Scanning by camera



The new way of scanning old negatives is gaining momentum nowadays. Countless DIY descriptions at web explain how to use DSLR or even point-and-shoot consumer camera for your films digitizing instead of film-scanner. Most of articles are about cardboard boxes with lamp, but there are devices, manufactured from metal, just like ‘Photo-Toaster’ by Cecil Williams. Most of pros, reading about such ‘inventions’, can only make a crooked grin because they know such brands, like ