Howtek vs Imacon vs Epson

[Michal Makowski]

I have founded something interesting.

http://blog.timparkin.co.uk/2008/08/scanner-comparison-epson-vs-imacon-vs.html

Michal

[Stephen Best]

Take a look at the "densewithgrass" sample. Move your mouse to and fro between "Show Imacon" and "Show Howtek". Notice the vertical displacement of the rock just right of centre? That's the cause of the CA. It will also contribute somewhat to softness. The Imacon needs servicing.

I don't know why people persist in generalizing based on tests like these.

[Michal Makowski]

The Imacon needs servicing.
Wow!!! Even superpro's, like Joe Cornish had servicing their scanners:D

[Lenny Eiger]

Just posted a comparison in the Howtek 4500 thread. Take a look..

Lenny

[timparkin]

Take a look at the "densewithgrass" sample. Move your mouse to and fro between "Show Imacon" and "Show Howtek". Notice the vertical displacement of the rock just right of centre? That's the cause of the CA. It will also contribute somewhat to softness. The Imacon needs servicing.

I don't know why people persist in generalizing based on tests like these.

I also got a scan made by another provider (Ian Scovell) and there was very little difference. His scanner had been meticulously set up by Hassleblad (Joe's was serviced by Hassleblad in the previous year also).

If you can't generalise based on real world tests, what can you generalise on ;-)

Tim

p.a. I've since bought my own Howtek and am getting a lot sharper results than those in the test.

[timparkin]

the url should be http://www.timparkin.co.uk/blog/scannercomparison I had the blog moved into my website and didn't get all the 301's in place..

I should also add that I'm happy to send a transparency to anyone with an Imacon that thinks they can get a better result.. (if they can include a raw file too)..

[Ken Lee]

I must have missed where film size is specified.

To get more out of a humble Epson scanner, you can switch from 4x5 to 5x7, a diagonal increase of 36%, and a 75% increase in film area.

Sometimes we forget that 4x5 is at the low end of Large Format.

It costs less to scan larger images with an affordable scanner, than scanning small images with an expensive scanner.

[mrladewig]

The Howtek will always get noticeably better shadow detail in dense images regardless of the resolution, and this may be important for some who shoot films like velvia with dense blacks. And if you want to shoot anything color besides Portra 160NC, 5X7 probably isn't going to be your best format for a reliable film supply.

[Ken Lee]

Oops - I forgot about color film. You're right !

[Ron Marshall]

Oops - I forgot about color film. You're right !

For b/w and a 3 to 4X enlargement I am happy with the output from my 4990.

But when I print anything larger, or in color, I always go for a DS.

[Ken Lee]

The 4990 delivers around 1600 spi. If you print at 600 dpi or more, then 2-3X is around the limit. (Even office laser printers print at 600 dpi, because otherwise, fonts are not smooth).

That's why I shoot more 5x7 these days: it gives some room for cropping if necessary, and one can make a fine 11x14 with only a 2X enlargement. For 4x5, an Epson 700 is a better choice: it delivers more like 2300 spi. With that, one can make up to 3-4X enlargements.

[Ron Marshall]

The 4990 delivers around 1600 spi. If you print at 600 dpi or more, then 2-3X is around the limit. (Even office laser printers print at 600 dpi, because otherwise, fonts are not smooth).

That's why I shoot more 5x7 these days: it gives some room for cropping if necessary, and one can make a fine 11x14 with only a 2X enlargement. For 4x5, an Epson 700 is a better choice: it delivers more like 2300 spi. With that, one can make up to 3-4X enlargements.

Ken, I think you may be confusing DPI and PPI; see the link:

http://www.rags-int-inc.com/PhotoTechStuff/Epson2200/

[SAShruby]

The 4990 delivers around 1600 spi. If you print at 600 dpi or more, then 2-3X is around the limit. (Even office laser printers print at 600 dpi, because otherwise, fonts are not smooth).

That's why I shoot more 5x7 these days: it gives some room for cropping if necessary, and one can make a fine 11x14 with only a 2X enlargement. For 4x5, an Epson 700 is a better choice: it delivers more like 2300 spi. With that, one can make up to 3-4X enlargements.

Based on my own Epson 4990 testing, with calibrated scanner, I scanned Transmission Resolution Guide from Stouffer in different ppi settings and the best resolution I got was @ 2300 ppi, above that I didn't see any improvements in sharpness. I scan @ 2400ppi which is the closest setting option. I can achieve 6.67 enlargement from scanned area if I set to 360dpi for printing, 8 @ 300dpi. Scan is wet mounted on the glass and option is full area transparency.

[Ken Lee]

"Ken, I think you may be confusing DPI and PPI; see the link: http://www.rags-int-inc.com/PhotoTechStuff/Epson2200/"

I may be wrong, but I think my terminology is consistent with that article.

Scanner terminology: spi = samples per inch

Printer terminology: dpi = dots per inch

Monitor terminology: ppi = pixels per inch

The scanner gives us spi, and the printer gives us dpi. I refer to scanner performance in terms of spi, and printer settings in terms of dpi.

[Ken Lee]

"Based on my own Epson 4990 testing... the best resolution I got was @ 2300 ppi"

The best resolution we get is when the scanner is set to "2300ppi", but what we actually get, is 1600 spi of resolution.

Scanner manufacturers are free to identify the settings on their scanners with any names they choose - like A, B, C - or Low, Medium,and High. However, actual measured resolution, is independent from the name of the setting, or what the manufacturer claims for that setting.

[Brian Ellis]

The 4990 delivers around 1600 spi. If you print at 600 dpi or more, then 2-3X is around the limit. (Even office laser printers print at 600 dpi, because otherwise, fonts are not smooth).

That's why I shoot more 5x7 these days: it gives some room for cropping if necessary, and one can make a fine 11x14 with only a 2X enlargement. For 4x5, an Epson 700 is a better choice: it delivers more like 2300 spi. With that, one can make up to 3-4X enlargements.

Ken - Every published test I've seen of the 4990 has shown that it's capable of actually resolving about 2100-2200 (ppi/spi, same thing different output device, generally used interchangeably to describe an image's resolution). Maybe there's something wrong with your scanner if you're only getting 1600. Or maybe the tests were wrong but I saw several and IIRC they all agreed. I believe the results of one were published in View Camera magazine but I could be wrong, it's been quite a whle.

Or maybe my memory is wrong which normally I'd say was the most likely possibility except that in this case I compared tests of the 4990 with the 700/750 when it came on the market, saw that their spi/ppi was about the same, and therefore passed on a purchase of a 700/750. At least I think I remember all that.

[Ken Lee]

Brian and Peter - I'm glad to be mistaken on this one !

I am relying on my memory - of testing done by Sandy King with a resolution target.

Thanks for the correction !

[Nathan Potter]

A couple of months back I did a resolution test on a new V750 pro using a chrome resolution mask and mentioned the results in another thread here (forget which one). At point of best focus (3mm above platen) I easily obtained 2500 spi non wet mount.

The important caveat in my results is that the glass resolution plate is of infinitely high contrast and so does not yield any information about a scanned subject of continuous tone - that is MTF kind of data. But the data I did obtain confirms a lot of previous results reported about the 750 resolving capability. I have been quite impressed by the clunky little machine given its consumer like quality.

Nate Potter, Austin TX.

