Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: The Art and Science of Reproducing Kodachrome  (Read 8911 times)
guyburns
Jr. Member
**
Offline Offline

Posts: 89


« on: September 26, 2012, 08:17:51 AM »
ReplyReply

After 18 months of experimenting, and asking questions on this forum, I have finally put together a PDF about how best to reproduce Kodachrome and other slide types. I purposely did not say "scan" as scanning is only one part of getting a Kodachrome slide into another format.

Cliff Rames was kind enough to proof read the early versions of the PDF and offer suggestions. He had this to say after the final proof, but I was unable to include his technical comments in the PDF so here they are (26 Aug 2012):

Quote
Hi Guy,

I read through your final version and it looks like a winner! I also tested your editing workflow on some of my own Kodachromes, and I must say that it is very well thought out and effective. It makes a lot of sense to me.  

Interestingly, in your Curves 2 color adjustments, page 62, bowing the curves up and down from the mid-point, has the main effect of aligning the Kodachrome characteristic curves. It is effectively the gamma adjustment for each curve needed to neutralize the grayscale, although it's not obvious.

To use my existing Linear scans with your workflow, it was necessary to first convert to a gamma > 1 space. You might want to add a note for those who have linear scans, that it's a necessity to convert to a gamma workspace, since curves and some of the other Photoshop tools are awkward to use on linear data. Of the standard editing spaces. only ProphotoRGB can contain all of the Kodachrome gamut. I can just imagine that your document will be read and applied by other people like me with linear scans, since in some circles (UseNet, filmscanners mailing list, etc.) the conventional wisdom has been that it's best to scan and profile linear. Under your constraints of avoiding color space conversion, scanning into a gamma space is a necessity as your research has shown. Not everyone is operating under the same constraints so I'm just suggesting a way to head-off potential criticism from experienced scan operators who come across your document.

Best regards,

Cliff

The PDF (234 pages) has been set up to allow easy and informative comparison of images by directly overlaying them in an interactive PDF. My findings may surprise some readers. The PDF (an incomplete version) can be downloaded here: http://www.mediafire.com/download.php?qc67n2gkdz3viyp
« Last Edit: September 26, 2012, 10:04:22 PM by guyburns » Logged
AFairley
Sr. Member
****
Offline Offline

Posts: 1165



« Reply #1 on: September 26, 2012, 12:13:57 PM »
ReplyReply

Thank you, Guy, since I have  been scanning favorite Kodachromes from the 1980s, this will be very useful information for me.
Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6945


WWW
« Reply #2 on: September 27, 2012, 09:49:38 AM »
ReplyReply

Hello Guy. I've downloaded your free PDF, and I shall purchase the complete version once I have returned to an internet connection over which I feel safe sending financial information. I have not had time to read the download yet, but obviously a subject of great interest to me, and as I wade into the work you've done, I'll most likely get in touch.

Regards,

Mark
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1163



WWW
« Reply #3 on: September 27, 2012, 11:50:05 AM »
ReplyReply

What intrigues and somewhat confounds me about the pdf is the mention of Kodachrome's bluish cast as an intended design in order to counteract the yellowish/amber light from a tungsten projector which that film type was meant to be viewed by, but accept and expect a flat bed scanner with its fluorescent/LED? light source as a good enough device to counteract/correct for that cast which I totally disagree with.

I didn't read all of the pdf but I thought it would be helpful if they covered an alternate capturing technique of backlighting Kodachrome slides with a similar halogen or tungsten type light source either on the scanner or with a high rez DSLR with a macro lens and see if it gives better color/detail results.

If that film type was designed for projection then it would stand to reason using a similar method for exposure would make color rendering and fidelity easier to attain. As usual I could be wrong on this but I always think it best especially with scanning film to grab all its data by any means necessary at the source instead of relying on software to bring it out.

