Wayne Fulton's book is worth every penny.
I do A2 prints (on an Epson 3000) from full-neg APS scans done on a Kodak
FD300 (2700 dpi I think). Regardless of the dot arithmetic I'm pretty
pleased with the results. The same setup gives me nice A2 prints from my
Casio 3.3 MP digicam where the dots supposedly ain't enough. It's quite a
Scanners have always given me grief in large lumps of blue sky because they
typically don't handle the gradual shift in colour well and the net effect
is a bit blocky. Select the sky and blur it works every time. I've even
been known to overwrite the actual image with a graded blue wedge except
that tends to be a little good to be true.
It is not unsual to see color banding in the monitor which isn't there in
the print. I think it's just an aliasing thing between the image dot
pattern and the monitor one. This is confirmed by the fact that rotating
the image 90 degrees makes the monitor banding go away.
At 13:56 21/11/01 -0600, you wrote:
>The best place to start is at http://www.scantips.com/
>Wayne Fulton's information is succinct and accurate and will give you much
good information. He also sells a book including more information, though
I have not bought it.
>Optimal resolution to be sent to an hp printer is 300dpi, and to an Epson
is 360dpi. A scan of a 35mm film image results, on the LS-30 at 2700spi
yields a 3800x2500 pixel image, approximately. Sending that to the printer
results in a 360dpi image at the 8x10 size, but only 220dpi at 11x17. If
the resulting 11x17 print looks good to you, then all is well.
Theoretically, a 4000spi filmscanner would result in a better print.
>The 'grain' is hard to deal with - enlarge to 100% and see what it looks
like. You may have to apply some noise filters in Photoshop or whatever
image program you are using - median, gaussian blur, dust & scratches, or
whatever works. Be sure to apply them in individual color channels only -
I find it helpful to change to LAB color space, apply the median filter in
the A and B channels, and dust & scratches in the L channel.
>Banding is generally a problem in broad expanses of the same color -
especially blue sky. A way around it is to select the sky using a mask of
some sort, then apply (again) a median filter in the appropriate channels,
or a gaussian blur. This will even out the colors and inhibit banding.
>----- Original Message -----
>From: "John Pendley" <email@example.com>
>Sent: Wednesday, November 21, 2001 11:14 AM
>Subject: Re: filmscanners: Novice scanner
>| At 09:42 AM 11/21/2001 -0600, you wrote:
>| >First question - how large do you like to print? If larger than 8 1/2 x
>| >11" then you probably should get a 4000spi scanner instead of the
>| >Acer. If not then the Acer should be fine.
>| I like prints larger than 8 1/2 x 11 if the resolution is there. I
>| math formula somewhere that produced the number 11 for the width of prints
>| with a 2720 scanner. I supposed that would mean an 11" square print, or
>| maybe an 11 x 17. You don't think the ScanWit will be satisfactory at
11 x 17?
>| >The grain that prints out - do you see it onscreen as well or only in the
>| It's hard to say. When I enlarge the color image that has all the red
>| dots, almost to the stage of seeing the pixels, there is definitely color
>| spotting in the shadowed part of the church wall, but it's not the same
>| dots that I get in a print. When I enlarge the B&W image on the screen,
>| more and more grain becomes visible.
>| >If only in the print then it sounds like a printer problem - try cleaning
>| >the nozzles with Epson's software.
>| There's definitely some kind of "grain" on the screen, though not exactly
>| what I'm getting in prints, at least in color, anyway. I have used
>| three utilities (clean nozzle, etc.) three or four times.
>| > If onscreen then it may be what is called "dust and scratches". You
>| > have probably read references to Digital ICE and to Vuescan's Dust and
>| > Scratches filter similar to ICE - they both help but they need an IR
>| > (infrared) channel to work. Another possible reason to switch scanners
>| > as the ScanWit 2720S does not have the IR channel though their 2740 does.
>| I didn't know that Vuescan had an ICE-type feature; as you say, it would
>| not help with the 2720, anyway.
>| >There is also an article on what is called Grain Aliasing at
>| >http://www.photoscientia.co.uk/Grain.htm - you might want to read that.
>| Thanks; I'll read it right after I finish here.
>| >Despite the discussions I would not be leery of the Nikon - I have the
>| >Nikon LS-30 and am very satisfied with it. I also would not worry about
>| >Polaroid's problems - the price is excellent with the $200 rebate, and
>| >David Hemingway of Polaroid participates in this group and has assured us
>| >the product is still being manufactured and rebates processed. I tend to
>| >trust his word on that.
>| That's good to know about David. At the moment, if I were to get a new
>| scanner, it would be between the Polaroid and the Canoscan 4000. The
>| might be a terrific scanner, but the price is a bit high for me.
>| >The banding problem is more particularized - when you print an image,
>| >format is the image in (PSD, TIFF, JPEG, or what?)? And what is the
>| >resolution that you are sending to the printer?
>| First, I have prints with no visible banding. They were made using the
>| same settings as the prints that do show banding. I have been printing @
>| 6.666 x 10, as I like the 2:3 format of 35mm. I've set the resolution at
>| 300 dpi, and all prints are from TIFF files. What seems even more
>| is that the slide that shows the most banding also shows bands in the scan
>| on the monitor.
>| >Good luck to you! I started late myself.
>| >Maris Lidaka Sr
>| Thanks for your willingness to help Maris. I hope that my answers to your
>| questions will give you some new ideas.