[SAShruby]

Ken, I think you may be confusing DPI and PPI; see the link:

http://www.rags-int-inc.com/PhotoTechStuff/Epson2200/

Ron,

This is great info. Now I do understand why people recomment to switch to 8-bit Grayscale. Printers today just cannot print great 16-bit tonal range without sacrifice of the resolution.

288ppi @ 2880x1440dpi would be my best option for Epson 7880, giving me 200 tonal ranges in grayscale [5x10x(3inks of grey+white base)], which provides acceptable optical resolution. I hope I got it right. If yes, I will try to print someting tonight and compare to my previous print which was printed as 720ppi @ 2880x1440dpi resolution. If my calculation is correct this seting gave me only (4*2*(3+1)=32 tonal ranges....

[sanking]

Brian and Peter - I'm glad to be mistaken on this one !

I am relying on my memory - of testing done by Sandy King with a resolution target.

Thanks for the correction !

Your memory is not wrong. I owned an Epson 4990 for several years and tested it with resolution targets at least three different times, each time making sure to place the target at the optimum scanning distance. The best resolution I ever got was around 1600 spi.

I now own an Epson V700 and have tested it several times. If you scan with the lower resolution lens (Epson calls it the "High Resolution Lens") the real resolution is around 1600 spi, about like the 4990. But if you scan with the "higher" resolution, which is engaged when you select "Film Holder" the real resolution is around 2300 spi.

Sandy

[Ken Lee]

Thank you !

Ever since being reminded that even office laser printers output at 600 dpi, I have been suspect of the 360 dpi output so popular among Epson users. I need to find out whether 720 is worth the extra pixels, on the paper I use.

If 720 is indeed a preferred setting, this suggests that one can enlarge around 2-3 times at most using a 4990... Ouch.

[rguinter]

I must have missed where film size is specified.

To get more out of a humble Epson scanner, you can switch from 4x5 to 5x7, a diagonal increase of 36%, and a 75% increase in film area.

Sometimes we forget that 4x5 is at the low end of Large Format.

It costs less to scan larger images with an affordable scanner, than scanning small images with an expensive scanner.

Ken: I have no experience with the original subject of the thread. But I noted your suggestion to switch up to 5x7 film.

Certainly a good idea when/where it can be done. But the feasibility of actually doing that is not so good any more.

I have a freezer full of current and vintage films in 4x5-inch. i.e., RDP, RDP-II, RDP-III, RVP, RVP-100, RAP, TP, RTP, RTP-II, VPL, VPS, TMX, TP, EPY, E100VS, HSI, Ektar-100, Adox-25, Efke-Aura, NSP, CDU-II... and probably some others I've forgotten.

Not many in this list can be gotten in 5x7.

For me and my hobby work the Epson 4990 has been an excellent setup and there has been no need to seek any higher resolution. When it crashes for good I'll upgrade at that time. Bob G.

[Brian Ellis]

Thank you !

Ever since being reminded that even office laser printers output at 600 dpi, I have been suspect of the 360 dpi output so popular among Epson users. I need to find out whether 720 is worth the extra pixels, on the paper I use.

If 720 is indeed a preferred setting, this suggests that one can enlarge around 2-3 times at most using a 4990... Ouch.

It's always seemed to me that the entire area of output settings is a morass of conflicting information. Some say you should send prints to the printer at the printer's native resolution (which is where your 360/720 numbers come from). But there doesn't seem to be agreement on what that resolution is for Epson pro printers such as my 3800 printer. Some seemingly reputable sources say it's 360ppi, other seemingly reputable sources say it's 720 (and then of course there are those who say it doesn't matter, that sending at an even multiple of the native resolution is as good as sending at the native resoltuion so if you send at 360 you're fine either way). Then the person in the article linked above said that his testing showed that the native resolution of his Epson 2200 printer was 288, not 360 or 720.

Then there's the more fundamental question of whether you should even try to send at the native resolution (or an even multiple) whatever that is. Some say the resampling usually needed to send at 360 (or 720 or 180 or 288) creates bigger problems than sending at the original resolution whatever that may be. Others disagree. And still others (e.g. the well-known and very reputable Eric Chan) take a middle course and say it depends on when you do your sharpening, that if you sharpen after resampling then resampling is fine.

I haven't bothered to provide cites to all this conflicting information, it's easily found just by doing an all-words search in Google on "epson 3800 native resolution) and reading the first four or five links that come up.

My own practice is to not resample, not try to send everything to the printer at the 3800's "native resolution" whatever that may be, and instead to just send at whatever the resolution is at the print size I want, as long as that resolution is at least 240 ppi which it almost always is. I've been happy with my prints doing that though I'd be happy to change methods if I could find any general agreement on a new and better method.

[Ron Marshall]

My own practice is to not resample, not try to send everything to the printer at the 3800's "native resolution" whatever that may be, and instead to just send at whatever the resolution is at the print size I want, as long as that resolution is at least 240 ppi which it almost always is. I've been happy with my prints doing that though I'd be happy to change methods if I could find any general agreement on a new and better method.

I try to send 360 ppi to my 3800, as in some images I can see a difference between 300 and 360. But for larger prints with a greater viewing distance 240 is fine.

I have tested several images using both 720 and 360 ppi files, (Epson 3800 at 2880 dpi) and never have seen a difference, except in spooling time.

[Darryl Baird]

Your memory is not wrong. I owned an Epson 4990 for several years and tested it with resolution targets at least three different times, each time making sure to place the target at the optimum scanning distance. The best resolution I ever got was around 1600 spi.

I now own an Epson V700 and have tested it several times. If you scan with the lower resolution lens (Epson calls it the "High Resolution Lens") the real resolution is around 1600 spi, about like the 4990. But if you scan with the "higher" resolution, which is engaged when you select "Film Holder" the real resolution is around 2300 spi.

Sandy
Sandy,
What is the resolution setting in the (Epson?) software that delivers this "real resolution" with the "film holder" setting? 2400 dpi or ???

thanks

[Darin Boville]

Sandy,
What is the resolution setting in the (Epson?) software that delivers this "real resolution" with the "film holder" setting? 2400 dpi or ???

thanks

And what setting, if any, in VueScan uses the higher res lens?

--Darin

[Ken Lee]

What is the resolution setting... ?

You scan a resolution target. You specify a 2400 spi scan, and you get a file where there are 2400 pixels per square inch.

Now, you look at the file and determine how much data can be distinguished from the scan. You find out that you can distinguish the target lines which correspond to 1600 spi.

The German web site www.filmscanner.info (http://www.filmscanner.info/en/EpsonPerfectionV700Photo.html) did a study of the Epson 700 in this way, using the classic USAF resolution target. As you can see in the article, they tell you that the actual resolution is closer to 2300 dpi. They show you the image of the target. It's all straightforward.

With the 4990, you can scan at any higher setting you like, and you will get a correspondingly larger file - but the extra pixels are just more... blur. You can still only distinguish 1600 spi. The same is true with the 700: there is an upper limit, and it's not the one claimed by Epson's advertising.

You might find it interesting to look on the same site. and see how they rate the Nikon Super Coolscan 9000 ED (http://www.filmscanner.info/en/NikonSuperCoolscan9000ED.html). They distinguish between vertical and horizontal resolution, and measure it quite precisely. The Nikon's actual resolution matches the advertised claim - a more common occurrence with dedicated film scanners, than with affordable flatbeds.

