Epson V700 Digital ice, not too good

[kevs-2323668]

Just started first serious testing of this scanner and it seems the ice adds weird moire type patters and some strange hues. How disappointing. Anyone else notice this?
And the ice on my 35mm Nikon scanner , by comparison, is so good.

The dust feature is not so great either -- in sense it changes the image too much.

I'm starting with Epson Scan, and was going to test the same images with Silverfast and maybe Vuescan if they have a demo to try. Do you think things will be better with those?

Otherwise, the scans seem to look pretty good on the monitor so far. The dust/ is not to horrific. These are color chromes. But I fear if I were to shoot BW negative.

Also curious: on Epson scan, anyway to to Not have the check box default to check unsharp mask?
Any way to get the scan into a color space I want from the get go?

[Frank Petronio]

Epson Scan > Advanced > look around, you can save it in ProPhoto RGB or Adobe RGB.

Two schools of thought. Some say adjust the prescan as far as you can even with the lousy previews that Epson Scan, Silverfast, or VueScan provide you. Others say to just scan the thing with all the "help" off and treat it as a raw file. I find the latter better and make my global adjustments in ACR/Lightroom, then edit details in Photoshop. Sometimes the prescan on auto is so far off that I do deviate from my advice and hit the grey/black/white eyedroppers but I don't go near any of the other crapware features offered.

I have not used Silverfast in recent times but I don't think you get a particularly better quality image from any of the scanning apps. To me Epson Scan is the most straightforward. In general all the scanning apps have that warm, 1996-vintage software feeling... i.e. they all suck. I do use VueScan on my 35mm film scanner but that's because the manufacturer's software is even more abysmal.

If you are not sure of what you are doing, a good way to figure out what works is to do a bunch of lower res fast scans with different settings and inspect them immediately for feedback. If you can't get it right after 4-5 scans then, umm, I don't know, ask someone else!

Also a big help is to give Lenny Eiger a contrasty, full-range scan with fine gradations and maybe some flesh tones - and then see what he does and you'll have a basis of comparison for knowing what you can get away with on the Epson and what kind of quality you are sacrificing/winning from a drum scan.

[Jon Shiu]

Never had those problems with ICE myself. Be aware that ICE does not work with Kodachrome, nor black and white film.

Jon

[dave_whatever]

In my experience of epson 4990 and v700 with epsonscan, silverfast and vuescan, I can say that I've alway found the ICE to be pretty poor compared to proper film scanners. I often have to clone out the "repairs" that the ICE has made. Doesn't seem to have a problem finding the dust, just that the algorythm for filling in the repair seems to be lacking.

[RedSun]

Film scanners have ICE too. I own a Minolta 5400 and it works great for 35mm.

[kevs-2323668]

Thanks guys, cool. Frank, yeah, I never do adjustments in scan software. In fact, how do you get that dang check-box not to default unsharped masked in the Epson to not check it?.

I'm using this for web images. I'll probably call Lenny or someone for drums for printing though.

I'm leaning towards not using the ice though I use the ice with no issue on the Nikon film scanner.

Surprisingly, the Epson can make a huge scan, 1GB, though I'm sure it wont blow up too well compared to a drum. Pity ice does not work. It really does not work -- and I'm scanning color chromes. Fortunately the dust is not too bad, but I'm fearful for BW negs down the road.

[rdenney]

I find the infrared pass (which is what algorithms like ICE use) is often slightly misregistered from the main scanning pass, so the the correction doesn't perfectly align with the dust image. I've given up on it with the Epson and have become as meticulous as possible at preventing dust in the first place.

It works fine with my Nikon, but the infrared channel on that is read in the same pass.

Rick "who still spends less time on spotting than with darkroom prints" Denney

[Jon Shiu]

I've had good luck scanning Velvia E6 film with ICE. Due to IR properties, Black and White film and Kodachrome do not work with ICE. However, C41 black and white should work.

If you are scanning chromes, make sure you move the black point back a bit so it is not clipped.

Jon

[kevs-2323668]

thanks Rick / Jon.

Rick what do you do to spot, does it help that much? Does does collect in the machine while you are scanning? Does BW neg seem to have 100x more dust on results that color and if so why?

