Lousy Scans from Epson 1680

In one previous post, I received advice on how to profile my Epson 1680.

Eventually I plan to do this for tranparencies so I can scan 4x5 Provia, but

thought I'd start with something simple, where I wasn't looking for exceptional

quality. I bought a Kodak Q-60 reflective target, downloaded the description

file, profiled the scanner in Vuescan, selected the ICC profile under

the "Color" tab, set DPI at 400 (then 800) and set off to convert some 4x6

snapshots to JPEGs. My goal was to feed the results into an online album

program and then have the album printed with the snapshots reproduced in

roughly actual size. The results were terrible, though. The color was ok

(depending on the original scanned) but the general image quality was not,

significantly less sharp than the original, washed out, and with shadows

relatively darker in the scan than in the original (causing odd effects under

the eyes in the subjects' faces, e.g.). I had read that the Epson 1680 was a

high-end prosumer art printer in its day (way back five years ago) and so I was

surprised. I didn't expect anything that would stand up to enlargement, and I

didn't expect that the scan would be indistinguishable from the original, but I

expected something that looked acceptable on its own rather than an image

instantly recognizable as poor. Any ideas on what I might be doing wrong?

Also, does this portend disaster for the transparency unit on the same machine,

where I would want better than merely acceptable scans? Is the problem maybe

Vuescan? (For what it's worth, used as a document scanner, which is how it

came to be in my office at work, I get crisp, clean results.) Thanks in advance.

Could be a focus position setting--I believe VS can override the basic Epson defaults. Try scanning the same picture using Epson Scan (downloadable from the Epson site if you don't have it). Make sure you have the correct setting for flat art (focus position 0). Don't compare color, just see if the scan is as sharp as you want. Make sure to set good resolution values. For a 4x6 print, I would set 4x6 at 240 dpi. for a starter value.

Thanks Les. It occurred to me to try different software, but for reasons I can't explain (and neither can Epson's tech people) Epson's scanner software won't run on my computer. That's how I got to Vuescan, which was inexpensive. I had always intended to use Silverfast before scanning tranparencies, so I think what I'll do is bite the bullet and get Silverfast now to see if it fixes the problem. (A version of Silverfast came with the scanner but was never installed on my computer; I'll track it down.) My guess is that something will work. Some research I did since my last post uncovered an old Shutterbug review of the 1680, where the reviewer said he got better results scanning 35mm slides and printing 13" prints from the scans than he could have from a wet darkroom. If that's true, or even close to true, I know that my unit is either damaged or not setup properly and I don't think it's damaged. I'll report back when I try Silverfast so that anyone searching the site might benefit (even if most have moved on from a 1680 scanner). Thanks again.

Hi

 

well I know I've got a different scanner, but can you perhaps post some sort of sample here?

 

Keep it to less than 500 pixels wide and give it a caption and it'll appear here.

 

I get this sort of result from my 4870<div></div>

for comparison same film scanned at 4000 dpi by a Nikon 8000<div></div>

just to clarify a few things:

<P>

 

<I>Eventually I plan to do this for tranparencies so I can scan

4x5 Provia

</I>

 

so are you having trouble scanning film or prints?

 

cos <BR>

<I>I bought a Kodak Q-60 reflective target,</I>

<BR>

isn't film

<P>

then:<BR>

<I>downloaded the description file, profiled the scanner in Vuescan, selected

the ICC profile under the "Color" tab, set DPI at 400 (then 800) and set

off to convert some 4x6 snapshots to JPEGs.

</I><BR>

less like you've tried film.

<P>

Do you have a sheet of provia to scan? Just give it a whirl if you do.

Unless you're interested in scanning prints that is.

<P>

Since this scanner also does 16 bit, definately put your scan into that

mode when using it (scan with vuescan and then open the tiff).

<P>

You will need some adjustments, but if you're used to looking at digital

camera images they're not as sharp.