For scanning color materials, we are limited, as it were, by the fuzziest of the 3 channels. Scanning b&w, we can sometimes get the sharpest results by using only the green channel (http://www.kenleegallery.com/html/tech/scanningGreen.html) - which is usually the best of the 3.

[sanking]

Sandy,
What is the resolution setting in the (Epson?) software that delivers this "real resolution" with the "film holder" setting? 2400 dpi or ???

thanks

Scanning at 3200 spi will give you nearly the full 2300 spi. Scanning at 4800 spi and 6400 spi will give a very slight improvement.

I can not help you with Vuescan as I hav not used it with the Epson V700, only the Epson application.

Sandy

[SAShruby]

Your memory is not wrong. I owned an Epson 4990 for several years and tested it with resolution targets at least three different times, each time making sure to place the target at the optimum scanning distance. The best resolution I ever got was around 1600 spi.

Sandy

No offense but I respectfully disagree with this observation. As I said before I was able to perform best scan at 2300spi/ppi as the best resolution on my scanner which is Epson 4990. I ran tests too and that's what I observed. I won't let myself deprive of additional information by scanning at poor 1600 spi.

[sanking]

No offense but I respectfully disagree with this observation. As I said before I was able to perform best scan at 2300spi/ppi as the best resolution on my scanner which is Epson 4990. I ran tests too and that's what I observed. I won't let myself deprive of additional information by scanning at poor 1600 spi.

Good for you, but I reported my test results, not yours. You have nothing to disagree with me about unless you are saying I did not observe what I said I observed. I have never claimed my results were anything but my own observations.

However, you may have misunderstood what I said. I did not say that the best results were obtained when scanning at 1600 spi, but that 1600 spi was the highest effective resolution I was able to observe, even when scanning at 2400 spi, 3200 spi and 4800 spi.

Sandy King

[timparkin]

Good for you, but I reported my test results, not yours. You have nothing to disagree with me about unless you are saying I did not observe what I said I observed. I have never claimed my results were anything but my own observations.

However, you may have misunderstood what I said. I did not say that the best results were obtained when scanning at 1600 spi, but that 1600 spi was the highest effective resolution I was able to observe, even when scanning at 2400 spi, 3200 spi and 4800 spi.

Sandy King

totally agreed Sandy - in fact in my test the best results are had when scanning at 6400dpi (just to give a high frequency noise base that is easily removed) although I reckon over 1600 dpi (it compares niceley with a 2000 dpi drum scan - in resolution terms anyway)

Tim

[timparkin]

http://static.timparkin.co.uk/static/scanner_comparison_1/epson-imacon-howtek-raw.jpg

This is best real world comparison I can come up with...

[Kirk Gittings]

A useful comparison, but as always only compares those particular scanners/operators. The problem is we never know how talented the operator is or in what shape the scanner is?

FWIW, I own an Epson 750, have access to the latest Imacon, and use a service for drum scans. Why the difference? Why use them all?

Anything quick and dirty I do at home on the Epson. I have had this Epson since they first came out and have experimented with it wet etc. ad nauseum. I find them useful but pretty low in quality-no matter how you use them. If I want to do a quality small print and am short of cash, I go for the Imacon because I can do them for free in the summers or pay a friend locally $15 anytime. IME I can get a better scan from the Imacon than I can ever get from my Epson. that si either me as an operator on the SAIC Imacons or a good local guy on his. If I am doing a print larger than 16x20 (or ever expect to need a print larger than that), I get a drum scan which run me about $150 including shipping. Currently I am using Lenny Eiger, on an Aztec Premium, for drum scans.

I have paid plenty of money over the years for commercial flatbed and drum scans that were not much better than the Epson. Its easier to find a good shrink than a good scanner operator (and about the same price per hour!).

No samples to show, but tons of personal comparisons along the way-I have tried every which way to skin this scanning cat and the above are my conclusions based on my requirements to do exhibition quality B&W from 4x5 over the last 6 years. i would love to think I could get more out of my Epson but I just haven't seen it-the difference even between it and an Imacon is tactile and the difference to a good drum scan is.....substantial. In the abyss between the Epson and a good drum scan I find the Imacon a notch above the Epson but still a few notches below a good drum scan.

[sanking]

IME I can get a better scan from the Imacon than I can ever get from my Epson. that si either me as an operator on the SAIC Imacons or a good local guy on his. If I am doing a print larger than 16x20 (or ever expect to need a print larger than that), I get a drum scan which run me about $150 including shipping. Currently I am using Lenny Eiger, on an Aztec Premium, for drum scans.



I had a friend's Imacon Flextight Precision IV here at the house for a few months last year and compared several medium format negatives with my Epson V700 and Eversmart Pro. I used the maximum optical resolution of each scanner, 3200 spi for the Precision IV, 3175 spi for the Eversmart, and 6400 spi with the Epson V700. I then compared results on the screen.

The Imacon scan was much better than the scan with the Epson V700, but not quite as good as the scan made with the Eversmart.

I also did the same comparison with a 4X5 negative, again using the maximum optical resolution of the three scanners, 2040 spi for the Imacon, 6400 spi for the Epson V700, and 3175 spi for the Eversmart. In this case the Epson scan was a tad better than the one made with the Precision IV, but the Eversmart Pro scan was vastly superior to both of the others.

If one were scanning 35mm negatives there is no doubt in my mind that the Precision IV at about 5400 spi would give a much superior scan to the V700 and Eversmart Pro, but its maximum resolution of 2040 spi puts it in the marginal category IMO for 4X5 film.

I was pretty disappointed for my friend with the results since he already had an Epson V700 and a Nikon LS-9000 and IMO the Imacon Flextight IV was not as good as the V700 for 4X5 film, and not as good as the LS-9000 for medium format film.

Sandy King

[Kirk Gittings]

Sandy, How were you scanning the 4x5 on the Epson?

[sanking]

Sandy, How were you scanning the 4x5 on the Epson?


I used a BetterScanning holder and adjusted the mount for best plane of focus, which is about 2.5mm above the glass bed. I mounted the negative to the underside of the AN glass, base side to the AN side of the glass, emulsion side down facing the CCD. I selected film holder to make sure the better lens was engaged.

I did not bother to fluid mount the negative with the V700. Fluid mounting will improve a scan with the V700, but does not improve resolution much, if at all.

Since I have the Eversmart Pro I normally use the V700 rather like you, for rough work and proofing. However, my opinion is that in a pinch the V700 is capable of a real good print up to about 4X in size from a good B&W negative.

Sandy King

[Kirk Gittings]

That is exactly how I used mine, but on mine the resolution is just not there.

[sanking]

That is exactly how I used mine, but on mine the resolution is just not there.

Kirk,

I tested two different V700 scanners with a high resolution target and both gave effective resolution of around 2300 spi when scanning at 6400 spi. And this figure is the same as the test done at ScanDig.

http://filmscanner.info/en/EpsonPerfectionV700Photo.html

You might have a bad apple, or perhaps since you are used to working with drum scans perhaps you just expect more sharpness than the V700 can give.

Sandy

[PenGun]

That is exactly how I used mine, but on mine the resolution is just not there.

Have you focused the Epson? A big difference between stock heights and the right height.

[Kirk Gittings]

You might have a bad apple, or perhaps since you are used to working with drum scans perhaps you just expect more sharpness than the V700 can give.

Possible on both counts.

[timparkin]

A useful comparison, but as always only compares those particular scanners/operators. The problem is we never know how talented the operator is or in what shape the scanner is?

