one started on a standard page size that
I've tested this too, by changing between various sizes of sheet, roll, and borderless roll.
The result is always the same, at least for me: borders other than 0 -> soft print, with the degree of softness varying depending on margins size.
By the way, now I've tested a 4800 under Tiger (not sure what the driver version is, sorry) and the bug affects it too.
-
has anyone tried to print some digital technical test prints, like a resolution test? surely the best way to test for this, not scanning in images.
I've printed some SMPTE test images, just to be sure that the issue isn't image dependent, but I'm not interested in doing measurements, and the softness is easily recognizable on the test image from Bill Atkinson I've always used.
-
I've attached a JPG version of the original image I'm using for my tests. It's already sharpened (0,5-160). Could you be so kind to try with this one?
[attachment=4123:attachment]
[a href=\"index.php?act=findpost&pid=158106\"][{POST_SNAPBACK}][/a]
It's not clear from your post if you're actually using the jpeg attached to your post for your tests? Given that the attached jpeg is a roughly 4"x6" print at 360ppi, I would think that the results would be highly dependent on how it was up-sized if making a large print.
Thanks,
-jk
-
It's not clear from your post if you're actually using the jpeg attached to your post for your tests? Given that the attached jpeg is a roughly 4"x6" print at 360ppi, I would think that the results would be highly dependent on how it was up-sized if making a large print.
Thanks,
-jk
[a href=\"index.php?act=findpost&pid=159925\"][{POST_SNAPBACK}][/a]
The original full version of that image can be downloaded from Bill Atkinson's public folder, it's called lab test page.tif, and is a 5520x6960 image.
That's the file I used when printing tests. There is enough small detail in several areas of the prints, including catchlights as well as text to clearly see when the problem is occurs.
-
It's not clear from your post if you're actually using the jpeg attached to your post for your tests? Given that the attached jpeg is a roughly 4"x6" print at 360ppi, I would think that the results would be highly dependent on how it was up-sized if making a large print.
I'm using a TIFF version of the image. I've uploaded a jpg conversion because I wasn't able to upload the original TIFF on this forum.
And I'm NOT resizing it. I'm printing it exactly as it is. It's a small image because there is no need to waste a big amount of paper and ink to do the testing. Just print it, choosing the right color profile for the paper you're using, twice: the first time on a default paper size (for example A4), the second time on a custom paper size created by yourself. There should be no difference, and indeed when printing under Windows there is no difference at all. Under OSX the results are the ones I've already reported.
-
The one thing I have _NOT_ seen mentioned is what starting page size was selected before going into the custom page size...if one started on a standard page size that represented a borderless printing and made a custom page size from that, I could see where the driver would be doing on the fly resize because that's what borderless page sizes do.
[a href=\"index.php?act=findpost&pid=159900\"][{POST_SNAPBACK}][/a]
I suppose this is possible. I"m not sure why, since it seems all a custom paper size should do is specify the size of the paper and the size of the printable area within the paper. It seems this should be an OS function, yet the problem seems to only affect Epson's x800 series printers (excluding the 3800).
-
And only Mac - not Windows.
-
Final test (for me anyway)
Last night I used the Lab Test Page.tif and created a smaller version by cropping and also copied some of the images with more detail so they would be on the page multiple times. I resized the image without interpolation to 360 dpi, and printed at native size 8 test prints on 16" Epson Prem. Glossy paper (roll).
Test 1 - Created a new custom page size of 16x12.5 inches with .5 inch manual borders on all 4 sides.
Test 2- Modified the previous custom page size, using the pop-up menu to select the printers default border size (which is 0 for the 9800).
Test 3 - Created a new custom page size of 16x13", with default border sizes (again which is 0).
Test 4 - Modified that custom page to have .5 inch borders all the way around.
All 4 of those tests required the image to be clipped, so to insure clipping had no impact, I cropped the image to fit within the page, and repeated the exact same 4 tests. All in all I created 4 custom pages, and modifed the borders of those 4 pages.
This is using Mac OS X Leopard and the new beta drivers available from Epson.