Jon, I don't screw with black points or anything in scanning software. Rick do move black point in the software?

ps Rick, Epson brings it into it's own profile and then I convert. I think Adobe 1998 is fine, don't think I need PHoto rgb for web images. Ditto 8 bit seems to be ok.

pss what deal with Silverfast? Clunky as hell. I needed to put in CD to launch and today it cannot find the scanner, and there is launcher and then the software. Can't get to first base with it.

[Sevo]

thanks Rick / Jon.
Rick what do you do to spot, does it help that much? Does does collect in the machine while you are scanning? Does BW neg seem to have 100x more dust on results that color and if so why?

No. IR de-spotting works in the IR band, to which the layers of chromogenic film are transparent, while dust and other solid particles block IR. The silver of black and white film is solid and IR blocking, so the entire image of (silver) black and white films shows up in the Digital ICE IR (dirt) channel rather than the dust only. Better implementations detect that problem and refuse to work, while older or worse ones remove the entire image...

Chromogenic black and white films (XP2 etc.) don't have that issue - but AFAIK none of them are available in sheet sizes.

[kevs-2323668]

Did not understand that at all SEvo, My question is if I don't use ICE, and I plan not to, BW neg would show more dust than color films / yes / no?

I cannot fathom Vue or Silverfast -- seems it will take hours to learn them, so I'm going with your opinion Frank and sticking with Epson scan -- (Rick if you think it's worth the battle to learn Vuescan let me know -- I just tried it and it preview the holders not the film -- very clumsy software.. so far)

[Sevo]

Did not understand that at all SEvo, My question is if I don't use ICE, and I plan not to, BW neg would show more dust than color films / yes / no?


That will entirely depend on your lab's cleanliness and proper exposure -exposure and processing discipline in colour tends to be higher, but if you are as accurate in your black and white handling, there is no difference.

[kevs-2323668]

Final note: it really is a farce. Epson should remove the digital ice. You can't have a feature that introduces artifacts worse than the ones it's meant to correct!

[rdenney]

Final note: it really is a farce. Epson should remove the digital ice. You can't have a feature that introduces artifacts worse than the ones it's meant to correct!

No argument from me.

I use the rubber stamp tool to spot dust. I just have to do it once. Do it after sharpening, though--some dust becomes more visible when sharpened.

Sevo means that some films look like dust to an ICE-type algorithm, including black-and-white films. ICE can't be used for B&W in any case. More dust on black and white? Not in my experience. Scanning (as was the case with enlarging) requires a cult of cleanliness. I inspect negatives in the holder under a strong light, viewed from the side such that the dust is really apparent, and remove it using a bellows blower (made for removing dust from laboratory equipment) and a camel-hair brush. I spend sometimes minutes dealing with dust before the negative goes into the (clean) scanner.

Smaller formats get enlarged more, and thus smaller specs of dust loom larger on the prints. That makes ICE and other IR-channel tools more valuable for smaller films scanners. I do not find that my life is ruled by dust with the V750.

Rick "prevention is better than cure" Denney

[Frank Petronio]

If I take 30 seconds with a Rocket Blower then I have 5 minutes of spotting. If I do less first, I do more later.

[kevs-2323668]

Thanks guys. Rick is the bellows blower better than air can?

[Sevo]

Final note: it really is a farce. Epson should remove the digital ice. You can't have a feature that introduces artifacts worse than the ones it's meant to correct!

I've never found it that bad - maybe recent versions of their scan software screwed up? Sure, it does not work on (silver) black and white nor on incompletely fixed film, and is not quite up to the Silverfast or Vuescan (nor even the Nikon Scan) implementations of such a filter, but at least the version supplied with my 4490 usually will generally do better than hurried de-spotting with the healing brush.

[kevs-2323668]

Follow up question, important:

My intern and I made about 90 scans with Epson scan. They started out ok, but then wow, they became really blown out, (overexposed highlights) and some really pink. They need to be redone, any ideas on how to prevent this? thanks!!

(and some jpegs I made from magazine articles a couple weeks ago were also too contrasty, so this seems to be an issue with this scanner, just getting an accurate exposure, what to do?)

[kevs-2323668]

PS, did a test, and everything is ok with Epson scan. I let an intern scan for 6 hours and most of his scan are blown out. How did he do this? I have no idea? Any ideas anyone. thanks! I cannot figure out the mistake made -- same settings.