Erm... well - the Imacon scanner is the best person offering scans in the UK and handles scans for one of the most published landscape photographers in the UK. Also, I've had the same transparency scanned by Imacon owned by the 'other' most publised landscape photographer in the UK (with similar results)

The Drum scan is at 4000dpi and you can compare that with one of your reference drum scans if you like (the picture is at 200%). I've compared the same transparency with two other drum scanners (favourably) and so I'm happy it's 'good', if not perfect

The Epson - well it doesn't matter. The results show a good comparison against the imacon so if the operator was anything less than 'good' then the results could be even better!

I should also add that I've used a 20x loupe and I can't see any additional detail in the transparency past the drum scan (which I'm presuming I would if it was a bad scan).

What concerns do you have about the operator/scanners? How could the results be read differently without presuming that two high profile photographers use significantly and equally compromised scanners?

[Kirk Gittings]

Nothing personal. There are just so many variables that I have learned to distrust scanner comparisons, including the ones on this website. A few examples.

I have had TERRIBLE scans done on pro flatbeds and drum scanners by "experts" highly recommended by big name photographers. Bad day, equipment out of whack-who knows? Does that say that their scanner is not as good as X scanner. No.

I have witnessed dramatically different interpretations of scanned targets-the industry standard for judging resolution?

I have now owned some 11 odd pro-summer flatbeds over the years and have spent countless hours trying to maximize their output to no avail by my standards. Does that mean they suck for everybody. No, it only means they don't meet my idiosyncratic standards.

IME Imacons get out of whack very easily-they need to be serviced every year and after any move.

Sharpening is a huge issue in comparisons, especially if Imacons are involved because a zero setting does not mean no sharpening (I can't remember the number but somethingg like -20 is actually 0 sharpening).

An actual experience-I have seen comparisons done that put Epson scans in the same ballpark as good drum scans. You ask for the original file with the exif data intact and never see it and the comparison is pulled from the website a few weeks later. HMMM.

Trying to judge results by 72 ppi images posted on forums is difficult at best. To me the Imacon sample has better resolution and shadow detail than the Epson, but in all truth-it is pretty hard to tell. The drum scan seems clearly better in both respects.

Plain and simple after years of dealing with this I just can't put much faith in scanner comparisons because there is little control of possible variables. Now granted a comparison that did control all the variables, new machines, factory tuned, one knowledgeable operator etc. would be difficult and expensive to set up, but until we see that kind of controlled test, comparisons can be helpful, but not definitive.

[timparkin]

Nothing personal. There are just so many variables that I have learned to distrust scanner comparisons, including the ones on this website. A few examples.

I have had TERRIBLE scans done on pro flatbeds and drum scanners by "experts" highly recommended by big name photographers. Bad day, equipment out of whack-who knows? Does that say that their scanner is not as good as X scanner. No.

I have witnessed dramatically different interpretations of scanned targets-the industry standard for judging resolution?

I have now owned some 11 odd pro-summer flatbeds over the years and have spent countless hours trying to maximize their output to no avail by my standards. Does that mean they suck for everybody. No, it only means they don't meet my idiosyncratic standards.

IME Imacons get out of whack very easily-they need to be serviced every year and after any move.

Sharpening is a huge issue in comparisons, especially if Imacons are involved because a zero setting does not mean no sharpening (I can't remember the number but somethingg like -20 is actually 0 sharpening).

An actual experience-I have seen comparisons done that put Epson scans in the same ballpark as good drum scans. You ask for the original file with the exif data intact and never see it and the comparison is pulled from the website a few weeks later. HMMM.

Trying to judge results by 72 ppi images posted on forums is difficult at best. To me the Imacon sample has better resolution and shadow detail than the Epson, but in all truth-it is pretty hard to tell. The drum scan seems clearly better in both respects.

Plain and simple after years of dealing with this I just can't put much faith in scanner comparisons because there is little control of possible variables. Now granted a comparison that did control all the variables, new machines, factory tuned, one knowledgeable operator etc. would be difficult and expensive to set up, but until we see that kind of controlled test, comparisons can be helpful, but not definitive.

I have to say that a test done by 'knowledable operators' is bogus (who says they are knowledgeable). A test done on 'factory tuned machines' is bogus (who has such machines or can guarantee scans done on them).

The only real tests are real world scans done by people who have access multiple scanners. Scanners that have been maintained by the manufacturers.

If we're trying to prove how *bad* something is then we're on a losing streak. However, tests to show something cheap is good when the tester has something expensive to use is more easily beleived.

I'm offering drum scans for a fee, why would I overstate how good an Epson is in this case? (also, a bad scan doesnt' mean a scanner is bad - a good scan does imply something else - you don't *fluke * good scans)

The Imacon comparisons were just done on the only available Imacon scanners I know who have a decent reputation.

Anyway - the best way of assessing how good certain scanners are is by looking around the internet and making your own judgement of the trustability of the reviewers (you definitely can't judge by asking the scanner manufacturers - bias - or by doing scans yourself - unknowledgable operator).

I got results from the Epson that made me worried about my Howtek for a moment - getting those results was a long winded process however (lots of focussing, 4x multi sampling, 6400dpi scans).

In my opinion, Epsons are great for 4x and you can push to 5-6x on scanner friendly trannies.

Where the drum scans really win are colour, Dmax and lack of halation/aberration.

Tim

p.s. Check the man sitting on the boat on the left and also the dividing lines in between the panes of glass in the small house.

[Richard Mahoney]

I have to say that a test done by 'knowledable operators' is bogus (who says they are knowledgeable). A test done on 'factory tuned machines' is bogus (who has such machines or can guarantee scans done on them).

...

Anyway - the best way of assessing how good certain scanners are is by looking around the internet and making your own judgement of the trustability of the reviewers (you definitely can't judge by asking the scanner manufacturers - bias - or by doing scans yourself - unknowledgable operator).

...

You know, Tim, this is exactly the opposite way that I would approach this whole issue. ... For me, the primary thing is to maintain the highest quality possible at each step along the path from the transparency to the print. Granted that a transparency is actually worth printing -- at this end, never a given ;) -- then the next absolutely critical phase is the (drum) scanning. When selecting someone to do the scanning I would *never* do a web search on all the various available machines.

The first thing I would do would be to go to someone whose work I respect and ask them to recommend someone. You see, the `operator' (if one must belittle someone with such a title) is primary and the machine they have chosen to work with just follows.

Once one is in front of someone known for high quality work, then one can just have a chat about how they approach their work in general and what they think may be best approach with specific transparencies.

This `non-web-based' approach has worked for me. When I first started my enquires I was completely unaware that just ten minutes down the road, in the middle of the countryside, working out of an office just across from the house, is one of the best fellows in the country. And as he is good at what he does then he has chosen a decent machine -- a Fuji Celsis 6250 (Crossfield). But really, this all just follows. In many ways, when you're dealing with people who care about what they are doing, it can be taken for granted, on trust. And it is not something I would have found on the web. My `operator' -- cringe -- doesn't have a `web-presence' per se, and neither, for that matter, does his Celsis.


Kind regards,

Richard

[sanking]

In my opinion, Epsons are great for 4x and you can push to 5-6x on scanner friendly trannies.

Where the drum scans really win are colour, Dmax and lack of halation/aberration.



