Converting/scanning 35 mm neg/slides - scanning options?

Hi everyone,

 

I just added a Nikon FE2 to my collection, have used one the past, and it will complement my D700. Also have lots of AI and AIS primes. The big question is the most practical cost-effective way to convert to .dng or .tiff and process in ACR.

 

Option #1 - send the film or negatives to The Darkroom or Precision-Photo in Austin, TX and have them develop and high res. 4000+ x 6000+ scans. Easiest, but I don't know how good is the quality and I think they are returned as .jpegs. Also - expensive - developing, scanning, and shipping can be in the $25-35 range plus cost of film. Adds up real fast, but I'd probably get some good quality files ready for large prints (12 x 18).

 

Option #2 - dust off a Coolscan V and see if it still works with some old Vuescan software and my newer McBook Pro. Painful. I remember the scanning days.

 

Option #3 - and the one I'm leaning towards. D700, Nikon ES-2, tripod, 60 mm micro. The lab scans would quickly equal the cost of the 60 mm micro. My question is this - would the D700-ES-2 combo give adequate resolution for a 12 x 18 canvas print? I've done 16 x 24 prints in the past from 35 mm slides using to Coolscan V, but really don't want to go back there.

 

Thanks for comments and insight.

 

Gene

Option #4:

 

1) Purchase an inexpensive (perhaps used) laptop running Windows. It does not need to be very fast or have more than 4 GB of memory.

2) Install VueScan for the drivers.

3) Install Nikon Scan 4.03

4) Hook the Coolscan V to the laptop for scanning.

5) Transfer the scans to your Mac for processing.

 

The Coolscan V is one of the better 35mm film scanner ever made (at least until you get to the professional models). It will be better than any camera/macro lens setup and certainly easier to use.

The D700 is an FX camera with 12.1 MP resolution. The resolution is on a par with that of a 35 mm slide, but you would get better results with more resolution. The sweet spot with regard to cost and quality is probably 24 MP, 6000x4000 pixels, equivalent to that from a Nikon Coolscan.

 

I use a Micro-Nikkor 55/2.8 with an PK-13 extension tube (to get 1:1 magnification). The ES-2 will fit that lens (52 mm) at the right extension if you use the longer 62 mm tube and a 52-62 mm adapter on the lens. the assembly is locked together and rigid. There is no need for a tripod, even at long exposure times.

 

If you have both slides and film strips, then the ES-2 is a good choice. However the original ES-1 is faster to use with slides only.

 

I have a Nikon LS-4000 scanner, but I prefer to use use a camera (Sony A7Riii, A9 or A7iii). The quality is the same or better with the camera (no banding), and it takes less than 5 minutes to digitize an entire roll of 36, compared to 1-2 hours with the scanner.

Lots to ponder here. About five years ago I donated my Coolscan V to a wildlife refuge. It's still there - they haven't used it - I have access to it. But the time factor in scanning - I'll have to give my D700 a try and see if it has enough res. I don't see buying a D850 (although it would be nice) as likely.

I've done pretty much every kind of image scanning. IMHO the best results still come from a dedicated film scanner of reasonable quality.

 

In my case, I originally used a Canoscan 4000, and when it died (and its connection-capabilities are also obsolete) I went to a Nikon CoolScan 9000.If you buy an old computer and dedicate it to the scanner, it's still the most efficient way to scan, unless you're only doing a couple of images at a time. (for a start into a long chain of posts on this subject see LINK ignore the gibberish, a by product of old file formats being incompatible with changes on this site)

My "stay at home" project has been digitizing several decades of 35mm negatives. In the past, I've used DigMyPics and was pleased with their service but the cost of outsourcing to digitizing services for this project is significant. My Nikon Coolscan (with a dedicated Windows XP computer) works well but the scan process is slow. When NegativeLab Pro came available to simplify negative->positive conversion as a Lightroom plug-in, a DSLR scanning process was viable. I have a 55mm Micro-Nikkor w/tubes and a copy stand. I built a light box using high-CRI LEDs and bought a couple of negative strip holders that slide in tracks on the light box.

 

I am pleased with project progress so far. Handling negatives is always a PITA (curl, sticking together, and removing the #!&@ paper strips some labs love to attach) for either the Coolscan or loading the filmstrip holder. The Coolscan takes much more time to scan a strip compared to sliding the filmstrip holder over the light box and clicking the shutter using WiFi remote on an iPad (about two seconds per image). If I was using the Coolscan, I would be only 25% through the project instead of 90%.