Scanner's fluorescent light source doesn't seem up to the task IMO. A very interesting and thorough study and valuable tomb of information on such a popular film type, though.
« Last Edit: September 27, 2012, 11:52:49 AM by tlooknbill » Logged
guyburns
Jr. Member
**
Offline Offline

Posts: 89


« Reply #4 on: September 27, 2012, 08:10:41 PM »
ReplyReply

tlooknbill: Yes, you'd think tungsten light would be the best solution. So early in 2011 my mate Baydon rigged up a Kodak Carousel and photographed slides with a digital camera by looking into the projector (see last page of PDF). But even though it sometimes gave the best results due to the camera's ability to capture shadow detail without flare (see the captures of Ref 05, page 99), the technique in general did not give results as good as from dedicated scanners. At least from my testing. And that testing involved sunlight, daylight, various colour temperature settings, and many hours of playing around. I gave it good go, but eventually came to the conclusion that I would not be able to get results, in general, as good as from the V700 or Coolscans.

Baydon is very happy with his method because it is so quick. He can capture 200 slides an hour, and most of them look pretty good. But compared to what can be achieved by other methods, I decided to to pass on the carousel technique.
« Last Edit: September 28, 2012, 10:59:10 PM by guyburns » Logged
WayneLarmon
Jr. Member
**
Offline Offline

Posts: 51


« Reply #5 on: September 27, 2012, 09:05:17 PM »
ReplyReply

Quote
....the technique in general did not give results as good as from dedicated scanners.

Have you seen the results that Peter Krogh (author of "The DAM Book") gets with "camera scanning"? (his phrase) See the camera scanning section of dpBestflow.com (and also note who funded dpBestflow.com.)   Also read the PDF that he prepared.

His tests show "camera scanning" to give slightly better results than a Nikon 8000 scanner at 4000 PPI.  You can download full resolution comparison images from here

But his information isn't directly addressing your concern--he doesn't pay any special attention to Kodachromes, so his results might not be contradicting the results you report.   He does say that often shadows need to be opened up after digitizing slides because they become more contrasty.   But he says "The Fill Light command in Lightroom and Shadow control in Aperture can do a good job with this task." (from the dpBestflow page.)  Again, you might be more exacting about shadow contrast than he is.

I have had good results with camera scanning (Canon 60D, Canon 60mm macro lens, PhotoSolve Extend-a-Slide), but I haven't tested anywhere nearly as exhaustively as you have.  And my comparisons are against a Plustek 7600i and an Epson V600.  I am curious why the Plustek 7200 you tested was so bad.  Luminous Landscape's Mark D Segal reviewed the 7600i against a Nikon 5000ED and also against an Epson V700 and found the 7600i to be a bit worse than the 5000ED and a bit better than the V700.  Unless the 7200 is drastically different from the 7600i, his results don't correlate with yours.

Thanks for the work you have done.  I am recovering from cataract surgery and my eyesight will be fuzzy until my new eyeglasses are ready.  When I can see clearly, I'll examine your paper closely.  Right now, I'm not too good with visual comparisons.  (Apologies for any typos that have crept into this message.)
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6945


WWW
« Reply #6 on: September 27, 2012, 09:30:25 PM »
ReplyReply

Wayne - the 7600i is a technically improved scanner over the 7200i, so what I said about the 7600i does not necessary convey to the 7200i. I never worked with a 7200i. The  8200i recently released in the US market (but has been available elsewhere for some time now) I understand has under-the-hood technical improvements over the 7600i, although its stated technical specifications are the same as those of the 7600i.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
WayneLarmon
Jr. Member
**
Offline Offline

Posts: 51


« Reply #7 on: September 28, 2012, 08:37:57 AM »
ReplyReply

Quote
The  8200i recently released in the US market (but has been available elsewhere for some time now) I understand has under-the-hood technical improvements over the 7600i...

A representative from Plustek posted on DPReview:

Quote
Hi Mark from Plustek here. A couple things....

The OpticFIlm 8xxx series scanners are the same hardware as the OpticFilm 74/7600 scanners. The only difference is SilverFast 8 comes in the box with the 8xxx scanners and you have to download it (free) for the 74/7600 scanners.
http://forums.dpreview.com/forums/read.asp?forum=1002&message=41824865

(Also note what he said about how they compare the upcoming Plustek OpticFilm 120 scanner with drum scanners.)