Pretty much agreed about your conclusions as they apply to the Epson V700/V750. Prints up to about 4X from B&W negatives look to me about as good as prints from drum scanners and high end flatbeds, and I have made comparisons with Howtek 4500, 6500, and a ColorGetter, plus my own Eversmart Pro.

That is for well exposed and developed B&W, however, not color transparency material. For color slides the Epson just does not cut the mustard IMO. But why would anyone be shooting color slides these days? Color negative film is sharper and easier to scan.

Sandy King

[onnect17]

Color negative film is sharper and easier to scan.

Sandy King

Sandy,
Could you clarify how/why color negative film is sharper?
Regarding the scanning, the information in color negs is compressed allowing increased lattitude per density, but you will waste 3 or 4 bits (typical) of data compared to the chromes.
Also, the howtek is one of the few scanners built with a pre-amp to switch between linear and log (very useful for color negs).
Regards.

[sanking]

Sandy,
Could you clarify how/why color negative film is sharper?
Regarding the scanning, the information in color negs is compressed allowing increased lattitude per density, but you will waste 3 or 4 bits (typical) of data compared to the chromes.
Also, the howtek is one of the few scanners built with a pre-amp to switch between linear and log (very useful for color negs).
Regards.

First, Howtek is a great scanner, no doubt about it. But my Eversmart Pro is just as good with B&W and color negatives as the Howtek 45. And more versatile since you can scan a negative up to 12X17" with full resolution, and also reflective material of the same size. For color transparency material the Howtek will do a better job.

Color negative film of the same ASA usually has higher resolution than slide film. There are exceptions but my experience is that this is usually the case. Also, the dynamic range of color negative film is much longer than that of slide film, at least three or four stops and perhaps much longer if you scan.

I have personally exposed Kodak Portra 160 in conditions where the SBR was greater than 15 and was able to capture the entire range with a scan. Typically the curve of color negative film levels off at a certain point and all highlight values beyond this point are compressed on the same line. You could not print this in the wet darkroom but if you scan it is possible to adjust the curve to increase contrast in this area. Pretty remarkable what you can achieve, especially if you adjust the colors in the conversion from RGB to B&W in PS. This kind of range is not even remotely possible with transparency film.

Have you tried Ektar 100? Way more dynamic range than any slide film, virtually grainless in any size print, and sharpness is way up there.

Sandy King

[onnect17]

First, Howtek is a great scanner, no doubt about it. But my Eversmart Pro is just as good with B&W and color negatives as the Howtek 45. And more versatile since you can scan a negative up to 12X17" with full resolution, and also reflective material of the same size. For color transparency material the Howtek will do a better job.

Color negative film of the same ASA usually has higher resolution than slide film. There are exceptions but my experience is that this is usually the case. Also, the dynamic range of color negative film is much longer than that of slide film, at least three or four stops and perhaps much longer if you scan.

I have personally exposed Kodak Portra 160 in conditions where the SBR was greater than 15 and was able to capture the entire range with a scan. Typically the curve of color negative film levels off at a certain point and all highlight values beyond this point are compressed on the same line. You could not print this in the wet darkroom but if you scan it is possible to adjust the curve to increase contrast in this area. Pretty remarkable what you can achieve, especially if you adjust the colors in the conversion from RGB to B&W in PS. This kind of range is not even remotely possible with transparency film.

Have you tried Ektar 100? Way more dynamic range than any slide film, virtually grainless in any size print, and sharpness is way up there.

Sandy King

Sandy,
I agree that chromes do not compete in SBR with color negs, but in situations where the 80% of the subject is contained in 2 or 3 stops I avoid color negs.
Regarding the grain, it's so fine in today's films with ISO <= 100 that I use 4000dpi without any problem.
I will order some Ektar 100 in 4x5 from B&H. In about 3 weeks the colors of the fall will peak in NH and will be a good test.
Regards,
Armando

[sanking]

Sandy,
I agree that chromes do not compete in SBR with color negs, but in situations where the 80% of the subject is contained in 2 or 3 stops I avoid color negs.
Regarding the grain, it's so fine in today's films with ISO <= 100 that I use 4000dpi without any problem.
I will order some Ektar 100 in 4x5 from B&H. In about 3 weeks the colors of the fall will peak in NH and will be a good test.
Regards,
Armando

Very unusual for me to photograph subjects where 80% of the scene is contained in two or three stops. Much more common to find range of 12 stops or more.

Ektar 100 on a low contrast day should give you great results with fall color. With full sun you might be better off with Portra 160 as it has higher dynamic range.

I am going to China in mid-October and plan to take 10-15 rolls of Ektar 100 with me to the Yellow Mountains, and will hope for overcast conditions.

Good luck with your fall colors at the peak in NH. Our colors in the mountains of North and South Carolina happen late October/early November. But it is as impressive as New England.

Sandy

[Richard Mahoney]

Sandy,


... That is for well exposed and developed B&W, however, not color transparency material. For color slides the Epson just does not cut the mustard IMO. ...

On the whole, that has been my experience ... and a case to point. Over the past week I've been quietly scanning a backlog of Astia for the web. Overall, not having especially high expectations, I've been relatively happy with the performance of the v700. I haven't been asking the 700 to produce scans comparable with my drum scans -- even after they've been significantly downsized for the web from their original 24"x30" -- so I haven't been disappointed. There have been, though, one or two exceptions, and I'd be lying if I said that I didn't find them profoundly irritating.

The image below is one of them. The original transparency, despite the fact that it's Astia, is highly saturated and contrasty, containing a great deal of edge definition, grass, gorse, fine branches, strongly marked shadows ... it is, for me, exactly the reason why I love transparencies, looking at it gives me a great deal of pleasure, something a negative never really comes close to. But the v700 scan ... the scan, despite all my attentions ... well ... it's an abomination, nothing close to the original.

I've scanned reasonably `flat' transparencies with the 700 without any real issue and they've been fine, but this particular image defeats me. The trouble is that I know that the fellow down the road with the Celsis could just knock it off without thinking, but, although I'm very tempted, I simply can't justify drum scans for the web -- and anyway, the thought of such a thing really is absurd. I'm hoping that despite the shocking state of this image someone will decide to order a print, cover the cost of a decent scan, and solve my problem ... but until then, I'll just have to accept the limitations of the v700. Hopefully though, I'll never be obliged to shoot accordingly :)


Best, Richard


http://camera-antipodea.indica-et-buddhica.com/portfolios/portfolio-one/cust-anglican-cemetary/cust-anglican-cemetery.jpg

4" x 5" Astia 100F, Schneider Super-Symmar HM 150/5.6, 1/15 s @ f/22

[Frank Petronio]

You can keep yourself up nights worrying about whether your scans are good enough, to the point where you say, "Screw it, it's not worth doing if I can't drum scan it". But if you follow that path you'll talk yourself out of bothering to shoot anything. You'll save a lot of time/money but you'll no longer be a photographer...

Or you can accept that the Epson's are pretty amazing little machines that are a tremendous value. Look at how far we've come from ten years ago, when many of us paid a lot of flim-flanning half-ass drum scanner operators to give us scans that weren't nearly as good as what we can do on our desktops now.

That's the nice thing about shooting film, you can always go back and rescan it later with something better, plus your editing skills and tools will (hopefully) improve with experience. It shouldn't paralyze you from using an Epson now.

And remember, commercial printing is no longer the most critical use of your scans. Making a file for 11x17 266dpi repro (trade magazine) is well within an Epson's capabilities. It's the fine art photographers, producing large, high-quality inkjets, who are the most critical users nowadays. Those are the people who need the best scans.

