Epson Stylus 1500w - Missing details at 1440 ppi

[Gubenco]

Hello, I am trying to print at maximum resolution with an Epson Stylus 1500w.
There is some missing lines (both y and x) to the squares and circles. With circle is even more interesting because it looks thicker in some parts where usually the detail is missing. It seems like there is a sort of moire effect where my 1440 ppi image meets a "non 1440dpi printer pattern". Diagonals look ok.
The effect is present on both Y and X axis with the same amount. They say there should be 5760 ppi on X axis I see nothing like this.
If I print 720ppi images there is no problem and dots are placed quite even.

I am sending 1440ppi aliased images to the printer and I have tried different setups all at highest resolution. I am curious if this is a known artifact or my printer is damaged. All nozzles look clean but the printer is not new. It is using an aftermarket CISS (it was installed by the vendor). I am using INK master dye inks.
Regards

 

[Ink stained Fingers]

Which program are you using to exactly print with 720 dpi resolution ? You may use the Roger Clarke test pattern as described here

https://www.printerknowledge.com/th...on-tests-epson-and-brother.13640/#post-118307

before you use more complex patterns - curved etc which will give you additional interpolation artifacts.

You may redo the head alignment before you do further testing.

Epson printers use a image pixel resolution of max 720 dpi, such dots have a color which needs to be rendered within these dots, and this happens with a higher resolution of max 5760 dpi x 2880 dpi or 1440dpi , this handles the individual colored ink dot placements within the 1/720 " image dot. You as a user do not have access to this individual droplet placement, that's driver/firmware internals. Canon printers typically run at 600dpi, and the ink droplets are placed at a max raster of 4800 or 9600 x 2400 droplets per inch.

 

[Gubenco]

Thank you @Ink stained Fingers for the test you provided.
I've put my printer to this trial, I do not know what to say... I definitely see some 1440 ppi capabilities. Of course I notice a pour text reproducing (even at 720ppi) and it has a lot of dot gain ( I have used a photo paper but for plain paper the efect is even worse). I won't restrict sending 720 ppi to the printer anyway (I do notice more details at 1440ppi but this is teh limit for my printer). What I truly liked here is that I have noticed at least one dot for every pixel, so the question remains "Why that co-linear pixels/dots are missing, in my circles and rectangles, for such a large area". I have tried to alter resolution up and down like 1439 - 1445 and I see some lines appearing and disappearing, It is definitely a moire effect. I have forget to mention that I am sending to printer from Adobe Photoshop and I use the Epson driver which doesn't specify the resolution (it says only Photo RPM (max dpi)). Is there any other driver with more control. I suspect here is the issue.

 

[maximilian59]

Could you please show the result of the latest nozzle test print? It looks like a recurring pattern.

 

[Gubenco]

Hello @maximilian59.
Here is the nozzle test and please believe me, those rectangles are the same image rotated at 2 and 5 degrees.

 

[Ink stained Fingers]

I'm not sure to understand why you try to print with 1440dpi, an Epson printer won't do that.

The printing software Qimage displays the active printer resolution depending on your driver settings, and you'll see that you don't need the RPM option activated to print with 720dpi. Qimage let you print either with a pre-selected dpi or the embedded dpi - embedded into the image.



Or you may use the alternate driver Turboprint which let you select as well the printing dpi.

I have tried to alter resolution up and down like 1439 - 1445 and I see some lines appearing and disappearing, It is definitely a moire effect

Yes, that's basic match with spatial frequencies - you get the difference and the sum of those, the higher frequencies cannot be resolved, and you get the moire effectc incl. extinction of particular lines , That's as it is.

You may do some more tests with the attached test files , rotate them if you like to get lots of additional patterns. The text file explains the basic structure of the dot/line pattern fields. And you can test the pattern output when you change parameters - dpi - driver options like edge smoothing, sharpening - alternate interpolation routines when turning the image - in Qimage. And if you don't want to print them all - you can print to an image file with Qimage and view the result on screen - including possible moire between the monitor and the printing dpi.

 

[maximilian59]

1440ppi aliased images

As I don’t know, what is an aliased image?
What are the native PPI of your file? If PPI and DPI differs, a recalculation ist done. If your picture has eg 720 PPI with wanted sizes, than a printer cannot put more information in it. With printing on higher DPI, you tell the driver to use more possible places to put a dot but not more pixels. Whether it is done by the dithering algorithm or not and where nobody knows. If your file is upscaled for printing before it is sent to the printer, there might be the first problem. This is then done by photoshop not the driver.
For example if you print with 180 DPI draft, it doesn’t matter whether your file has 720 PPI, the printer driver recalculated them to the wanted resolution. Also in the Fuji files is written, don’t alter the size! Only use different printer settings.
Cheers,
Maximilian

 

[Ink stained Fingers]

Also in the Fuji files is written, don’t alter the size!

That's pretty important - there may only be one method - the seldom used 'pixel resize' method - you magnify a pixel by x2 or x3 without any influence by neighboring pixels .
Which print resolution can you effectively expect - image a sequence of line pairs black and white - with the width of 1/720" each - the ink of the dots in the black (or colored) lines spread by 50% to the left and right and just close the gap/cover up the gap and the gap is not visible anymore. And you now print with 1/360" width - the ink spread remains half of the 1/720" so you get at the end a gap width of 1/720" left which remains visible.
This matches about the eff. resolution I have seen over the years on plenty of printers - Canon or Epson which does not make much of a difference in this respect - most of these printers do about an eff. resolution of 300 - 350 dpi, this varies slightly with the paper type, the coating controls the width of the ink spread/dot gain. And the visibility of the gap varies with the print orientation - horizontal or vertical. The printhead cannot print slanted/tilted lines directly , there is some interpolation software in between before the actual print.

 

[Gubenco]

As I don’t know, what is an aliased image?
What are the native PPI of your file?

I use 1440 ppi, Aliased image are generated without filtering and pixel interpolation is made using nearest neighbor method.

@ Ink stained Fingers
I have tried Qimage and I think there may be "some" potential. It printed the missing details wit much smaller dots I suppose those are the 1.5 pl droplets and I think I have seen them for the first time in a real print. (please take a look to the attached image).
I have used the Overdrive -1440dpi and ti worked in a satisfactory way but the frustrating issue is that I could not print the entire image with these settings, only 1-2 square cm or something no more. If I print a larger image it prints as previously with missing parts,with larger droplets and simpler blacks. When it prints well, it has also a more rich black color with CMYK. I do have also a driver error in qimage when the driver takes over the settings (attached). I can not control this part, It looks like the printer has a limited budget of small droplets and it switches automatically to lower resolution mode. when Epson driver sees to much work, it panics and switches to dumb mode. The printing settings and the source images are the same except it works only when I crop the image.

 

[Ink stained Fingers]

I'm still not clear why you would need to generate and print an image file with a 1440dpi setting. And I never have seen complications with Qimage as you describe , I don't know whether it's a memory/file size problem or anything else software or system related, I just would not think that it ' panics and switches to dumb mode' on its own.