Thanks, Chris. So far I've tried only scanning reflective material; 4x6 snapshots, as I haven't yet unpacked the transparency unit yet. I'll try a TIFF scan (feel stupid for not having tried that already); thanks for the suggestion. And I'll try Silverfast. I'll post results if I learn something that I think might be useful. As for your posts, thanks. Also, I liked the comparison to the Nikon 8000. The Nikon scan looks sharper (and I would expect it to be) but it's also smaller. Is it sharper in fact?

In scanning reflective materials ie prints there is only so much info in a print. Thus our Epson 1200, 2400, 3200, and 4800 dpi units all scan prints about the same; typically in the 600 region; with 300 sometimes good enough; 1200 for the rare tack sharp contact print. Maybe there is some auto setting going on with your scanner; here our old usb 1200U epson from the year 2000 scans prints just fine and is total overkill in the resolution area. We manually do the prescan and setting of levels; and even use the stock epson software. The retired Epson 600 that ran on a parallel port was decent too; just slow. Our older 1990's low cost/value?:) Mustek's had poor shadow detail for prints.<BR><BR> 4x6" prints dont always hold all the info in a 35mm negative; one is also getting a truncated dynamic range of the negative by scanning prints. <BR><BR>RE: <i> The color was ok (depending on the original scanned) but the general image quality was not, <b>significantly less sharp than the original</b>, washed out, and with shadows relatively darker in the scan than in the original (causing odd effects under the eyes in the subjects' faces, e.g.). </i><BR><BR>are you calling *the original* the print you scanned; or the negative that made the print; that you didnt scan?

Lobalobo

<P>

be careful evaluating those images (8000 and 4870 scan) as I didn't pay "strict" attention to the exact dimensions. I've redone this below (JPG lossless crops to reduce losses).

<P>

Note that the epson image was 'resampled' back to 4000 dpi <B>after</B> it was sharpened (1.8 pixel radius, 80%)

<P>

also, like Kelly says, prints suck as scan items. Scan something sharp like laser printed pages to get an idea of the scanner sharpness. Pull out a good x8 loupe and look at your source print ... I bet you'll find it looks soft too

<P>

:-)<div></div>

Thanks, Chris and Kelly. I agree with what all say (not that I would be in a position to argue anyway). My scanned prints look significantly worse than the original print. I did this: printed (on an Epson 890) a JPEG from a run-of-the-mill point & shoot pocket camera (Kodak V70, 5mp); then scanned the print and printed the scan on the same printer. I wasn't worried about color (though color matched perfectly, as it turned out); but the print from scan was blurry, not only compared to the original but generally; the blurriness jumped out at you (not a mere softness in the edges). I'm going to try Silverfast, which I understand can focus this scanner, and a TIFF. Will report back.

 

In the meantime, I'm interested (for long run purposes) in Chris' comparison above. What I've read is that for 35mm or MF slides, a good flatbed scanner (say the Epson v750) can match the sharpness of a (twice the price) good negative scanner (say the Nikon Coolscan 9000) but only after the former is sharpened, and the price paid is in sharpening artifacts. In the above, though, saturation aside (which is impossible to judge on a monitor anyway) these two scans look pretty much identical, with perhaps less noise on the bottom one. Is that your impression as well? Would you say that a devoted film scanner isn't worth the money. (I ask for this reason. Once I can figure out how to configure the Epson 1680 properly, I plan eventually to profile and use it for 4x5 Provia, taken with a good Fujinon lens mounted on an old press camera. But in the long run, I hope to get a good MF camera and scan that film; for MF I had thought about the Nikon 9000, which I can afford, but just barely. Would I be wasting my money?)

Ok, lets see.