 

Gear preparation and workflow is critical for this DIY approach. I focus on the grain and stop down to F/8 which gives DOF for residual buckling of the film in the holder. I am amazed at the amount of dust that shows up on the Nikon Z sensor that is always pointed down on the copy stand (I always thought dust would drift downwards) so I have gone through a couple of sensor cleaning sessions. Dust on the negatives has not been a problem except for a handful of envelopes - once I see it, anti-static brushes have worked fine.

 

Once I get the images transferred to my Lightroom workstation, NegativeLab Pro does the conversion to positive. I am pleased with the results and I like the color rendition better than the Coolscan but that may be due to the conversion profile NegativeLab uses. I'll go back and use Coolscan on selected images for comparison. One thing I found was my standard lens (35mm F/2) in the film days was relatively soft. (A reason to tote the 4x5 for landscape work in those days).

 

About 60% of my project time is spent in Lightroom. NegativeLab Pro takes about two seconds per image to convert but I multi-task during batch conversions. The bulk of the time is adding keywords and rating to the images. The essential keyword is the lab envelope number for that film roll so I can find the negative if I need to re-scan. Before scanning the first negative of a roll, I write the number and date from the lab envelope on a strip of paper, load it into the filmstrip holder, and photograph it. With each Lightroom import, that keyword image is first so I can tag the import correctly. Once the negatives are converted, I rate and add keywords.

 

NegativeLab Pro conversion is sensitive to the amount of non-image (i.e. negative edges) in the image it converts. There is a setting for percentage of image edge to ignore during conversion, but I get best results by cropping followed by un-converting and converting the negative again. The conversion of the cropped area is usually right on target for color balance.

 

The color negative conversion segment of the project is almost complete. When my attic archeology digs turn up ancient B&W negatives, I'll modify the lightbox for large format and start on those. Slides are easy and since I can pre-filter the images on a light table, that segment of the project will go faster.

When copying negative color film, it's a good idea to set a fixed white balance with an empty film holder. Otherwise the camera may overreact to the orange mask. I concur with cropping the image before conversion to remove the orange, unexposed borders. I have had good luck using the back-end of SilverFast software for the conversion. Negativelab sounds like its worth a closer look.

I posted these recently in another forum.

1) A small crop from T-max 100 - one of the finest-grained films available - digitised at greater than 6000 ppi:

2) A DSLR shot using a Canon 12 Mp 5D - a very similar spec to the D700. Taken from the same camera position using the same lens:

The crop from both is the same small area of the frame. I know which image I'd rather use for a 12x18" print.

 

So if you think film is going to get you a better print than a direct digital shot, think again.

 

However, 'scanning' with your AA filtered 12Mp D700 isn't going to get anywhere near to the above clarity from film. I used a 24Mp camera with no AA filter, and shot only a small section of the negative to give an effective 6000 ppi or more scan.

 

Also, not worth buying the 60 AF Micro-Nikkor just for film copying IMO. You'll get just as good, if not better results from an Ai 55mm Micro-Nikkor, which can be got used for about half the price.

I used to own a D700. The resolution is below that of fine grained 35mm films. That is to be expected with just 12mp. Scanning on a Nikon 5000, I found my Astia 100F and films like it came in around 18mp and matched and sometimes exceeded my 18mp Canon 7D. Another issue of course with b&w films is that a 12mp DSLR with a Bayer interpolated image is even less than a real 12mp mono image.

I used to own a D700. The resolution is below that of fine grained 35mm films. That is to be expected with just 12mp......

Can we see your comparison shots then please Dave?

 

Full-frame digital to 35mm film? And not on TechPan or Copex, but using a film with a viable ISO speed.

Funny how everything comes back to Kodak Technical Pan, designed for microfilm and military reconnaissance. It was typically shot at ISO 25, and required a special developer to achieve anything close to a usable tonal range. I'll bet your dad had lots of Tech Pan photos in the family album.

Funny how everything comes back to Kodak Technical Pan, designed for microfilm and military reconnaissance. It was typically shot at ISO 25, and required a special developer to achieve anything close to a usable tonal range. I'll bet your dad had lots of Tech Pan photos in the family album.

