Photo CD vs Desktop Scanner

I have a Nikon Coolscan III (LS-30), which is roughly comparable to an LS-2000. My observation is that PhotoCD has greater dynamic range, which is sometimes useful for images with high contrast. The Coolscan produces less noise, because you may scan an image repeatedly, averaging the results.

So, it seems that the best solution is to have the same image scanned repeatedly by a PhotoCD scanner. :-) In truth, it does not matter much. Using the right software and a bit of knowledge about digital image processing, both alternatives give you excellent results. Without the right software tools, neither alternative will produce good results.

Resolution is hardly a problem with PhotoCD, but handling slides {compared to negs} is a problem. Virtually any consumer desktop scanner is a better bet to scan slides than PhotoCD because of the Kodak unit's problem with dynamic range.

Joe - Not to confuse things, but I'm also trying to make a decision regarding Pro Photo CD vs Scanner. It appears the two responses you received are conflicting - one says CD has better dynamic range, the other says Photo CD scans of "slides" is not ideal because of a problem with dynamic ranger?

To tag on another related question, what about the increased resolution available with Pro Photo CD's versus the particular scanner mentioned? Wouldn't that allow for a much better print, especially in the size range that Joe indicates? I had some slides scanned to base 16 (2048x3072) that I printed at 11x16". Now that gets my dpi down to about 185 at that image size and everything I've read says you should try to stay at about 300dpi to get a first rate print. I've been thinking about trying some base 64 scans (4096x6144)and print at the same size. Would you expect to see noticable a difference in the prints between base 16 and base 64? There's a noticable difference in scan price:>) It would appear that one might need to go to a 4000 dpi scanner to equal Pro Photo CD for prints in the 11x16 size range??

Jim

I have had some none-to-satisfactory experiences with pro-photo CD. While the DPI is theoretically 4000, I have found that in practice, the optics do not keep up with the pixels, and the image is soft. For 35mm, I find that I get just as much real detail from a regular Photo CD, and then doing a bi-cubic resizing to a larger size BEFORE doing any unsharp masking. I have had this experience with two different labs, so I don't think it was operator failure for the Pro-CD.

I have found the better scanners (Nikon, Minolta, Poloroid) to offer better sharpness than either PhotoCD solution. The bit-depth is largely a factor of operator settings, but I can always do better on the Nikon 2000 than what I have gotten on either Pro or regular CD.

Finally, I am not down on the Pro-Photo CDs all the way. Although I am not at all impressed with the results for 35mm, they work well for medium format. Again, my impression is that the optics limit their ability to deal with 35mm, but they work well in larger formats.

Of course, none of the alternatives currently available have the bit depth that I get with drum scans (when made by competent operators). But a bunch of drum scans run into real money, and have the downside of putting the original film through some abuse.

Summary:

- Photo CD - very respectable scans - Good Desktop - somewhat better scans - both sharpness and bit depth - Pro-Photo CD - Little better than Photo CD for 35mm (often worse in my experience) - Drum Scan - Very nice, but pricey

By the way, Nikon has announced a next-generation scanner that claims to have stats close to current drum scanners. I believe that it will be out in April, so it will be interesting to look at.

Best, GP

For 11x17 use, Kodak Pro Photo CD (72 MB) scans are more than adequate, while the regular 18 MB (non-"Pro") are not quite up to the task.

PhotoCD scans are softer than drum scans (I run a DaiNippon SCREEN drum scanner, it does optical unsharp masking on- the-fly as part of the scanning process.) Despite owning a drum, I prefer to use Pro Photo CD scans when scanning from negatives, or if an original chrome is grainy.

PhtoCD scans are not sharpened, they are raw scans - they are not meant to be used for pre-press or inkjet prints "as is." The amount of USM should be controlled by the end-user in Photoshop, either after the import, or during import via the Kodak Acquire module, downloadable from Kodak's website.

For example, using a ProPhotoCD scan we recently produced an 18 x 24 inch poster of a sprint car, there would have been no advantage to using our drum scanner since the only fine detail not picked up was grain. This is to a 175 line screen, using a 300 ppi file.