[sanking]

IME Imacons get out of whack very easily-they need to be serviced every year and after any move.

Sharpening is a huge issue in comparisons, especially if Imacons are involved because a zero setting does not mean no sharpening (I can't remember the number but somethingg like -20 is actually 0 sharpening).



Kirk,

Is there any particular reason Imacons go out of whack easily and need to be serviced frequently?

I agree that experience is a big issue in making comparisons. In my own comparisons I always leave off sharpening completely, though as you suggest it is not always obvious how to do so and operator knowledge of the specific machine is very important.

However, sharpening does not change very much the results of resolution tests with targets. Regardless of what you do with sharpening once you get the contrast set sharpening is not a big factor.

On the other hand, knowing how to get the best overall scan out of any given scanner depends a great deal on experience with the specific scanner. For example, I used to believe that scanning at optical resolution of 3175 spi would give me the best scan with my Eversmart Pro. However, this scanner actually has resolution of 3175 spi X 8200 spi. The first figure is determined by the width of the sensor, number of pixel points, and distance from the CCD to the material being scanned, the second by the number of pixel points picked up in the travel of the CCD over the material. Now it happens that the lens of the Eversmart Pro is actually able to transfer a high percentage of the 8200 spi figure to the CCD so this is real resolution in one direction. That means that when you scan at 8200 spi the software has to interpolate in only one direction, and the result is a much better scan than if you one were to just scan at 3175 spi.

Unfortunately this kind of capability is unique to very high end flatbeds with outstanding lenses. Your consumer Epson is not even capable of recording all of the pixel points across the width of the sensor.

I mention this primarily to emphasize a point you made, and that is the fact that experience with the equipment is very important to getting the best results from any scan.

Attached you will find a target that illustrates the greater resolution in the direction of the travel of the CCD. The optical resolution, if you evaluate both the horizonal and vertical pairs, is about 70 lp/mm, but if you evaluate only the vertical bars it is well over 120 lp/mm.

Sandy King

[Kirk Gittings]

Is there any particular reason Imacons go out of whack easily and need to be serviced frequently?


Sandy, Do I know exactly why? Not really. I've never owned one just used them. I just know that SAIC has theirs calibrated and cleaned once a year and always after a move. Towards the end of that year you can tell the difference. In the summer when I use them I always wait till after the tech makes his visit. Dust on the sensor is one thing that has been explained to me and focus calibration another. Now these are student/faculty used machines, but there are 4 of them plus pro flatbeds and numerous Epson 750s to share the load.

[timparkin]

You know, Tim, this is exactly the opposite way that I would approach this whole issue. ... ..
<snip>
The first thing I would do would be to go to someone whose work I respect and ask them to recommend someone. You see, the `operator' (if one must belittle someone with such a title) is primary and the machine they have chosen to work with just follows.

Once one is in front of someone known for high quality work, then one can just have a chat about how they approach their work in general and what they think may be best approach with specific transparencies.


Hi Richard,

I agree that if you have access to someone who you trust completely in there knowledge of the best drum scanners and operators then you will get a good recommendation. The problem is, how do you assess who to trust without having knowledge enough about drum scanning to do the assessment.

I was responding to Kirks comments about using 'factory calibrated' machines and 'expert operators'. This scenario is so outside the capabilities of most people that even if the best scanner was chosen, it might be a 'formula one' scanner that falls apart without minute by minute maintenance and calibration. I'd rather have a scanner that performs very highly but doesn't need much expert knowledge..

Out of the 'top photographers' I know, only one or two have drum scans made of their work so the mass opinion of people would be that a drum scan isn't worth it. Most get scans made on 'workshop' Imacons whose provenance and maintenance is poor at best and the results can be little better than a well set up Epson.

However, looking at the comparisons on the Internet, the vast majority of comparisons show drum scanners offer a huge benefit, and with the best benefits in colour, contrast, tone, etc. Without the internet I would be paying for Imacon scans. With the Internet I've bought my own drum scanner. Hope you understand where I'm coming from here?

I understand that your personal path probably led you to a similar result - it might well not have though, depending on who you talked to..

Regards

Tim

p.s. A few of those top photographers I was talking about scan using Epsons and the work looks absolutely fine to customers at prints up to 20x24... We have to remember that past a certain point, the only people we'll be impressing is ourselves and other photographers.

[Richard Mahoney]

Gidday Tim,

Thanks for getting back and kicking off this thread. We may all be going on about the old Epsons again but I for one have been picking up a lot along the way ...



I agree that if you have access to someone who you trust completely in there knowledge of the best drum scanners and operators then you will get a good recommendation. The problem is, how do you assess who to trust without having knowledge enough about drum scanning to do the assessment.

New Zealand is a small place and if someone isn't actually related to you, chances are they know your relations. This can help if one wants to find someone who can do something. One just has to start asking around ... In my case the main drive came from dissatisfaction with blocked up shadow detail. Shooting transparencies outside is limiting enough but finding that one is loosing significant detail due to the scanning is ... well ... its enough to make one ... ;)

So I just started asking around for someone who could get into the shadows and it turned out that there wasn't really any choice: it was either taking one's work to the fellow down the road or accepting that one wasn't ever going to be able to adequately print one's images, period. As to trust, well I don't know much about scanning, but I still trust the fellow I was sent to. Why? Well a number of things, but perhaps most importantly -- I've seen his work. Everything else to the side, that's enough for me.



I was responding to Kirks comments about using 'factory calibrated' machines and 'expert operators'. This scenario is so outside the capabilities of most people that even if the best scanner was chosen, it might be a 'formula one' scanner that falls apart without minute by minute maintenance and calibration. I'd rather have a scanner that performs very highly but doesn't need much expert knowledge..

All I can say to this is that perhaps they don't make scanners as they used to. From what I can gather, the Celsis has been used constantly for close to 15 years, nothing significant has ever gone wrong with it and it's showing no signs of packing it in.




Out of the 'top photographers' I know, only one or two have drum scans made of their work so the mass opinion of people would be that a drum scan isn't worth it. Most get scans made on 'workshop' Imacons whose provenance and maintenance is poor at best and the results can be little better than a well set up Epson.

This is sad to hear. Cost to the side, why would one willingly cripple one's work with a poor quality scan which couldn't lead to anything but a poor reproduction? In addition to photographic scans the crowd I deal with also reproduces paintings and so on: 4x5 transparency --> drum scan --> Giclee inkjet onto Hahnemühle German Etching. I imagine that a significant falling away in quality from the original to the reproduction is considered completely unacceptable. For this sort of attention I wouldn't consider NZD 150 for an optical A1 scan (USD 110 / EUR 84 / GBP 70) `not worth it'.



However, looking at the comparisons on the Internet, the vast majority of comparisons show drum scanners offer a huge benefit, and with the best benefits in colour, contrast, tone, etc. Without the internet I would be paying for Imacon scans. With the Internet I've bought my own drum scanner. Hope you understand where I'm coming from here?.

Absolutely, and other things being equal, I'd be very tempted to do likewise. Best of luck with the beast.


Kind regards,

Richard

[Joerg Krusche]

Sandy,

great you attached the USAF test chart .. about the only scanner where the manufacturer does give realistic .. even conservative resolution specs .. that is great !