 

And it had an extended red response curve. This removed or reduced blemishes (zits, pimples) on Caucasian skin, which was an advantage, a big advantage, in the days before Photoshop and spot healing brushes and the like.

 

Tech Pan was a great film for portraits, formal or informal. The low ISO was not a problem if you used flash. But it is probably not the film to shoot you child's Junior High School basketball game. :-)

Can we see your comparison shots then please Dave?

 

Full-frame digital to 35mm film? And not on TechPan or Copex, but using a film with a viable ISO speed.

Hi!

 

This is Gene, the OP. I'm going to find out real soon. As soon as the Memorial Day crowds clear out from South Padre Island, I'll go out with the FE2 and D700 and do some surf and sky/clouds photography. It may take some time, send out the film, developing, then I'll scan on Coolscan V. Right now I have three rolls of Superia 400 in the house - I prefer Portra 160 and 400, but that will have to wait. One thing I've noticed, and this has nothing to do with resolution or image quality, is just how much easier it is to frame and composition with the smaller FE2 and a manual focus prime.

 

Gene

Can we see your comparison shots then please Dave?

 

Full-frame digital to 35mm film? And not on TechPan or Copex, but using a film with a viable ISO speed.

 

did tech pan and Velvia way back. The 18mp was processed normally. The fil. Had no contrast or sharpening applied. The film was basically identical to 18mp...so yes, better than 12 mp just like I said. Look at the map section of Nigeria where the Velvia showed the blue river....but the digital sensor didn’t pick that up...lost in Bayer interpolation. I obtained the same results with Astia 100F, Provia, Ektachrome, etc. 12mp is not enough for fine grained 35mm film.

To my eye, the Canon 7D image is the sharpest of the set. This is clearly evidenced in the text, as well as other details. The Canon seems perfectly capable of rendering blue rivers.

 

Having "scanned" hundreds of slides with a camera, it becomes obvious that any lack of sharpness is in the original image (slide), not the digital rendering. If the copy is grain-sharp, it can't get any better. Look at a Velvia slide with a 20x magnifier, comparable to pixel-peeping a 24 MP image, and see for yourself. There are several reasons for this phenomena. Digital image have higher acuity, even at normal (6:1) contrast ratios. Film may do better at absolute resolution, but only with high contrast (1000:1) resolution targets. There have been vast improvements in lenses prompted by this acuity, and focusing is much more precise, particularly with mirrorless cameras or live-view.

 

Digital cameras will never have the color quality of film. Balderdash! Digital cameras are mostly very accurate, whereas film distorts and exaggerates colors in a way some find more appealing. The same landscape taken with a digital camera and Velvia look quite different. But if you copy the Velvia slide with a digital camera, it looks like Velvia.

The

To my eye, the Canon 7D image is the sharpest of the set. This is clearly evidenced in the text, as well as other details. The Canon seems perfectly capable of rendering blue rivers.

 

Having "scanned" hundreds of slides with a camera, it becomes obvious that any lack of sharpness is in the original image (slide), not the digital rendering. If the copy is grain-sharp, it can't get any better. Look at a Velvia slide with a 20x magnifier, comparable to pixel-peeping a 24 MP image, and see for yourself. There are several reasons for this phenomena. Digital image have higher acuity, even at normal (6:1) contrast ratios. Film may do better at absolute resolution, but only with high contrast (1000:1) resolution targets. There have been vast improvements in lenses prompted by this acuity, and focusing is much more precise, particularly with mirrorless cameras or live-view.

 

Digital cameras will never have the color quality of film. Balderdash! Digital cameras are mostly very accurate, whereas film distorts and exaggerates colors in a way some find more appealing. The same landscape taken with a digital camera and Velvia look quite different. But if you copy the Velvia slide with a digital camera, it looks like Velvia.

The blue river in Nigeria is missing on the digital sot...so no...it clearly CANNOT pull that detail out. And as I said, the film images are unsharpened....the issue here was resolution...not acutance. The film clearly resolves more than 12 mp as it is giving the 18 a run for its money. That was the point. I boosted the contrast of the film image and here it is clearer the resolving power between the two is basically the same. Thus, 12 mp being sufficient is incorrect.

Edited May 25, 2020 by Dave Luttmann

Oh that river! It's there in the digital image, just not blue.

Oh that river! It's there in the digital image, just not blue.