<P>

<I>In the meantime, I'm interested (for long run purposes) in Chris'

comparison above. What I've read is that for 35mm or MF slides, a good

flatbed scanner (say the Epson v750) can match the sharpness of a (twice

the price) good negative scanner (say the Nikon Coolscan 9000)

</I><P>

 

I don't know about "match the sharpness" but <B>I</B> think

that a dedicated film scanner (especially the Nikons' like LS-40,

LS-4000, LS-50 and LS-5000 <B>are better by at least twice</B>.

<P>

now, the question becomes will that be worth it to <B>you</B>?

<P>

I've bought an Epson because I scan 4x5 <U>not because they are better

or significantly cheaper</U> (when I bought the epson it was $700 and a

Nikon LS-50 was $1000).

<P>

I'm now seriously considering buying a Nikon (on eBay and probably

eBaying again) <U>just for scanning my 35mm stuff that I alreay have</U>.

This is covered in a moment.

<P>

 

<I>but only after the former is sharpened, and the price paid is in sharpening

artifacts. </I>

 

careful sharpening should not give noticable 'artifacts', so if you're seeing

them (IMHO) you've gone too far.

<P>

Looking at the above posted samples notice that you can actually see the

stem of the branch hanging vertically over the car windscreen in the Nikon

scan, <U>but it is not visible in the Epson</U>?

<P>

 

<I>In the above, though, saturation aside (which is impossible to

judge on a monitor anyway) these two scans look pretty much identical,

</I>

<P>

A good point, and importantly <U>will that detail be lost in printing</U>,

I've found that it is.

<P>

I've found that I can't see the advantages that I get when printing from the Epson

VS the Nikon unless being very critical and printing very large

(well for 35mm that is). The reason I want the Nikon is the ease of use.

 

<UL>

<LI>I can load the film into the strip feeder with no hassle comapred to

fiddling around with a tray for the Epson, this reduces the stresses and

hassle in scanning

<LI>the Nikon takes up very little desk space. I can (and do with my older Nikon

LS-20) keep it in the bottom draw, and just pull it out and use it when I need.

The Espon being a flat bed is very deskspace hungry and so requires me to clear my desk

to make space to use it (or get a chair to sit it on).

<LI>ICE works VERY VERY well on the Nikons I mentioned above, much better

than the Epson. Since most of my 35mm is colour this means less spotting and

faster scans.

</UL>

 

<P>

 

<I>Would you say that a devoted film scanner isn't worth the money. (I</I>

<P>

if you fit into any of the above categorys, then <B>yes</B>.

<P>

 

 

<P>

 

<I>Once I can figure out how to configure the Epson 1680

properly, I plan eventually to profile and use it for 4x5 Provia, taken

with a good Fujinon lens mounted on an old press camera.

</I>

 

personally if not doing product photography, where you need to ensure colour match

consistently time and time again, why bother with profiling? Keep in mind that you

might be doing art not commercial art, so coloimetery is not as essential as

<I>getting the look you like</I>. I feel that the purpose of profiling the

scanner is to ensure consistency, so that each bureau will be putting out

results from the same film that look the same when compared (under ideal compairson

environments) with the same film stock.

<P>

 

<I>But in the long run, I hope to get a good MF camera and scan that film; </I>

<P>

 

ok, so are you talking 120 / 220 roll film when you say MF, or are you meaning

4x5 sheet film? You mention a press camera (like speed graphic I assume) so

I'm thinking sheets here ... but if you're meaning a 120 roll film camera

I'd suggest you think carefully about getting a full frame DSLR instead. The

differences you'll find between scans of 120 (at certainly 645 formats and perhaps

6x6 or 6x7 formats) will not be too far away from what you get from the 5D

(and obviously the 1D II and later)

 

the above Nikon 8000 scan was done by a friend of mine who has a stack of good

MF gear in his cupboard and isn't used since he bought his 5D.

 

<P>

Personally I've found that while I get more pixels from a 4000 dpi 35mm scan that

I need them to compete with the results of my 10D. The DSLR makes images which

I can print at 200 DPI and compare with scans which need to be printed at 300 DPI

to not look soft. The DSLR image is really able to <i>skip a generation</i> in

production meaning that the images are 'cleaner' at the limits.