Any significant detail was easily extracted with USM on the luminance channel, then converted back to CMYK for printing. For your Epson, back to RGB would work as well, and be easier to handle in terms of file size and printer RIPping. So go for it, and if they don't believe you, show them your prints. Or ask to see theirs.

"Although I am not at all impressed with the results for 35mm, they work well for medium format. Again, my impression is that the optics limit their ability to deal with 35mm, but they work well in larger formats."

This sort of logic is becoming pretty prevalent on Photonet, although I am somewhat confused as to why. Why would anyone use a larger format and then compromise on the scan? If you are going to get benefits from using a larger format, you are going to have to input, process, and output the image at the same absolute resolution as the 35mm format. Otherwise just use 35mm and reap all the benefits of cost, ease-of-use, etc. that that format offers.

For the record, although 300 dpi is commonly accepted nowadays, in truth it becomes woefully inadequate when compared to reasonable optically/chemically produced prints. At any print size.

This is slightly off topic, but before you run out and buy a Coolscan 2000, you should know that Nikon just annouced the Coolscan 4000 (see their web-site for details). This means that the 2000 will probably drop to around $1000 once the new scanner hits the market.

hi joe.

just a consideration... while the coolscan is no cheap toy, you will quickly recoup your initial cost versus paying over and over again and through the nose for something that costs kodak pennies. not to mention the versatility you get with the scanner.

myself i haven't been too impressed with the dynamic range of the 2 photo CDs i've had made. but then again, kodak seems to be tripping over its own toes lately with most of their services.

go with the scanner. you'll be glad you did.

I'll post some examples to make my point on this debate. (see attached image).

The top image is a 1:1 crop (with no post scan processing or USM) from a 35 Superia-Reala neg ProPhotoCD scanned at 6144x4096, which is well and beyond the capability of most of the scanners mentioned here (and my optics). Superia-Reala is one of the finest grained films currently on the market, and as you can tell from the scan, the grain is extremely sharp, if not in excess of the film's actual resolution. If anybody in this thread has a home film scanner that can equal this resolution, I'd like to see it posted. This is typical of my results with ProPhotoCD from various labs, although every once in awhile I get a 35mm scan slightly out of focus.

The middle image was from 35mm RG 100, and pulled straight from a PhotoCD with no adjustments to contrast or density. Again, this is typical of my ProPhotoCD results from negatives. Contrast and density are about as perfect as I could expect from a commercial film scan.

The bottom image is from a ProPhotoCD scan from a perfectly exposed, 35mm, Fuji Astia chrome. The image is dark, murky, and the histogram shows poor dynamic range and lack of density information. Again, this is typical of my results from various labs from ProPhotoCD. While some phto.netters have claimed better results from ProPhotoCD from slides, nobody disputes that the Kodak device has poor dynamic range, which is critical for slide scanning and not so critical for neg work. In short, if you want to scan 35mm slides, you should look other than PhotoCD.

Whew! A lot of conflicting advice. A lot of bad advice too!

You need only know two things. Resolution and dmax. In both cases the Nikon LS-2000 is superior - in theory and actual tests. At 2,800 dpi and a dmax of 3.6 CD photos aren't even in the same ball game. Years ago I ran a very exhaustive test using the same slide and a LS-1000 vs Kodak CD. There are quite a few threads in the archives on this matter.

If you are dead serious purchase a new Nikon LS-4000. It resolves 4000 dpi and has a dmax of 4.2. This is superior to most very expensive drum scans!!!!!!! In order of resolution and dmax, from least to best:

1- Flat bed scanner with slide convertor (dmax 2.6 to 3.0)

2- Photo CD (dmax 2.8 to 3.0)

3- Lesser slide scanners (dmax 3.0 to 3.4)

4- Pro Photo CD (dmax 3.0 to 3.4)

5- A few slide scanners such as LS-1000 or LS-2000 (dmax 3.4 to 3.6)

6- Drum scan (dmax 3.6 to 4.1)

7- The new Nikon LS-4000 (dmax 4.2)

8= The new Nikon LS-8000 (because it takes larger chromes)(dmax 4.2)

It resolves 4000 dpi and has a dmax of 4.2. This is superior to most very expensive drum scans!!!!!!!

