Ah - it wasn't clear from your initial post - & 4800 is an oft cited resolution on Epsons. Carry on as before!Originally Posted by RodinalDuchamp
Ah - it wasn't clear from your initial post - & 4800 is an oft cited resolution on Epsons. Carry on as before!
Thank you! And all of your posts were very insightful I appreciate your responses.Ah - it wasn't clear from your initial post - & 4800 is an oft cited resolution on Epsons. Carry on as before!
I use 4200 on my V850.
To obtain a very good result with an EPSON we have to scan at higher resolution and later downsampling if we want.
While a Pro scanner tends to deliver a digitally optimized image with the EPSONs we always have some job to do in Ps to get a sound image, the sharpening and the downsampling (use "bicubic, ideal for reductions" choice) will be critical.
The V700 - 850 has two fixed scan widths, 5.9" with the highres lens, or 8" with the lowres lens that is used automaticly when the scanner detects the area guide.
With selected 5.9" scan width the EPSON delivers 2400 average effective dpi in all formats, from 35mm to 5x7", it delivers exactly the same performance in all those formats.
While for 35mm 2400 effective can be low in some situations, for sheets this is a really good performace, and for MF it's more that correct.
Interneg says that the 2048 hardware dpi of the X5 (1800 effective) in 4x5" are much better than the 2400 effective of the EPSON. I don't agree, absolutely, with that. We have discussed that too much, to the point that I think it's time to make a side by side comparison to enlight what an EPSON does or not in LF, as it is nearly the single suitable tool that most of us can get, without going to very old gear that may involve severe service shortcomings.
So I'll prepare a side by side, including Velvia with multiexposure...
Interneg, if you want I'll send you the test sheets, give me an address by PM.
If I were to scan at 6400dpi and say do a 2x2 binning at scan time. I can reduce noise, but will I retain quality of 6400? The main purpose/goal would be to have a smaller file, but with quality of larger file if possible.
I've not an answer, you should test it and see, but let me describe some involved factors:
With the V850 if you scan at 6400dpi you don't have a 6400 effective dpi quality, you can get a maximum 2900 effective dpi in the sensor direction and 2300 in the movement direction, some passes will delliver a 10% less effective rating depending on the sensor position on the target or vibrations, or who knows...
That excessive dpi brings on excessive file sizes (that today are less excessive than in the past), but it also prevents aliasing artifacts allowing for a more effective sharpening later.
A sort of binning is what scanner does when scanning is not performed at maximum resolution. The scanner can do that in two ways 1) it can use a feature of the sensor that averages the analog value of several pixels to deliver a single analog value for them to the A/D converter, or 2) it can average digitalized values of individual pixels, I don't know how the V850 does it... but if using a binning in the scanner software then the binning is made digitally for sure.
A binning made "after the A/D conversion" will lower the A/D conversion noise by averaging several A/D conversions, compared to making an analog average in the sensor and then making a single conversion.
I guess that A/D noise is low, but it would be interesting to measure it in a dense area. To make that measurement I would use a gray smooth target, and I'd place it well out of focus on the scanner, then we would measure the variance of the hue/lightness of a crop and compare both ways (binning vs scanner lower dpi).
Let me describe my workflow with the V:
> Use a SSD disk (it can be 64 GB only, cheap) for recording scanned images. Or clone your system disk to a SSD, before that move big data to regular HDD if not wanting to pay a big SSD. Your problems with big files will evaporate, and Ps will start fast.
> I scan at a resolution that will be higher than the one that I want to use in Ps edition. Always I take all histogram so I override the suggested DR end points on the histrogram, I always scan 16 bit. Save always TIFF because if not only 8 bits are saved.
> Edit always in 16Bits per channel, I perform a mild USM sharpening and then I reduce to the edition size selecting "Bicubic, ideal for reductions". Then I save a copy as my scanned original in TIFF.
> Edit the image. Save a second copy Edited, 16 bits, top quality, in TIFF.
> Make release image. Proofing, adjustments for the display medium (for print or monitor or...).
> Image size reduction to the diplay medium (pixels in the monitor, printer ppi), using bicubic (ideal for reductions). Another USM touch, a bit depending on the viewing distance. 8 bit per channel conversion. Save a third TIFF copy. Save a 4rt copy in jpg if necessary.
Some people say that printers have very good algorithms and they prefer to not send a pixel for each printer pixel, but a +140% larger ppi and allowing printer optimize the size reduction. I think that those people are not weird, but anyway what will happen will depend on the particular printer, so for best prediction of the result I prefer sending one to one pix.
im sorry, i haven't read the whole thread but
your digital book will end up costing more than your hand printed book easily
if you add it everything .. scanning, tweaking, printing &c
unless you get light jet/ digital c prints made of all your prints and don't go overboard with your scanning / tweeking.
a lot of people think that you can really tell the difference when things are printed small
but to be honest unless you are printing something gigantic i really wouldn't go overboard
( unless you want to of course ! )
making a 1-off hand made book is not hard, it can be done easily without much fuss.
japanese bound books are the easiest to make, and you don't even need to stitch them
you can make it a "post binding" . the only "tricky" part of the whole operation is
gluing/pasting the top edge of your prints to a receiver page. but even that isn't hard..
the next hard thing is adding thickness to the spine where the binding will be so
the block of the book doesn't fan and it lies flat ( unless you want it to fan than it is even easier ! )
you have to add thickness to compensate for the thickness of the prints.
there are online tutorials on making japanese bound books, there are easy to follow books
on making them too that can be picked up for not too much $$ through online sellers like amazon..
closed spine is a little more time consuming. you need to make a stitching rack and sew the pages
together and glue them into a spine but its not really that hard .. i've been making open and closed spine books
since i was IDK 15 or 16 years old .. it just takes a little effort.
library supply houses like gaylord brothers will easily sell you the materials you need
or places like harcourt bindery in boston. there are book presses bought cheap but you can just use
something heavy ( look for a book called books boxes and portfolios )
you can get book-board to make your covers from art supply stores
and they will even cut them down to size... wheat paste is cheap and easy ( lineco sells it
also purchased online cheap and easy ) ...
IDK the idea that pigment printed images cost less than darkroom prints is a myth ..
light jet is cheap as dirt by comparison and RC prints glue easily to rag paper
have fun with your project !
john
a digital image also can be printed in photopaper with a lambda, frontier or lightjet...
but a hand printed book it's something very special !
yup
cheap and easy and no fuss if that is the route one wants to go ..
IDK maybe its me but scanning at 4800dpi &c for a 5x7 or even 8x10 print seems like overkill..
and
yup
John, my problem is another one: the amount of paper boxes I've to spend to get a nice resultyup
Let me add that lightjets also are used to print (rather C) BW photopaper, Ilford Lab Direct prints in RC (gloss and semi), until 10" they use a Frontier and a Lightjet from there, but Bob prints also FB with the Lambda, allowing for a top Q result.
Posting Permissions
Reply With Quote
Bookmarks