Paul-
Thanks for your comments.
>
>>>> 1. Does anyone know what the resolving power of the Epson printers
>>>> are
>>>> (4 lp/mm 10lp/mm, 20 lp/mm)?
>>
>>> Depends on the printer. You can do the math...if it's a 720DPI
>>> printer,
>>> that's 360lp/inch, divide that by 25.4, or 14.7 lp/mm.
>
>> Yes, I can do the arithmetic but that seems to be a theoretical
>> maximum. What if the line pairs are grey. I thought that the 720DPI
>> is 720 printer dots per inch and it takes several printer dots to make
>> a grey. Doesn't dithering reduce this number.
>
> I get a range of about 7 to about 12 if I'm lucky. The file size
> increase
> from 360 to 720 lp/mm adds very little. I'm assuming the printer is
> set to
> 1440.
>
> The Epson 3000 was not quite as good as my 1160 and the more modern
> printers
> I've tested, but close enough that in the real world, it would
> probably not
> matter.
>
> The gray and black ink resolution is, as I recall, usually about the
> same.
> Of course, the lighter ink would be harder to see, so it might not
> look as
> sharp.
>
>> I have printed the 21step greayscale that someone provided (Paul
>> Roark?) which has small line pair patterns on it. At 3.6 lp/mm the
>> contrast is high and the lines obvious, at 7.2 the contrast is very
>> low
>> the lines are not distinct at all. Under a loupe the vertical lines
>> are much worse than the horizontal. In practical terms is seems to me
>> that the resolving power is much less that 14.7.
>
> That sounds about right. Be careful to check the dpi of the test file.
> It's supposed to be 360. One was originally put in the Files section
> that
> had a resolution of only 300 dpi, which would throw off the numbers.
It was the 300 dpi one. With the 360 dpi file the 7.2 lines resolved
easily. With this test and another test file I made my numbers are
similar to yours. Greater than seven and less than 10.
>
>> It also seems to me that the printers resolution is dwarfed by every
>> other upstream component and I wonder whether you could even tell the
>> difference between a good lens and an adequate lens in an inkjet
>> print.
>> Or whether, for a fixed print size, you could tell the difference
>> between a 4000dpi scan and a 2000dpi scan?
>
> In my enlargements -- 16 x 20 and 22 x 28 -- the printer is usually
> not the
> limiting factor. Only my best negatives from the best medium format
> cameras
> can deliver enough resolution through my Nikon 8000 at 4000 dpi to get
> near
> perfect sharpness across the printed enlargement. The show I'm now
> preparing for has some of my old 16x20 silver prints in it also.
> These were
> shot with Zeiss medium format lenses (Rollei SL66) on Tmax 100, with a
> tripod, and enlarged through Apo-Rodagon lenses. They are, generally,
> not
> as sharp as the digital inkjet prints.
This is very interesting! If you are enlarging be a factor of 7-10 and
the print is limiting at say 7 lp/mm then you need to deliver about 50
to 70 lp/mm from the scanner/lens/film before the printer is the
limiting factor. That's quite a high number.
>
>>
>>>> 4. Do papers have a significant impact on resolution/sharpness
>>>> number
>>>> of tones rendered and tonal transition
>
> I have not measured the sharpness of the top RC papers, but my sense
> of it
> is that they show a lot more of the digital artifacts. Part of this
> (probably most of it) is that the increased RC contrast makes defects
> visible that were hidden on matte papers.
>
>> For the numerically inclined it would be interesting to construct
>> MTF(like) charts for paper/printer/driver combinations.
>
> I wish I could. I don't know how to measure the relative contrast of
> high-resolution targets. I suspect it takes some expensive equipment.
Yes, I thought about problems of measurement only after writing it. It
would be interesting though.