Nikon's marketing divison can claim what they want, but other CCD scanners (Flextight, LeafScan) have claimed near 4.0 or better Dmax have been trashed pretty bad in side by side comparisons against even mid line drums like Howtek. Drums are also able to operate at their d-range limit and produce clean scans, while CCD based scanners are not happy operating at their electronic limit and begin to produce a lot of noise.

I have owned an LS-2000 for two years. I like it. I have never had a photoCD made so I can't give conflicting opinions :)

The difference I perceive is that with a desktop scanner you do the work but have control over the process, albeit within the limitations of the scanner, software, etc. With PhotoCD you do not do the work but have little control over the process.

This can be important when you have under-exposed a slide and still need to pull all you can out of it. Then, owning your own scanner can come into its own.

Forget Dmax and resolution issues, if the results match the time and money expended, then you have what you want. That's how I feel about the LS-2000.

A word of warning on CCD scanners. Do not believe the manufacturers value for DMAX (or much else). Often it is theoretical and measurements by many suggest that scanners with qouted DMAX 3.6 perform as per scanners with quoted DMAX 3.4. You are best to search the web for review sites and/or try the units out for yourself.

As others have said, the LS2000 is now yesterday's product. As to how good the replacements are, we will have to wait for user reviews.

I have a Microtek 4000T which gives roughly 4000 * 6000 DPI, and noise free scans with one pass suitable for a 11 * 17 print. Many would say that a 2700 DPI scanner such as the LS2000 does not have the resolution for such a large print. Also I found the LS30 had too much noise, and multi-pass scanning took too long (~15 minutes for a 16 pass scan). I think the LS2000 has similar noise characteristic.

The Minolta Dual Scan II is well reviewed and cheap if ~2700 DPI is enough.

Thank you, everyone, for the considerate advice. I will hold out for the Nikon 4000 desktop scanner for all my transparencies. Unless my prints look (upon reasonably close inspection) just like a chemical pro lab, I will not be happy. Now, if I can just figure out how to produce hi quality images using photo shop. It is as much art as science. Time for a new thread! Best to all. Joe C

Wow Scott, even after AutoLevels and AutoContrast that PhotoCD scan of Astia is pretty bad. I couldn't resist trying, and having already wasted my time, might as well post the results. Which is more color accurate?

Sorry, here's the JPEG.

Here's my $.02 worth. I tried the Pro-CD route and was not happy about the dirt on the scans, and out of focus results I got back. I was also paying for a full roll of film even thou I would only have a couple of images per roll worth reprinting. I then purchased a Polaroid 4000, better, but now the dynamic range was not as good. Still not what I wanted. Finally I obtained a full working copy of Silverfast software on the Polaroid. Doing 8 pass scanning of my images I now have increadable range and only have to scan those images I want to reprint. I have total control of the prescan settings for each film type I use and it recongnizes the Color Management I have in place. I use this for prints up to 13x19 on my Epson 1270, anything bigger and it is off to the drum scanner, however I can now include a digital file with the color and sharpness that I want that they can use as a baseline when doing their scans.

Certainly not a solution for everyone, but it does work for me.

-harry

One last tidbit of info I received last evening - news to me but maybe not to others on this forum. The photography group to which I belong had our meeting at, and toured, the facility of Reed Photo-Imaging, a large pro lab here in the Denver metro area. During a conversation with Gary Reed he informed me that Kodak Photo CD's are expected to go away - maybe as soon as one year. He said that Kodak has informed their lab that they will no longer provide service or parts for the Kodak CD scanner (I think they have the 4050) starting early this spring (not exactly sure if he mentioned a date). Essentially, Reed Photo expects to keep cranking out CD's until the machine breaks down, then discontinue it. He said that Kodak MAY be bringing out a replacement system, but it didn't sound like that was really certain yet. So.....it sounds like the Kodak Photo CD beast as we know it today is going to be short lived at best. I guess the only certainty in life is the certainty of change.

Jim

Sorry Bill, there just isn't enough density information there to work with. Never had good slide scans PhotoCD, and doubtfull I ever will.

While I won't exactly miss the Kodak system, I do have to admit Kodak did a good job with the PhotoCD compression format.