Now my question .. in order to obtain that very high resolution on the axis of the moving bed/stepper motor (?) you enter as desired resolution 8200 SPI .. equivalent to ca. 2700% at 12 dpi/mm output .. ? .. I just want to make sure I understand what you say.

File size then is according to entered resolution ?

I understand that you want to show how to go to the extremes of the Creo..

best regards,

Joerg

[sanking]

Sandy,

great you attached the USAF test chart .. about the only scanner where the manufacturer does give realistic .. even conservative resolution specs .. that is great !

Now my question .. in order to obtain that very high resolution on the axis of the moving bed/stepper motor (?) you enter as desired resolution 8200 SPI .. equivalent to ca. 2700% at 12 dpi/mm output .. ? .. I just want to make sure I understand what you say.

File size then is according to entered resolution ?

I understand that you want to show how to go to the extremes of the Creo..

best regards,

Joerg

Hi Joerg,

Yes, I entered 8200 spi resolution in making the scan. That means that the scanner picks up 8200 pixel points in the direction of movement of the CCD, and 3175 spi across the width of the CCD.

Now, I was a little generous in the assertion that the lens of the Eversmart was capable of optically capturing all 8200 of those points. If you look at the chart carefully you will see that the last pair that can be read in one direction is either 6-6 or 7-1. 6-6 would give resolution of 115 lp/mm (or effective resolution of 5800 spi) while 7-1 indicates real resolution of 120 lp/mm (or effective 6100 spi).

Course, I might have been able to get more with a better scan, but not bad at all as is I think.

I would of course never scan anything larger than a 6X7 cm format negative at 8200 spi as the resulting file size would break the bank. However, if you position the 6X7 cm negative in the right direction, i.e. with the long direction in the direction of travel of the CCD, the scans are not long at all, less than 15 minutes. That is because the width of the CCD covers the entire negative in one direction. It took me a while to learn to take advantage of this to shorten scan times, but there it is.

Sandy King

[Joerg Krusche]

Sandy,

thank you for your explanation .. I learned quite a bit from you !

best,

Joerg

[Brian Ellis]

A useful comparison, but as always only compares those particular scanners/operators. The problem is we never know how talented the operator is or in what shape the scanner is?

FWIW, I own an Epson 750, have access to the latest Imacon, and use a service for drum scans. Why the difference? Why use them all?

Anything quick and dirty I do at home on the Epson. I have had this Epson since they first came out and have experimented with it wet etc. ad nauseum. I find them useful but pretty low in quality-no matter how you use them. If I want to do a quality small print and am short of cash, I go for the Imacon because I can do them for free in the summers or pay a friend locally $15 anytime. IME I can get a better scan from the Imacon than I can ever get from my Epson. that si either me as an operator on the SAIC Imacons or a good local guy on his. If I am doing a print larger than 16x20 (or ever expect to need a print larger than that), I get a drum scan which run me about $150 including shipping. Currently I am using Lenny Eiger, on an Aztec Premium, for drum scans.

I have paid plenty of money over the years for commercial flatbed and drum scans that were not much better than the Epson. Its easier to find a good shrink than a good scanner operator (and about the same price per hour!).

No samples to show, but tons of personal comparisons along the way-I have tried every which way to skin this scanning cat and the above are my conclusions based on my requirements to do exhibition quality B&W from 4x5 over the last 6 years. i would love to think I could get more out of my Epson but I just haven't seen it-the difference even between it and an Imacon is tactile and the difference to a good drum scan is.....substantial. In the abyss between the Epson and a good drum scan I find the Imacon a notch above the Epson but still a few notches below a good drum scan.

Does the type of photograph never enter into the equation, i.e. no matter what the image there's always this "abyss" between the Epson and a good drum scan?

I ask because in my limited experience with drum scans the only place I've seen any real noticeable difference between them and my Epson 4990, with prints in the 16x20 range or smaller, has been in darkest shadow areas that contain detail. With the Epson I lose some or all that detail, with the drum scans I didn't.

But there's been no other noticeable difference and unless that shadow detail is critical to the image, I don't care. It isn't worth it to me to pay $100 or more for a scan that's "better" than I can get from my Epson if the area in which it's better is unimportant to the image. But if you see this "abyss" throughout the scan, i.e. in all areas of the image, then maybe the person who did my drum scans (not Lenny) didn't know what he was doing.

[Brian Ellis]

I had a friend's Imacon Flextight Precision IV here at the house for a few months last year and compared several medium format negatives with my Epson V700 and Eversmart Pro. I used the maximum optical resolution of each scanner, 3200 spi for the Precision IV, 3175 spi for the Eversmart, and 6400 spi with the Epson V700. I then compared results on the screen.

The Imacon scan was much better than the scan with the Epson V700, but not quite as good as the scan made with the Eversmart.

I also did the same comparison with a 4X5 negative, again using the maximum optical resolution of the three scanners, 2040 spi for the Imacon, 6400 spi for the Epson V700, and 3175 spi for the Eversmart. In this case the Epson scan was a tad better than the one made with the Precision IV, but the Eversmart Pro scan was vastly superior to both of the others.

If one were scanning 35mm negatives there is no doubt in my mind that the Precision IV at about 5400 spi would give a much superior scan to the V700 and Eversmart Pro, but its maximum resolution of 2040 spi puts it in the marginal category IMO for 4X5 film.

I was pretty disappointed for my friend with the results since he already had an Epson V700 and a Nikon LS-9000 and IMO the Imacon Flextight IV was not as good as the V700 for 4X5 film, and not as good as the LS-9000 for medium format film.

Sandy King

My 4990 stinks with anything smaller than 4x5. So except to run a few tests I never used it with my 6x7 negatives (back before digital replaced my 6x7 system and mooted the whole point). And I never even bothered trying 35mm with it, I figured that if I couldn't make a decent sized good print from a 6x7 scanned negative I certainly couldn't from 35mm.

[Peter De Smidt]

<snip>

I ask because in my limited experience with drum scans the only place I've seen any real noticeable difference between them and my Epson 4990, with prints in the 16x20 range or smaller, has been in darkest shadow areas that contain detail. With the Epson I lose some or all that detail, with the drum scans I didn't.

<snip>.

And if you take that into account when exposing and developing your film, you should be able to get as much shadow detail as you'd like in your images, even using an Epson scan.

[Brian Ellis]

And if you take that into account when exposing and developing your film, you should be able to get as much shadow detail as you'd like in your images, even using an Epson scan.

Oh I can, absolutely. I can grossly overexpose (e.g. place the darkest important shadows on Zone V) to take the limitations of the scanner into account. That certainly will retain all detail in the darkest shadows. But then the brighter midtones (Zone VI) and highlights (Zone VII and VIII) become so bright and blown out that even N-2 development can't fix them. And areas that should have been shadows and darker midtones (e.g. Zone III and IV) cease to be shadows and become bright midtones in the negative instead. All of which results in a negative that's difficult if not impossible to print well.

[Kirk Gittings]

Does the type of photograph never enter into the equation, i.e. no matter what the image there's always this "abyss" between the Epson and a good drum scan?

I ask because in my limited experience with drum scans the only place I've seen any real noticeable difference between them and my Epson 4990, with prints in the 16x20 range or smaller, has been in darkest shadow areas that contain detail. With the Epson I lose some or all that detail, with the drum scans I didn't.