<P>

Have a peek at my <A HREF="http://home.people.net.au/~cjeastwd/digital/" target="_blank">pages here</A>

and the last link on that page "<A HREF="http://home.people.net.au/~cjeastwd/digital/thoughts.htm" target="_blank">thoughts page</A>"

covers some of this sorts of 'musing'.

<P>

<I>thought about the Nikon 9000, which I can afford, but just barely. Would I

be wasting my money?)

</I>

<P>

it seems that many photographers who are using digital are abandoning 120. Digital

capture is better than film in so many respects. 22Megapixel MF backs are starting

to appear in the second hand market too (as newer models attract their owners to

upgrade). 4x5 Scanning backs are far more expensive and require much more storage.

So until we get to the point where

 

<UL>

<LI>storing 300MB per image (storing a RAW capture) <U>and</U>

<LI>moving it fast from capture device to storage system <U>and</U>

<LI>keeping this stuff smaller and lighter than (say) 2 packs of readyloads <U>and</U>

<LI>having its cost low enough that you won't jump off that cliff after it should you drop it

</UL>

<P>

I think LF film will still have a niche as an in the field image recording media,

not to mention effective low priced optical storage media. As you might have guessed

by my keeping my film scanner in my draw, I use my scanner as a kind of "optical

data storage drive" that' optimised for image storage. Like many MO drives (although

there's not much M in there) it has "removable media"

<P>

:-)

 

 

whoops, in replying to a negative question:

<HR width=50%>

<P>

<I>Would you say that a devoted film scanner <b>isn't</b> worth the money. (I</I>

<P>

if you fit into any of the above categorys, then <B>yes</B>.

<P>

<HR width=50%>

<P>

I mistakenly answered with a positive, it should read:

<P>

<I>Would you say that a devoted film scanner isn't worth the money. (I</I>

<P>

if you <b>do not fit</b> into any of the above categorys, then <B>yes</B> it is not worth the money.

<P>

4x6 color prints today are sometimes not as good as ones made say 5 to 10 years ago via optical methods.They vary from dime store to dime store; pro lab to pro lab. 1 hour c41 processing today often develops the film; then scans it; then prints a digital image on the 4x6" print; with optical printing ancient history. Some machines are set up in an average to poor way and one can see pixelation in the 4x6" print from scanning. Another point s a 4x6" print is a 4x enlargement from 35mm; and a print might hold only 7 10 lines per mm; thus the upward limit of the print caps the flatbed scanner; not the scanner itself.

Hi

<P>

since I'm currently fooling around with an Epson 3200, thought I'd give you a sample of what it can do with a sheet of 4x5 ADOX 100 iso film<div></div>

attached<div></div>

Thanks to both Chris and Kelly. On my issue with family snapshots, Kelly hit the nail on the head, as they say. The dime-store prints I was trying to copy were simply of too low quality to get acceptable results. It occurred to me that I should try to scan and save the reflective IT8 target (presumably a high-quality print) and that result was acceptable, as was scanning negatives, each with Silverfast (quite a step up from VueScan). Neither result was outstanding, but color negatives (as opposed to slides) are notoriously difficult, I take it, and the camera was a point-and-shoot and so the negative itself is not razor shop. Problem solved here, or so I think, and thanks.

 

As for the long term and scanning medium format, I do realize, as Chris says, that I can probably get as good or better results with a full-format DSLR, but I had in mind scanning MF film instead until I MF digital backs come down in price into my range. And then I'd have experience with the camera and lenses, which I could just continue to use. That said, everything Chris says makes sense (except that I don't think the Nikon MF scanner, the 9000, I believe, fits in a drawer--I was surprised that it was enormous when I looked at it in a photo shop). Thanks again.

