Does MF color slides scanning in 24 bit still make sense
The main issue with camera capture vs drum scanning is not resolution nor dynamic range, it’s flare. Before discussing what I mean, let’s look at the question of resolution: Using a high quality, short focal length enlarger lens and sufficient stitching, any high-MP camera can easily equal or exceed the so-called “sampling resolution” (pixels per inch) of any drum scanner, for whatever that’s worth. Pick your pixel count and you can achieve it, subject to a little care in the stitching. But how much resolution do you need? The sharpest low-ISO slide films were capable of perhaps 60 line pairs/mm at 40% MTF, which, according to Nyquist, needs a sampling rate of about 6,000 dpi, give or take. But good luck finding a film original with such extraordinary image detail. It would have to have been captured (a.) on Kodachrome 25 or similar, (b.) with a very fine macro optic at optimum aperture and (c.) on a perfectly stable tripod, etc. Most of the images we think of as “tack-sharp” have about half that real resolution if they’re lucky, and we love them. Bottom line - whatever sampling resolution you decide you need, you can get it by camera stitching. Albeit with patience. Of course there’s the question of whether camera-captured pixels are as "sharp” as drum scanner samples. I won’t get into that, except to say that you’d be surprised how cheap some drum scanner lenses were. The notion that Crosfield, Hell, Screen and ICG spent the farm on high quality apochromatic microscope lenses is a little exaggerated. Next let’s look at dynamic range. Most good slide films have a dynamic range of at least 4.0 (max. - min. dye density) which corresponds to an f-stop range around 14. Interestingly, most drum scanners struggled to get anywhere near 4.0 DR. The log amp (logarithmic photomultiplier tube amplifier) in Crosfield scanners was artificially limited to about 3.0 but I was able to squeeze about 4.3 out of ICG scanners, with a bit of PMT noise. Few digital cameras can approach a 14 stop range in one shot, but with careful HDR, they can easily exceed it. However this is where flare begins to muddy the waters (or shadows). The flare problem. The one area in which drum scanners easily trounce flat-bed scanners is gross optical flare. The only flare you’ll find in a drum scanner is within a few microns of the scanning spot, the actual flare radius depending on the scanning plane optics. However all line-at-a-time (a.k.a. “CDD” or “flat bed”) scanners suffer from a much more severe problem that I call “linear flare", manifested as streaks of unwanted lightness where light and dark image areas meet. What makes this flare offensive is that it only occurs at right angles to the scanning direction. These directional flare lines are caused by the fact that light image areas must be illuminated at the same time as dark areas in the same scan line are being sampled. Drum scanners don’t do this. Camera captures also suffer from optical flare in the lens itself, but lens flare is often less noticeable (unless you go looking for it) than flat-bed scanner flare, because (a.) it’s “radial” and therefore symmetrical around the bright detail and (b.) it is often mistaken for flare in the original, i.e. caused by the original taking lens. Unfortunately, if you rely on HDR to expand your camera-copy dynamic range, the copy-camera’s lens flare can actually be magnified. That’s because the HDR process typically relies on some form of blurred masking to blend light and dark exposures. The blurred mask can itself contribute to ghostly shadow anomalies if you’re not careful. The good news is that scanner-produced flare (whether linear or radial) is mostly noticeable when you have to exhume dark shadow detail, i.e. open up an under-exposed transparency. Normally-exposed slide scans usually look fine. OK. That’s enough typing for a cold New Jersey Sunday. ........................................................ Don Hutcheson, President HutchColor, LLC Washington, NJ USA don@hutchcolor.com M: 908-500-0341 ........................................................ On Jan 17, 2016, at 15:00 , colorsync-users-request@lists.apple.com wrote:
Send Colorsync-users mailing list submissions to colorsync-users@lists.apple.com
To subscribe or unsubscribe via the World Wide Web, visit https://lists.apple.com/mailman/listinfo/colorsync-users or, via email, send a message with subject or body 'help' to colorsync-users-request@lists.apple.com
You can reach the person managing the list at colorsync-users-owner@lists.apple.com
When replying, please edit your Subject line so it is more specific than "Re: Contents of Colorsync-users digest..."
