Profile Plotter graph question

[mikemeister_admin]

Hi All, I have attaced a zip of the output PSD file from the profile plotter graph action template, I am interested in what people make of it.

I use an ordinary Dell CRT monitor and an Epson R800 with Epson Ultra Glossy paper.  I loaded paper profiles from epson, then I profiled my screen & printer using a Spyder & PrintFix Pro.  This profiling process has helped me avoid me from printing outside the colour gamut.  Yellows were a particular problem that are now manageable (perhaps not perfect) through post processing.  The one thing that I could never find a workflow to address is printing shadow detail, this was really only a big problem when darks were the subject matter. For example, photos of gorilla's in Uganda.  Shifting the black point up on the curves helps a little, but it is far from perfect.

I am suspecting that the printer has to go back to Epson and this is not a profiling/calibration problem.  Any ideas?

[I've added a gif of the same file for convenience - Mike]

[-default]

The profile plot looks fine.  There are no severe jaggies, and it adds magenta to the image, which is characteristic of Epson printers, which tend to print cool out of the box.

I would not suspect a problem with your printer at this point.

Your problem with shadow detai is not going to be helped by this profile because of relatively level region along the left (black) edge, which will tend to compress shadow detail. You're basically losing the bottom 1/8 of your tonal range, printing as pure black. 

Also, the red+green blip in the shadows bugs me.  The profile is saying "the blackest areas of the calibration target are too blue, so I'm going to add yellow light to them to get the color closer, even though that screws up the blacks.  Rich blacks are one of the assets of this excellent printer, and adding yellow to the shadows is not helpful.  To get rid of this, use Black Point Compensation when plotting (or printing with) your profile, and always use Relative Colorimetric intent.

It's far more likely the difficulty is either with your calibration device, or your procedure.  Make sure you work in a relatively dark room.  It is standard procedure to use a second sheet of the same paper underneath your calibration print.  Using a very dark surface, and keeping the room dark, may help with the Red-Green blip in the shadows as well.   Do not use a colored surface under the print.

If you are able to read Lab values from the spectrometer, you may want to invest in one of Wolf Faust's color targets to do a sanity check.  Faust is doing us a service by making these targets available for $6 and the quality is excellent.

[mikemeister_admin]

Hi, I am in the office at the moment so I can only provide limited information for the time being.

The printing I use is Relative with black point compensation.  The calibration was done with multiple peices of white paper under the chart whilst in the presence of modest light, so I will re-scan the 749 squares in the dark tonight.

TIA, Anthony.

[-default]

I would start by experimenting with the smaller, 150 square target.

The mystery, IMHO, is finding out why the calibration software wants to add yellow to the shadows.  If you can solve that issue, you may get rid of that flat spot in the shadows.

If your device is capable of reading a direct Lab value, that would be a good prelude to doing a full calibration, since you could see if the Lab value changed for the darkest patch.

[mikemeister_admin]

Thanks, I guess since the values are so extreme that such a precise profile is not required for this analysis.

The calibration device is a humble point an click mersuring device made by datacolor (see below).  I am unsure as to how I could retreive usuable data strings from each click.

Just a thought, could this also be because the white point in the device cradle requires a clean?

[-default]

Very suave looking device, I must say. 

It's not the calibration tile - that would cause the white areas to go out of whack.  Worth looking on the lens of the spectro, though.  A bit of semi-transparent bluish material - ink or fluff -  on the lens would be exactly the kind of problem that would cause this.

BTW - another obvious cause for a bad calibration would be scanning the wrong swatch at some point.  This is remarkably easy to do, and an error like that can perturb the entire profile in unpredicatble ways.  Does the profiling software give you a deltaE reading after it calculates the profile?  If there was a very large number - more than 6 or 7 - a mis-scanned swatch is a common cause for this.

[mikemeister_admin]

Hi, I'll check for those physical aspects (ie. fluff) tonight.

The Sypder3Print software that I am using has a chart to show which swatch you are reading as you go, where one half of a square is the color and the half is the spectrometer colour reading.  The variance displated is purely visual (ie. not in RGB HEX digits).

It should be noted that I do see some significant difference between the swatch sample colour and what is returned.  Not understanding the science behind icc/icm, I had been put that variance down as part of the profiling process, after all.. things were much worse before i got this a screen and print reader.

It would be disappointing if my spectrometer unit were also faulty.  I mean how many, checks and balances does one have to introduce to print a picture acurately.  ;)

[mikemeister_admin]

Hi,

I have found where to export the 729 LAB values (with line numbers) to text and I have attached them in this message.  I am unsure how to translate the three columns of values, the first looks like a percentage.
  There is still too much ambient light to do another spectrometer scan, I will be back later for an update.

Regards,

Anthony

[-default]

The darkest patch I see is patch #1, with a value Lab(14.05, -1.58,-4.27).  The 14.05 Lightness value is too high.  A glossy paper should measure 2.0 or less.  I would not bother to scan any more patches until this is solved.

Next on the agenda is to test your spectro.  Turn off the room lights, point the instrument out into empty space, and see how low a Lightness value you get - hopefully close to zero.  If it's daylight, don't sweat it.  Turn off the room lights and use a relatively dark area, such as under the desk or wrap the device in cloth.

If the spectro tests OK, I would start looking at the paper.

BTW - the a and b values indicate a fairly hefty blue color, which is consistent with a blue object in the optical path.  The profile is trying to compensate for this by adding red and green to the shadow.

[mikemeister_admin]

Hi,

Sorry I couldn't get it done last night because I got back too late to scan 150 squares.  What I will do tongiht is get some Epson Semi-Gloss paper and try calibrating against the paper drivers for that as well.  I thought the Epson Ultra Gloss Photo A4 available here in Australia would have been the same as Premium Gloss Photo Letter available in the US, so I used a the US PGPP driver on their website.  I remembered I could never find an Ultra Gloss Photo Paper driver ANYWHERE on the internet despite the packaging saying it supported the R800.  Perhaps the Ultra version has more blue in order to give it that extra shine and maybe because it is more glossy it requires less ink acheive saturation (hence the clipping)?

I am not much of a fan of semi-gloss, so my long term prints won't go on this paper type but it will at least be interesting to see the results.

US Paper driver link: http://www.epson.com/cgi-bin/Store/EditorialAnnouncement.jsp?oid=51716719

(the Australian driver link page is hopelessly out of date)