But there's been no other noticeable difference and unless that shadow detail is critical to the image, I don't care. It isn't worth it to me to pay $100 or more for a scan that's "better" than I can get from my Epson if the area in which it's better is unimportant to the image. But if you see this "abyss" throughout the scan, i.e. in all areas of the image, then maybe the person who did my drum scans (not Lenny) didn't know what he was doing.

Frankly most of my serious efforts with scanners and scans revolve around my B&W work. Of course I have, prior to digital, had thousands of my commercial color transparencies scanned for magazines, books, posters etc. but I was rarely involved in that. For a couple of years, just prior to digital, I did do allot of my own scanning for commercial publication because I was behind the times and everyone just wanted files. I never found that a scanner like an Epson (in my hands) could deliver what a commercial drum scan could deliver on transparencies for publication. They suffered in the shadow areas, highlight separation, color fidelity etc. I never had any complaints from clients, but I could see the difference.

It is certainly true IME that the major differences of home flatbeds and pro drum scans lies in resolution and in the renderings of the extremes of the tonal range. My 750 cannot pull detail out of deep shadows anywhere close to a good drum scan and I believe at the other end of the scale in the highlights that there is some flaring that takes place in home flatbeds that obscures fine detail in dense highlights.


And if you take that into account when exposing and developing your film, you should be able to get as much shadow detail as you'd like in your images, even using an Epson scan. Peter De Smidt

I believe this is part of the key to maximizing tonal control from the Epson scanners. Shoot and develop for the limitations of the scanner. With a drum scan I just shoot normally using basic ZS controls, but to maximize a scan on an Epson I would place the low values higher and develop the highlights lower.

BTW, I visited the printing service I use for large digital prints yesterday, who was working on a print of mine that I had never printed in silver or ink larger than 16x20 before (TRI-X in HC-110). I have never owned a printer that would print over 17" wide. This is a print 24x33-actually the largest art print I have ever made and is for an upcoming museum show. It was scanned by me on a well maintained Imacon at SAIC and looked superb at that size. The grain (or grain clumping I don't suppose that an Imacon can actually resolve grain) was sharp and tight and looked great even on close inspection with my reading glasses, requiring no additional sharpening than the original Imacon capture sharpening (I don't remember the setting). This was never a "look" that I could get on large prints with my Epson. As a backup I had Lenny do a drum scan of the same image in case it didn't hold up at that size. Lenny's scan pulled more deep shadow detail out of the negative, but I didn't need it-not a wasted expense anyway-I am archiving quality drum scans of all my best images from the past 30 years.

http://sitemanager.sitewelder.com/users/KirkGittings2359/images/KirkGittings2359674332.jpg

I know Lenny's drum scan has more deep shadow detail than the Imacon, but I tend to print this pretty deep anyway. There is a fair amount of work in the file and with a tight deadline looming, I was not wanting to start over with a new scan unless I had to. But in terms of resolution the Imacon could handle this size and on this image the tones were fine too.

[sanking]

This is a print 24x33-actually the largest art print I have ever made and is for an upcoming museum show. It was scanned by me on a well maintained Imacon at SAIC and looked superb at that size. The grain (or grain clumping I don't suppose that an Imacon can actually resolve grain) was sharp and tight and looked great even on close inspection with my reading glasses, requiring no additional sharpening than the original Imacon capture sharpening (I don't remember the setting). This was never a "look" that I could get on large prints with my Epson. As a backup I had Lenny do a drum scan of the same image in case it didn't hold up at that size. Lenny's scan pulled more deep shadow detail out of the negative, but I didn't need it-not a wasted expense anyway-I am archiving quality drum scans of all my best images from the past 30 years.



Kirk,

Regardless of how it was scanned that is a very nice image. The balance between the open window and the light on the wings of the angel is striking.

Sandy King

[Kirk Gittings]

Thanks Sandy.

[Peter De Smidt]

Oh I can, absolutely. I can grossly overexpose (e.g. place the darkest important shadows on Zone V) to take the limitations of the scanner into account. That certainly will retain all detail in the darkest shadows. But then the brighter midtones (Zone VI) and highlights (Zone VII and VIII) become so bright and blown out that even N-2 development can't fix them. And areas that should have been shadows and darker midtones (e.g. Zone III and IV) cease to be shadows and become bright midtones in the negative instead. All of which results in a negative that's difficult if not impossible to print well.

Wow. My guess is that there's an issue with your scanner or technique. I've not had an Epson flatbed, but I did own a Canon 9950F, another consumer scanner. With it I could get better info from the toe of a film than I could with wet printing. I could safely double my exposure index with no trouble. My guess is that you're clipping shadow detail somehow. Consumer scanners, in my experience, are challenged by high densities but not low ones.

[Lenny Eiger]

As a backup I had Lenny do a drum scan of the same image in case it didn't hold up at that size. Lenny's scan pulled more deep shadow detail out of the negative, but I didn't need it-not a wasted expense anyway-I am archiving quality drum scans of all my best images from the past 30 years.


Thanks, Kirk. I have always held that certain printing styles needed different kinds of scanning. If you peruse Kirk's site, a worthwhile investment of time, you will see that he has a fairly contrasty printing style. He may need some good scanning for his highlights... depending on the neg, but he doesn't need it for the shadows, which he treats differently.

It will depend on what you how you like to print whether or not you need the extra detail....

Lenny

[Kirk Gittings]

Thanks, Kirk. I have always held that certain printing styles needed different kinds of scanning. If you peruse Kirk's site, a worthwhile investment of time, you will see that he has a fairly contrasty printing style. He may need some good scanning for his highlights... depending on the neg, but he doesn't need it for the shadows, which he treats differently.

It will depend on what you how you like to print whether or not you need the extra detail....

Lenny

Like you supply me, I prefer to start with a full information scan. From there I can print down the blacks if I want to, but at least I have that choice from a good scan!

[onnect17]

First, Howtek is a great scanner, no doubt about it. But my Eversmart Pro is just as good with B&W and color negatives as the Howtek 45. And more versatile since you can scan a negative up to 12X17" with full resolution, and also reflective material of the same size. For color transparency material the Howtek will do a better job.

Color negative film of the same ASA usually has higher resolution than slide film. There are exceptions but my experience is that this is usually the case. Also, the dynamic range of color negative film is much longer than that of slide film, at least three or four stops and perhaps much longer if you scan.

I have personally exposed Kodak Portra 160 in conditions where the SBR was greater than 15 and was able to capture the entire range with a scan. Typically the curve of color negative film levels off at a certain point and all highlight values beyond this point are compressed on the same line. You could not print this in the wet darkroom but if you scan it is possible to adjust the curve to increase contrast in this area. Pretty remarkable what you can achieve, especially if you adjust the colors in the conversion from RGB to B&W in PS. This kind of range is not even remotely possible with transparency film.

Have you tried Ektar 100? Way more dynamic range than any slide film, virtually grainless in any size print, and sharpness is way up there.

Sandy King

Hi Sandy,
Thanks for recommending the Ektar. I got a couple of shots with the 6x17 of the foliage in NH and I'm really happy with the results (even with the holder being some kind of a POS, more than 1mm off).
I printed one of the negatives at 40"x120" on canvas and looks great.
Regarding the Imacon vs Howtek, I can compare only the 343 to the d4000. I don't think the wet mounting is worth it for B&W in the Howtek. Now, IMHO the howtek have the advantage of a cleaner light source and individual paths for R, G, and B; so the color quality is a lot better in the Howtek.
My two cents.
Armando