Today's Topics:
1. Re: Does MF color slides scanning in 24 bit still make sense (Paul Schilliger) 2. Re: Does MF color slides scanning in 24 bit still make sense (Mike Strickler)
----------------------------------------------------------------------
Message: 1 Date: Sat, 16 Jan 2016 22:37:53 +0100 From: Paul Schilliger <pschilliger@sunrise.ch> To: Ernst Dinkla <info@pigment-print.com> Cc: "colorsync-users@lists.apple.com" <colorsync-users@lists.apple.com> Subject: Re: Does MF color slides scanning in 24 bit still make sense Message-ID: <569AB831.5080607@sunrise.ch> Content-Type: text/plain; charset="ISO-8859-1"
------------------------------
Message: 2 Date: Sat, 16 Jan 2016 20:36:41 -0800 From: Mike Strickler <info@mspgraphics.com> To: colorsync-users@lists.apple.com Subject: Re: Does MF color slides scanning in 24 bit still make sense Message-ID: <6A7CB0F8-5615-442E-A079-815430E6B2A1@mspgraphics.com> Content-Type: text/plain; charset=us-ascii
I think one should be clear about the intended purpose of such DSLR captures. Expectations should probably be set at slide shows on computer displays and small prints--IOW good quality digital archiving. Sharp prints at 10X or greater enlargement may require higher spatial and tonal resolution (and lower flare) than these cameras can achieve. Best results with a camera will be had with a dedicated copying lens such as the Apo-Rodagon D and a slide copier. Enlarging lenses can also do a good job. Be mindful that the the focus error alone with mounted slides may seriously compromise results. A PMT drum scanner will, on the other hand, fix the entire film orignal at a precise distance from the lens, which is normally an apochromatic microscope objective, diffraction-limited at below f/1. 36 MP captured on such a device is a rather different thing from what you will get on a Nikon D800. A proper drum scan for a 10X enlargement of a medium format film will begin at 4,000 dpi, yielding over 60 MP, and will reveal the subtlest gradations and grain structure. A compromise between speed and quality might be found in a dedicated slide scanner such as the Nikon Coolscan. But even here wet mounting is recommended, and this again takes too much time for most people trying to record many images. My preference would be to drum scan any image to be critically printed and shoot the rest on a slide copier using a copy lens and decent DSLR.
------------------------------
_______________________________________________ Colorsync-users mailing list Colorsync-users@lists.apple.com https://lists.apple.com/mailman/listinfo/colorsync-users
End of Colorsync-users Digest, Vol 13, Issue 10 ***********************************************
To create less flare in desktop scanners would require other sensor types than the usual linear or 3 linear CCDs. Even if the halogen, CCFL or linear LEDS are replaced by something like a flying spot illumination there will be neighbour sensor wells receiving light that should not go there. I have thought about different geometries in the optical path, with for example the linear CCD at an angle to the scan direction + a flying spot illumination but then the usual narrow split mirror path would not be possible and the data stream more complex Googling for the use of RGB LEDstrips (+ Ulbricht sphere etc) to make an ideal light source for slide copying I came across this thread where they use the RGB LEDstrips and more: http://www.dpreview.com/forums/thread/3546695 They aim at LF films there, so step and repeat copying. I think they should mask the film more, if the videos show actual practice. I use wet mounting on the Coolscan 8000 and was probably the first one to do that. In a camera/film copying system for B&W 35mm film strips I think it should be possible to make a narrow glass channel filled with fluid that allows a faster wet mounting/embedding when strips are pulled through frame by frame and still keep focus then by an AF system. In movie film copying other wetting systems are used (but with flying spot illumination which reduces flare). The Silverfast and Vuescan multi sampling settings for several scanners do deliver if dynamic range has to be increased. On optics, the best 1:1 macro resolution lenses are probably the Printing-Nikkor lenses. Marco Cavina is quite sure that the larger Coolscan 8000 has that lens design integrated as well, explains part of the pricing. It can not be the film holder + frame switching, worst part in the design of that scanner. http://www.marcocavina.com/articoli_fotografici/Nikon_EL-Nikkor_lenses/00_pa... Bottom page, in Italian but Google does its best in the translation. Met vriendelijke groet, Ernst Dinkla Dinkla Grafische Techniek Quad, piëzografie, giclée www.pigment-print.com On Sun, Jan 17, 2016 at 11:35 PM, Don Hutcheson <don@hutchcolor.com> wrote:
The main issue with camera capture vs drum scanning is not resolution nor dynamic range, it’s flare.