Same map. Digital missed it. Sorry...but everything I said is correct. 12mp is not enough

I would recommend using the Coolscan as you already have it even if you have to get a separate computer for scanning.

If you don't I would recommend to to use the D700.

If the OP purchases a macro lens setup for the D700, as suggested in my first post, it can be used on any Nikon DSLR or any of the mirrorless, high resolution cameras on the market. It won't wear out or go out of style for this intended use. The only maintenance issue is oil migration to the diaphragm, making it stick or become sluggish. That can be fixed by any competent repair service. Mine was done nearly 20 years ago, using (presumably) better grease.

 

At 12 MP you may or may not see dye clouds. At 24 MP, dye clouds are clearly visible, and at higher resolutions, the dye clouds become clearer. The film image cannot be any sharper than the grain or dye clouds allow, and is usually much less.

 

Here's the rub. If your digital camera makes copies of slides indistinguishable from the original or LS-5000 scans, what is the incentive to keep shooting film at $20/roll with processing? If you have a large collection of old slides and negatives you'd like to digitize, you probably won't take the time to do it with a scanner. With a DSLR (or mirrorless) you can do several hundred in a couple of hours and process the results at your leisure. Slides require little or no post-processing. Negatives are something else (another time, another thread).

 

There is a big difference between 12 MP and 18 MP, and an 18 MP camera (e.g. Leica M9) without an anti-aliasing filter is the equivalent of 24 MP with AA. Except for a few pockets or resistance, the film vs digital battle was decided 15 years ago. That said, if photography is your pleasure, not your bread-and-butter, do what gives you the most satisfaction. It's not necessarily one or the other.

 

Resolution is subject to the root-sum-square rule of uncertainties. If the film, lens and digital process have the same resolution, the results will less than the weakest link. Velvia has a maximum resolution of about 160 lp/mm, measured against a USAF metallized glass target with 1000:1 or better contrast. In normal practice, the contrast is much lower (6:1), and the resolution is on the order or 60-80 lp/mm, the rough equivalent of 6 MP.

Edited May 26, 2020 by Ed_Ingold

Bayer errors are mainly due to color interpolation, which can be as simple as nearest-neighbor priority. All elements of the Bayer array contribute to luminance, which is the main determinant of spacial resolution. It doesn't surprise me that details of rivers are rendered as lines of a slightly darker shade than the surrounding color. The river itself is clearly visible, as well as the text labels. If color detail is important, the solution is to use a higher resolution camera. There's a double benefit to that. Once you get beyond 24 MP spacial aliasing is much less important, because it occurs in repetitive details too small to be easily visible. For a commercial photographer, that might be in certain fabrics. For landscapes, watch out for picket fences near the horizon.

Look at the map section of Nigeria where the Velvia showed the blue river....but the digital sensor didn’t pick that up...lost in Bayer interpolation. I obtained the same results with Astia 100F, Provia, Ektachrome, etc. 12mp is not enough for fine grained 35mm film.

Not this again!

If you read through the original thread carefully, you'll find out that a different map was used, by a different poster, for the so-called 'comparison'.

did tech pan and Velvia way back. The 18mp was processed normally. The fil. Had no contrast or sharpening applied. The film was basically identical to 18mp...so yes, better than 12 mp just like I said. Look at the map section of Nigeria where the Velvia showed the blue river....but the digital sensor didn’t pick that up...lost in Bayer interpolation. I obtained the same results with Astia 100F, Provia, Ektachrome, etc. 12mp is not enough for fine grained 35mm film.[ATTACH=full]1343101[/ATTACH]

Looks like the Canon by a hair over the tech pan.

BTW, still waiting for your own tests or examples to be shown Dave.

Those map pictures were originally posted 10 years ago by member 'danielleetaylor' in this old thread, also using example pictures by someone called Les Sarile.

 

If you read through that thread, danielleetaylor freely admits that his copy of the map isn't the same one as was used for Les's film tests. This can clearly be seen where colours and details differ between the examples. This completely nullifies any comments about a 'blue river' being missing and suchlike.

 

And to nitpick, isn't it against forum rules to post someone else's pictures without credit?

Edited May 28, 2020 by rodeo_joe|1

Tech Pan was a great film for portraits, formal or informal. The low ISO was not a problem if you used flash.

The finicky and erratic processing needed definitely was a problem though.