Replacement 4900 Gamut Volume Smaller than Original 4900

[Shane Webster]

I received a replacement 4900 yesterday that I set up, performed a nozzle check, aligned the print head and printed new IGF targets for profiling.  I've now read the patches and have an interesting result.  The gamut volume of the new printer is 914,514 and the gamut volume of the old printer was 935,147, a difference of about 2.2%.  As far as I can tell--other than the printer--everything was the same-- same paper lot, same test chart, same device measuring the patches, same program creating the profile, same settings on profile creation, and same dry down time for the prints.  I wouldn't think there would be that much difference between two 4900 printers.  I'm thinking about re-printing the test chart just in case the printer needed to run through some ink but that doesn't really sound like the issue, but I don't know.  I've attached two pictures from ColorThink that shows the gamut difference--all lower luminance colors.  Any thoughts on why there should be a difference or is the percent difference not material in the overall scheme of things?

[Mark D Segal]

The problem may be the measuring device. I know they can produce different readings of apparently the same thing, but to what extent I'm not sure. Did you let it warm-up the same amount of time before use for the two printers? When was the last time you had the device checked and recalibrated? What device is it, BTW?

[Shane Webster]

It's an iSiS.  I did let it warm up.  I actually measured two charts and averaged them together.  The worst ten percent only had a difference of 0.27% with the sigma worst ten percent being 0.07% (best 90% is at 0.04%).  While I haven't had it calibrated, I profiled the IGF for my previous 4900 about a month ago and it hasn't been used too many times since.

In re-printing the test chart, I've realized there were some differences.  In an attempt to use the same test chart for both PMP and i1P, I saved my patch set from i1P and generated a test chart with MT.  The test chart generated from MT matched, in layout, the test chart generated by i1P.  Initially I read both charts, averaged them with MT and then created the profile within i1P based on the average (gamut 914,024).  I then read one of the charts within i1P (since we can't average two measurements in i1P but that's another thread) and created a profile from just it (gamut 914,454)--I did this in case the issue was from using a measurement file from outside i1P. 

From this, I see a couple of potential variables-- 1) The test chart generated from MT may be somewhat different from the test chart generated from i1P (I wouldn't think so since it was based on the patch set from i1P, but it is a difference).  To remove this variable, I'm re-printing the test chart generated by i1P; 2) The profile created with i1P measuring the chart was read from a 2nd scan of one of my test charts which, I suppose, could generate different results if some scuffing occurred while reading the chart the first time through (especially given the sensitive nature of IGF).  In reading the chart from item 1, this variable should remove itself.  Unfortunately, I cannot open the CGATS file or the CxF file I can save from i1P in MT (again, something for another thread) to compare the results.  I also created an additional profile in PMP that is off by about 19,000 colors, but the old PMP profile was created from one of Bill Atkinson's charts so I'm re-printing that one as well in order to re-create a profile within PMP based on the same testchart previously used. 

All of this switching between PMP and i1P is really aggravating me since I can't have both my PMP and i1P dongles connected to my computer at the same time and must swap them out.

[Shane Webster]

After some fun with excel and text wrangler, I managed to get the CGATS information from i1P in a format MT could use for comparison purposes.  The worst 10% Delta E Average between the new 4900 and old 4900 is 1.1% with a total of 0.62%.  The maximum is 2.44% with the best 90% being 1.01%.

[Mark D Segal]

Not being an expert in how these comparative measurements should cohere, but it seems to me you're in good territory.

[Shane Webster]

Yes, I don't know what the expected Delta E tolerance is between two similar Epson professional printers, but the average does seem low.

[Ernst Dinkla]

Next week it may have a wider gamut than the old one. Shaken the carts thoroughly, humidity at print time on the same level, the old printer also measured the same day you installed it ?

If it keeps that lower gamut maybe they changed the pigment load to give you 100& nozzle quality :-)  Small price for more convenience.


met vriendelijke groeten, Ernst
Try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/