Thanks drew, I appreciate the effort you are putting into this matter. I'm not insulted about the integrity issue - that matter is behind us. As for the quality of my data, given the large sample size, the systematic manner in which it was collected and the relative simplicity of the collection and calculation procedure, I think that its reliability is quite OK. I doubt Vistek was basing their advertising on my work, because what they quote is a different standard print size, and their ratio 4800/4000 as I recollect is better than my results warrant. So they've done something beyond reading Luminous Landscape. I would like to explore this with them, and living in Toronto as I do, when time permits I can drop in there and ask a few pointed questions. If I do, I'll report here what I learn.
Turning to your quality testing idea, the procedure you suggest is interesting and the adjudication approach you suggest is scientific - like blind-tasting contests of good wines. There are, however, four issues with it:
(1) choice of media and drivers needs to be consistent between all prints otherwise the comparisons would be invalid. For the 4800, a change from matte to non-matte would cost about 150 dollars in ink round-trip for the change between MK and PK inks. For the 4000 you don't face this cost, but the quality difference on non-matte media is APPARANT in respect of bronzing and gloss differential.
(2) One needs to be quite deliberate about the choice of images for such a test, because if you want to stress test for true quality differentials between these machines we would both need to think hard about the kinds of image characteristics required and be consistent about it.
(3) As there may be differences of workflow between us it may be hard to tell what is due to the printer and what to other aspects of the workflow.
Finally (4) if the test were confined to matte media, there would be very little apparent quality difference between them on account of the printers. I have made general comparisons of image quality between the 4000 and the 4800 on Epson Enhanced Matte, and frankly, I would be very hard put to see any eye-popping improvement of the 4800 over the 4000. They both produce stunning results. Black and whites may be a bit more neutral but that is about it. I have also test printed the Gretag McBeth color printer test page on both machines, and apart from slightly more saturation of the colors and slightly more neutral rendition of the neutrals favouring the 4800, on matte paper even these test charts are within spitting distance of eachother quality-wise.
Bottom line, I'm not sure such quality testing would be conclusive, but let me think about it a bit more.