Although in a previous test I was able to see a difference similar to the OP, there is no visible difference in any of these 8 tests to the eye (even with my reading glasses on). I also couldn't find a significant difference using a 10x loupe. If I scanned the images at 1200dpi and compared them, I could see areas that appeared very slightly sharper when comparing a custom border to a 0" border - but this occurred on both prints as compared to each other, rather than one print being sharper in all areas, and even at this resolution those differences may actually have just been me looking too hard.
At this point, at least under Leopard with the new beta drivers, I cannot duplicate or cause a soft print due to a custom page setup. The original test I did which did have an obvious difference I believe was done with the prior version of the 9800 driver.
Curious if the OP has tried using Leopard and the new beta drivers, and what the results might be.
-
I am running Leopard (10.5.1) with the old drivers and a 7800. I will run some tests tonight and then install the new beta driver and run the test again.
-
I've tested the beta drivers as soon as they were released, with the usual result. Just for fun, I've done a bit of code and data analysis (in my former life I was a software engineer) and now I can confirm that the data sent to the printer is different when printing on custom rather than default paper sizes. This does NOT happen under Windows. Note that I'm no longer talking about visually inspecting the prints, I'm talking about examining the raw data stream. Since I have no more time to spare and Epson support has been less than non-existent, I've set up a Windows XP machine, connected to my studio network, and I'm now printing from it. This also solves the issue of unwanted color shifts, wich was never officially acknowleged nor addressed (the usual workaround was to use ColorSync to change the printer default color profile to Generic RGB).
I was thinking about buying an 11880, but I will do so only after personally testing it by printing my usual image under OSX and then examining the prints. Hoping that at least the color shifts issue has really been solved.
-
Ares, this I believe is the first time you mention colour shifts. I have not heard complaints about that, and I know quite a few Mac OSX users operating Epson professional printers without such complaints (there are complaints about the unreliable stickiness of printer settings under OSX, and perhaps those kind of glitches are causing the colour shifts - as you most likely know, when printing with Mac OSX you are best advised to verify the integrity of all the settings at least twice before pressing the print button.
The aspect of it that is most troubling is your comments on support from Epson. I don't know which of their worldwide branches you are dealing with, but here in North America they're pretty good. They virtually closed down their Canadian operation and they provide Canada support from the USA. So far the track record on this side of the Atlantic it appears that on the whole their Pro-Graphics people are responsive and helpful. One gets the impression over here at least that with growing competition for the machines, they want their customers to know they are supported. If this state of mind is not being replicated elsewhere it is something that Epson America and Epson HQ in Japan should be told about.
-
Ares, this I believe is the first time you mention colour shifts. I have not heard complaints about that, and I know quite a few Mac OSX users operating Epson professional printers without such complaints
It has been mentioned even on this forum some time ago:
http://luminous-landscape.com/forum/index....hl=salmon+color (http://luminous-landscape.com/forum/index.php?showtopic=13790&hl=salmon+color)
It's called "salmon color" because the most noticeable shift happens in yellow/orange region. The easiest way to discover if you are affected is to print the first A4 sheet of Bill Atkinson's 1728 target. Since the color patches are sequential, an unwanted shift is immediately recognizable. I've had this problem and solved it by the workaround I've already described. Never heard a word from Epson about it.
Now, the default color profile for Epson LFP has always been "Standard". In ColorSync this is the printer profile with the blue dot next to it.
But with Leopard drivers the default profile has changed to PLPP250. Why? Because this solves the issue? Or because this masks the issue, since PLPP250 is one of the profiles with the larger gamut? It is believed that the color shift is due to an unwanted profile conversion of the already prematched data, so converting to a profile with a larger gamut would have less impact on color appearance.
The aspect of it that is most troubling is your comments on support from Epson. I don't know which of their worldwide branches you are dealing with, but here in North America they're pretty good. They virtually closed down their Canadian operation and
Maybe I'll try to contact them to see if they are interested in my findings.
Just to clarify: I have NOTHING against Epson printers, or Epson itself. I've spent thousand of euros on Epson LFPs and I've been extremely satisfied. I've printed on almost every paper and every canvas on the market, never a problem, the straight paper path is wonderful, never had issues of banding or head strikes, output quality has been (and is) always impeccable.
