• Welcome to Photrio!
    Registration is fast and free. Join today to unlock search, see fewer ads, and access all forum features.
    Click here to sign up

Pros and Cons of ICE

I've only ever been happy with ICE on my Nikon scanners, slows my Epson down to buggery and introduces some occasional funky artifacts in the fine detail.

If I could get something like the LS-4000 which scanned 4x5 I'd be rapped!

[note: this meaning

rap 2 (rp)
tr.v. rapt or rapped (rpt), rapping, raps Archaic
1. past participle rapt To enchant or seize with rapture.
2. To snatch.
]
 
After I asked the question earlier this morning I scanned some negatives from the film I processed last night. I decided to try the "ICE" and see what happened. As I was looking over the image removing dust specs I notice these crazy artifacts! Then I realized why "ICE" may not be so good???
 
The ICE on my V500 is pretty much useless. Clearly I don't know how to use it...
 
I don't use ICE with my Epson scanners, neither do I use the Dust Removal feature. I've found ICE reduces sharpness without significantly removing dust or scratches in any way, and the Dust Removal feature adds strange artifacts to my scans -- they look like slivers of glass . . . or ice, come to think of it.
 
There's no way around digi-spotting if you want high quality scans...make sure you get the best, cleanest scan possible whether that means dust-brushes, cleaning glass with microfibre cloth, wet-mounting...whatever works for you.
 
Folks

I suspect that the problems lie in the difficulty of mapping the IR channel to the other channels. Epson clearly has a problem with this even mapping the R G and B which are on the same physical scanhead. If you look at my article here you can see that at a pixel level the Epson can't get it exactly right.

As Nikon scanners don't seem to have this problem (neither the mapping problems or the ice artifacts) I suspect that this is the source of the problem.
 
I don't use ICE or dust removal on my epson 4990. Tried once and it was soooo slow. If I do a careful job of dusting off the neg and cleaning the scanner glass it's not much effort to spot the digital image.
 
I have used digital dust etc. removal (briefly) on both Canon and Epson scanners. As long as you don't want to print too large, it sort of works - the artifacts it creates are too small to be easily seen.
I scan 5x4 negs. and print them BIG - up to 8x and I have never founf any form of digital blemish removal to be satisfactory.
I now scan as clean as I can get then (after contrast, brightness and sharpness treatments) choose 'view actual pixels' and 'healing tool' in photoshop and spend at least a couple of hours spotting each image. After all, you only have to do it once for each image.
 
How good ICE is seems to depend on the implementer. ICE with epson scan is moderately ok, with SilverFast is kinda screwy, and with Vuescan is about the best.

The reality is the ICE can only detect the dust/scratches. It's the software that repairs the damage. Like everything else, software can be good or bad. Autopilot can only do so much. Ultimately it's up to you if you want to do the repair or the machine.

I would much rather that ICE just show me where the damage is and let me do the healing, as I can be more intuative about it.
 
Hi

How good ICE is seems to depend on the implementer. ICE with epson scan is moderately ok, with SilverFast is kinda screwy, and with Vuescan is about the best.

ok ... interesting!

I now have another reason to retest vuescan
 
How does one enable ICE with VueScan?

Filters tab, select box called infrared clean. If you can't see it, your selected scanner does not support it. I always use heavy, as it is much better than SF or Epson Scan on artifacts.
 
I have used FARE with my Canon 9950F flatbed scanner with quite some success for scanning prints (I don't use the scanner for negatives any more due to its ridiculous low true resolution - 1200 ppi - despite the blatantly false claims of Canon of 4800 ppi). FARE is Canons implementation of using an IR channel to combat dust.

The FARE feature seems to be doing a pretty fair job when set in its LOW setting. None of the "blurring" effects I have seen reported for Epson scanners. It only targets the dust spots, sometimes leaving a barely perceptible artefact, but in LOW setting, generally not an issue.

I do more or less RAW scans, with sharpening and grain reduction set to NONE.

Marco
 
I would much rather that ICE just show me where the damage is and let me do the healing, as I can be more intuative about it.
I don't know about the other scanning software, but vuescan allows you to save the scan with 4 color channels, the fourth being the infrared channel which supposedly shows the scratches. If your photo editing software can import and display such an image, you're set.
 
I no longer use ICE with my Epson as it is too darn slow and inconsistent.

I'm now using the new Silverfast Ai Studio with iSRD, which is quicker and without the artifacts. I scan the file as 64-bit HDRi, which scan both the color information as well as the infrared channel, and use Silverfast HDR Studio to process the scan. in HDR studio, I select iSRD, and either the default setting of 12 or if the negative is especially dirty, as high as 18.

ICE on my Epson V750 adds up to 90 minutes on some 6x7 scans. Using iSRD, it takes a matter of minutes.
 
I have a Nikon 9000 which I did not like the results with ICE. In fact I have never liked ICE on any scanner I have used. I am now scanning Kodachrome using SilverFast's Kodachrome workflow and have discovered that I do not have to do any digital spotting. Go figure. Speed is not bad.
 
Interestingly enough, Ice on my coolscanV seems to reduce grain effects on very fine grain films such as Astia and Velvia, without losing much detail. In fact after sharpening with Focus Magic, I get the same or better detail with Ice than without it.
Note that this is only for very fine grain positive film.
If I try the same with Ektar, I have to set Ice on lowest setting or none at all as it seems to affect the result more.
One more argument to scan colour negatives in positive mode with Nikonscan?