Before discussing what I mean, let’s look at the question of resolution:
Using a high quality, short focal length enlarger lens and sufficient stitching, any high-MP camera can easily equal or exceed the so-called “sampling resolution” (pixels per inch) of any drum scanner, for whatever that’s worth. Pick your pixel count and you can achieve it, subject to a little care in the stitching. But how much resolution do you need? The sharpest low-ISO slide films were capable of perhaps 60 line pairs/mm at 40% MTF, which, according to Nyquist, needs a sampling rate of about 6,000 dpi, give or take. But good luck finding a film original with such extraordinary image detail. It would have to have been captured (a.) on Kodachrome 25 or similar, (b.) with a very fine macro optic at optimum aperture and (c.) on a perfectly stable tripod, etc. Most of the images we think of as “tack-sharp” have about half that real resolution if they’re lucky, and we love them. Bottom line - whatever sampling resolution you decide you need, you can get it by camera stitching. Albeit with patience.
Of course there’s the question of whether camera-captured pixels are as "sharp” as drum scanner samples. I won’t get into that, except to say that you’d be surprised how cheap some drum scanner lenses were. The notion that Crosfield, Hell, Screen and ICG spent the farm on high quality apochromatic microscope lenses is a little exaggerated.
Next let’s look at dynamic range.
Most good slide films have a dynamic range of at least 4.0 (max. - min. dye density) which corresponds to an f-stop range around 14. Interestingly, most drum scanners struggled to get anywhere near 4.0 DR. The log amp (logarithmic photomultiplier tube amplifier) in Crosfield scanners was artificially limited to about 3.0 but I was able to squeeze about 4.3 out of ICG scanners, with a bit of PMT noise. Few digital cameras can approach a 14 stop range in one shot, but with careful HDR, they can easily exceed it. However this is where flare begins to muddy the waters (or shadows).
The flare problem.
The one area in which drum scanners easily trounce flat-bed scanners is gross optical flare. The only flare you’ll find in a drum scanner is within a few microns of the scanning spot, the actual flare radius depending on the scanning plane optics. However all line-at-a-time (a.k.a. “CDD” or “flat bed”) scanners suffer from a much more severe problem that I call “linear flare", manifested as streaks of unwanted lightness where light and dark image areas meet. What makes this flare offensive is that it only occurs at right angles to the scanning direction. These directional flare lines are caused by the fact that light image areas must be illuminated at the same time as dark areas in the same scan line are being sampled. Drum scanners don’t do this.
Camera captures also suffer from optical flare in the lens itself, but lens flare is often less noticeable (unless you go looking for it) than flat-bed scanner flare, because (a.) it’s “radial” and therefore symmetrical around the bright detail and (b.) it is often mistaken for flare in the original, i.e. caused by the original taking lens. Unfortunately, if you rely on HDR to expand your camera-copy dynamic range, the copy-camera’s lens flare can actually be magnified. That’s because the HDR process typically relies on some form of blurred masking to blend light and dark exposures. The blurred mask can itself contribute to ghostly shadow anomalies if you’re not careful.