I posted an article on the internet; Epson 1640SUP vs PhotoCD, which might help to answer your questions. The Epson isn't nearly as good as the Nikon Coolscan 2000, but it sure produces better-exposed scans than Photo CD. Photo CD has the edge, however, in sharpness. I've found that one cannot rely on technicians to make proper exposures for PhotoCD scans, especially with difficult slides. If you have the time to do your own scanning then get a scanner. I'm sure you'll tend to get better overall results.

I just had to add my $0.02 worth. PhotoCD is good, yet expensive and somewhat overrated. It does take more adjustment than should be necessary for the price. Why the color balance issues? It would seem the easiest thing for them to standardize.

Several photoCD vendors have told me that for 35mm the 6th resolution (18mB) is not a viable improvement over the 5 mB (5th res.) scan. They are sure that the optical resolution capability of the film, scanner, and camera's optics are exceeded. I cannot qualify this with any real world experience, but would like to hear any comments from people who've used the ProPhotoCD highest resolution with 35mm. The increased dynamic range from a proscan would seem to only benefit slide film users. Any thoughts on this?

Has anyone compared the new Kodak Film scanner with a PhotoCD? Does it output to YCC files (.pcd) as well?

I do not work for Dale Labs in Florida, but I have had them process and scan to their proprietary CD format on 15 occasions. They produce a maximum of 18 megs from 35mm negatives or slides. You might send a roll or negative and a roll of slide film in to test them out. I have been pleased with the results. Cheers, Bob

In all of the discussion above, I only see one small reference to a fundamental difference between PhotoCD and other scanning technologies. That is, the encoding format, which in PhotoCD appears to be a properly thought out system of preserving the original film's information by bearing in mind the relationship between the scanning process, the physical properties of film and the limitations inherent in digitisation.

The White Papers available on the PhotoCD web page touch on this, and are a far cry from the usual dross you find on the Kodak web site (i.e. along the lines of "PhotoCD gives you the colors you've come to expect from Kodak...)

(link)

In particular, it refers to encoding values of luminance beyond 100% which initially seems counter-intuitive, but given some time to work through the explanation, this it does make sense. The PhotoCD system allows encodings of RGB values from the range of -0.2 to 2; whereas intuitively it would seem there is no point encoding values outside of 0 to 1.0.

Add in the fact that RGB values & Gammas of screens in the PC world are absurdly undefined and variable (c.f. higher end machines such as Silicon Graphics, even Macintoshes), and I have this deep suspicion that there is more to Photo CD than meets the eye. I have found from my own (admittedly uncontrolled) tests that Fuji frontier slide scans can lose detail at either end of the dynamic range, most frequently in the highlights, whereas PhotoCD scans do rather better.

I am by no means an expert in the arcane details of digital image representation (and it is actually a huge subject the more you go into it - you end up delving into the details of human colour perception to really understand why some systems will be better than others), but there does seem to be a basic difference in design strategy.

Ok, the photo cd scans look strange when you read them straight off the CD without colour correcting them (see my before & after photo attached) but that's presumably due to their information-preserving encoding system. But it does not matter one jot what the unmanipulated scans look like - you might as well complain that you don't like colour negative film because "it's all orange..."

Also, some of the other posts mentioned other scanners which have a built-in Unsharp Mask done "at the time of scanning". I am suspicious of any system that pulls this sort of "sharpening" stunt. I can do that in myself at the final stage of PhotoShopping - it doesn't add any information and in fact destroys some. As a further aside, this is one of the problems comparing different digital still cameras: many of them play around with the raw image internally so that they "look sharper" when they print out in the shop and impress the unsuspecting. Unfortunately, few camera shop assistants grasp this point and stare blankly at you when you mention in-camera image enhancement.

So, this has become a bit of a rave, but the gist of it is that PhotoCD seems to have a solid optical & color-engineering basis, whereas straight RGB encodings seem a bit simple-minded by comparison.

That said, if the Photocd scanners are getting a bit dated in terms of sensors and optics in general, then that may be reason to move on, but resolution by itself is only a small part of the issue. It's no good having a loads of pixels if each pixel is wrong.