Chris: Since my last post I looked at the pages you linked; thanks for the very useful information. A friend of mine who is an experienced amateur photographer and his friend, who makes a living shooting landscapes (or at least successfully sells her work), agree with your assessment that a 4x5 scanned on a competent flatbed scanner (high-end under $1,000) produce better images from landscape photography than even the best full-sesnor DSLR. I wonder, though, about a comparison of a MF digital back, the most affordable of which (though not quite in my price range at the moment) is the Mamiya 22mp back; no more mp than the highest resolution Canon, but the bigger sensor seems to matter, say some reviews I've read. What do you think?

"a 4x5 scanned on a competent flatbed scanner ... produce better images from landscape photography than even the best full-sesnor DSLR."

 

By far. And really all that competent of a flatbed isn't necessary either.

 

A 4x5 sheet of film has about 80MP of easily extractable image information. If the image can only be captured in one shot and the goal is to print large, there's still really no reasonable alternative to film yet. The real reason though to shoot with a view camera (and sheet film) is the amount of control provided, both in movements and in development.

 

If the subject is static, digital scanning backs are an option. This can exceed film in quality: resolution, tonality, and "grain." The poor man's alternative is to digitally stitch multiple images; this actually works excellently.

Lobalobo

 

Robert is right on the numbers. Actually the numbers are useful in understanding things too. Sheet film is 5 inches wide, so a scan of this at only 600 dpi (barely challenging any flatbed) its 3000 pixels. This is close to what cameras such as 10D give). So even doubling this to 1200 is still within the capture of flatbeds, and making a 6000 pixel image. Thats up there with a 1Ds III (and costing you less than US$10,000).

 

LF cameras however are not a recipe for image clarity, as you can easily under archive the systems ability. For example, recently I was doing some testing of a new film type and development system. I took the image below with my Toho and a 90mm lens, at the same time I took an exposure with my 10D (EF24 f2.8 @ 400ISO).

 

Now, ignoring these scalings, the 4x5 camera image was much softer than the 10D image was (despite being many more pixels). Careful examination of the image showed that during the exposure of the film (10 seconds) some camera movement had happened, this is evident in the elliptical shape of the power LED on the small stereo (bottom right section)

 

Without this reference I'd be left scratching my head as to why the image from film was so much worse than the one from the 10D. the wall paper pattern is clear on the 10D image (colour insert) while not on the large format image.

 

I'm mentioning this because as you reach for higher 'clarity' its important also to revise and refine EVERY part of you image making procedure. If you are after the high definition possible with LF or MF backs then you must make sure you don't smear pixels across the recording media with vibration or movement. In this instance (based on inspecting the film carefully) I'm blaming either; front and back standards vibration, or tripod vibration from the floor.

 

hope this helps<div></div>

Hi!

 

I should mention that I have an Epson 1600 and 1680. These were the top of the line pro graphic arts Epsons of their day. Mine give very sharp results indeed. George DeWolfe makes and sells (or made and sold) 16x20 and much larger prints from 4x5 scanned with these scanners. I find that when I scan a good 4x6 print, I can print another one which is virtually indistinguishable from the original. (Enlargement from prints is another story--possible but far from optimal.) I have not scanned 35mm with them, althought I do have a tranny head; but I get very good results from 2 1/4 film.

Thanks again to all. I will keep all this in mind, including Chris' warning about keeping the camera still (and using good lenses,etc.) On Les' last point, I wonder still why I can't scan a print and get one almost indistinguishable from the original at the same size. As I mentione above, the originals are themselves low quality, but the scanned images are much worse. I'm beginning to worry about the scanner, but more likely I'm still having problems with the settings. As I mentioned, I'm getting acceptable 35mm scans, but even these are not outstanding and perhaps the problem is with color management. I'm going to post on a new thread about that, as I'm experience some oddness with Silverfast, but don't want to change subjects in this thread. Anyone interested in contributing there, I would be grateful (and hope one day to know enough to provide information on this site rather than constantly ask for it).