The good news is that scanner-produced flare (whether linear or radial) is mostly noticeable when you have to exhume dark shadow detail, i.e. open up an under-exposed transparency. Normally-exposed slide scans usually look fine.
OK. That’s enough typing for a cold New Jersey Sunday.
........................................................ Don Hutcheson, President HutchColor, LLC Washington, NJ USA don@hutchcolor.com M: 908-500-0341 ........................................................
On Jan 17, 2016, at 15:00 , colorsync-users-request@lists.apple.com wrote:
Send Colorsync-users mailing list submissions to colorsync-users@lists.apple.com
To subscribe or unsubscribe via the World Wide Web, visit https://lists.apple.com/mailman/listinfo/colorsync-users or, via email, send a message with subject or body 'help' to colorsync-users-request@lists.apple.com
You can reach the person managing the list at colorsync-users-owner@lists.apple.com
When replying, please edit your Subject line so it is more specific than "Re: Contents of Colorsync-users digest..."
Today's Topics:
1. Re: Does MF color slides scanning in 24 bit still make sense (Paul Schilliger) 2. Re: Does MF color slides scanning in 24 bit still make sense (Mike Strickler)
----------------------------------------------------------------------
Message: 1 Date: Sat, 16 Jan 2016 22:37:53 +0100 From: Paul Schilliger <pschilliger@sunrise.ch> To: Ernst Dinkla <info@pigment-print.com> Cc: "colorsync-users@lists.apple.com" <colorsync-users@lists.apple.com> Subject: Re: Does MF color slides scanning in 24 bit still make sense Message-ID: <569AB831.5080607@sunrise.ch> Content-Type: text/plain; charset="ISO-8859-1"
------------------------------
Message: 2 Date: Sat, 16 Jan 2016 20:36:41 -0800 From: Mike Strickler <info@mspgraphics.com> To: colorsync-users@lists.apple.com Subject: Re: Does MF color slides scanning in 24 bit still make sense Message-ID: <6A7CB0F8-5615-442E-A079-815430E6B2A1@mspgraphics.com> Content-Type: text/plain; charset=us-ascii
I think one should be clear about the intended purpose of such DSLR captures. Expectations should probably be set at slide shows on computer displays and small prints--IOW good quality digital archiving. Sharp prints at 10X or greater enlargement may require higher spatial and tonal resolution (and lower flare) than these cameras can achieve. Best results with a camera will be had with a dedicated copying lens such as the Apo-Rodagon D and a slide copier. Enlarging lenses can also do a good job. Be mindful that the the focus error alone with mounted slides may seriously compromise results. A PMT drum scanner will, on the other hand, fix the entire film orignal at a precise distance from the lens, which is normally an apochromatic microscope objective, diffraction-limited at below f/1. 36 MP captured on such a device is a rather different thing from what you will get on a Nikon D800. A proper drum scan for a 10X enlargement of a medium format film will begin at 4,000 dpi, yielding over 60 MP, and will reveal the subtlest gradations and grain structure. A compromise between speed and quality might be found in a dedicated slide scanner such as the Nikon Coolscan. But even here wet mounting is recommended, and this again takes too much time for most people trying to record many images. My preference would be to drum scan any image to be critically printed and shoot the rest on a slide copier using a copy lens and decent DSLR.
------------------------------
_______________________________________________ Colorsync-users mailing list Colorsync-users@lists.apple.com https://lists.apple.com/mailman/listinfo/colorsync-users
End of Colorsync-users Digest, Vol 13, Issue 10 ***********************************************
_______________________________________________ Do not post admin requests to the list. They will be ignored. Colorsync-users mailing list (Colorsync-users@lists.apple.com) Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/colorsync-users/ernst.dinkla%40gmail...
This email sent to ernst.dinkla@gmail.com
participants (2)
-
Don Hutcheson
-
Ernst Dinkla