It's only that Epson support for OSX is lacking, and, at least here in Europe, they are not interested at all in improving it. Just another example: the version of LFP Remote Panel utility available on Epson european web sites is still the original 1.0b wich doesn't work neither with Leopard drivers nor with the latest Tiger drivers. I've had to download the newer version available from Epson USA site. Unfortunately this version doesn't work with Leopard drivers, so for now I have to use the Windows version instead. Note that on european sites the Leopard driver is not a beta one, is presented as final (version 6.11), so we have the driver and not a working LFP Remote Utility. And again, no mention of the problem by Epson. One could waste hours trying to figure out why LFP Remote Utility doesn't communicate with the printer.
-
I've tested the beta drivers as soon as they were released, with the usual result. Just for fun, I've done a bit of code and data analysis (in my former life I was a software engineer) and now I can confirm that the data sent to the printer is different when printing on custom rather than default paper sizes. This does NOT happen under Windows. Note that I'm no longer talking about visually inspecting the prints, I'm talking about examining the raw data stream.
[a href=\"index.php?act=findpost&pid=160482\"][{POST_SNAPBACK}][/a]
So you examined the raw data stream under both platforms? It would seem the streams would be different to some point, to account for various border sizes etc. but that might not affect the actual printed data. I am curious how you "intercepted" the data to examine it.
So your saying your work around of 0 borders doesn't work now?
-
It has been mentioned even on this forum some time ago:
http://luminous-landscape.com/forum/index....hl=salmon+color (http://luminous-landscape.com/forum/index.php?showtopic=13790&hl=salmon+color)
It's called "salmon color" because the most noticeable shift happens in yellow/orange region. The easiest way to discover if you are affected is to print the first A4 sheet of Bill Atkinson's 1728 target. Since the color patches are sequential, an unwanted shift is immediately recognizable. I've had this problem and solved it by the workaround I've already described. Never heard a word from Epson about it.
Now, the default color profile for Epson LFP has always been "Standard". In ColorSync this is the printer profile with the blue dot next to it.
But with Leopard drivers the default profile has changed to PLPP250. Why? Because this solves the issue? Or because this masks the issue, since PLPP250 is one of the profiles with the larger gamut? It is believed that the color shift is due to an unwanted profile conversion of the already prematched data, so converting to a profile with a larger gamut would have less impact on color appearance.
Maybe I'll try to contact them to see if they are interested in my findings.
Just to clarify: I have NOTHING against Epson printers, or Epson itself. I've spent thousand of euros on Epson LFPs and I've been extremely satisfied. I've printed on almost every paper and every canvas on the market, never a problem, the straight paper path is wonderful, never had issues of banding or head strikes, output quality has been (and is) always impeccable.
It's only that Epson support for OSX is lacking, and, at least here in Europe, they are not interested at all in improving it. Just another example: the version of LFP Remote Panel utility available on Epson european web sites is still the original 1.0b wich doesn't work neither with Leopard drivers nor with the latest Tiger drivers. I've had to download the newer version available from Epson USA site. Unfortunately this version doesn't work with Leopard drivers, so for now I have to use the Windows version instead. Note that on european sites the Leopard driver is not a beta one, is presented as final (version 6.11), so we have the driver and not a working LFP Remote Utility. And again, no mention of the problem by Epson. One could waste hours trying to figure out why LFP Remote Utility doesn't communicate with the printer.
[a href=\"index.php?act=findpost&pid=160582\"][{POST_SNAPBACK}][/a]
Thanks for bringing that thread to my attention. I would have ignored because I use Windows XP and have never had a colour management problem on any Epson printer or using any version of the Epson driver since the launch of the 4000 printer.
I don't know how much of a mandate or freedom of action Epson in Europe has to actually fix problems related to operating systems. I suspect there is most likely more analytic infrastructure in EPson USA and certainly Epson Japan for dealing with this stuff. But the very least they should do is take note of the issues, forward them to where they would be evaluated, and give you appropriate feedback. It really shouldn't be necessary to run two computer systems to achieve a satisfactory print.