I am working through my family's negative and slide archives and what I am interested in are solutions that are cost effective and exist as products that are on the market.  I'm trying to make a workflow that I can share with other members of my extended family and anything that involves chasing down equipment that can't be bought brand new is a non-starter.   I really appreciated your review of the 7600i.  I've been very happy with mine.  (For the price.)

But I am still interested in camera scanning, because I think (but can't prove) that my 60D RAW files are giving me better color than what my 7600i yields.  (I know that looking at 60D conversions at 100% often makes feel happy about the tonality, but looking at 7600i conversions at the same magnification makes me sigh.  Generally.)  And camera scanning is a whole lot faster than my 7600i is.  But there isn't much to choose from for mounting slides and negatives in front of a camera/macro lens.   That is currently in production.  I've had good luck with the PhotoSolve Extend-a-slide with slides, but the ergonomics aren't optimal,  And I haven't found the Photosolve negative carrier to be workable for 35mm.  (I was able to modify the 35mm negative carrier to work for 110 negatives though.  110 negatives are problematic with conventional scanners.  I am very happy that I was able to rescue family heirloom images from the 1970s and 80s that otherwise only existed as overexposed prints.)

And I am still going with Peter Krough's contention that camera scanning, when implemented optimally, yields better results than can be achieved with Nikon (or Imacon) scanners.  It is the "when implemented optimally" part that is tricky.

Back to Kodachrome.  Here is a link to an interesting discussion on how Kodachrome was processed, and how this affects the blue cast we get when scanning.  The posts by Alan Marcus, the former Quality Control Manager at Dyancolor's Aurora, Illinois are especially interesting.  He was trained by Kodak on how to process Kodachrome.
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6945


WWW
« Reply #8 on: September 28, 2012, 09:17:11 AM »
ReplyReply

Yes, I had previously seen Mr. Druziak's advice on the 7600/8200 comparison. I'm very much looking forward to see IQ from the 120 after its commercial release. This could be an important event for the scanning community.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
AFairley
Sr. Member
****
Offline Offline

Posts: 1165



« Reply #9 on: September 28, 2012, 11:14:54 AM »
ReplyReply

Re: Camera scanning -- with Kodachromes (64 and 200), I can get results from my 16MP m4/3 OM-D + ZD 50mm macro that easily equal the Coolscan IV in terms of IQ and DR, and are so close to the Coolscan V as to be a non-issue.  And I am not even getting the full use of the 4/3 sensor because of the aspect ration mismatch.  Also, the "out of the box" color with the OM-D matches the chromes more closely (I shoot RAW, so that's the Adobe defaults being applied).  Plus, the camera "scans" don't have the problem with flare/highlight blooming my (mirror-cleaned) Coolscans have that that limits the amount of tonal adjustment you can do without resorting to serious masking.  Note that DXO gives the OM-D a DR rating of 12.something, which helps a lot compared to older DSLRs with less dynamic range.  I would expect even better results with more recent offerings like the D800/D600, which test in the DR 14.x range.

That said, at the end of the day, I do have to do some masking to correct color casts in the highlights and shadows.  I'm curious to read Guy's adustment techiques.  Guy, re your book, I have sent you a PM.
« Last Edit: September 28, 2012, 11:20:24 AM by AFairley » Logged

RFPhotography
Guest
« Reply #10 on: September 29, 2012, 06:06:30 PM »
ReplyReply

Reading this thread for the first time I'm actually surprised that anyone has anything good to say about Plustek scanners. For years they've been pretty bad.  What's changed?
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6945


WWW
« Reply #11 on: September 29, 2012, 11:46:18 PM »
ReplyReply

Bob,

What was the latest model you used and what has been your experience with it?

For sure, taking about the 7600i, this is not a Nikon Coolscan 5000, but (a) it's about one-third the price bundled with SilverFast Ai and a profiling target of what the Nikon sold for (bundled with Nikon Scan) when the Nikon was available at normal retail prices, (b) the Plustek is a manufactured, current product and (c) it can produce decent scans - at least the one I tested back in mid-2010 - http://www.luminous-landscape.com/reviews/scanners/plustek.shtml.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad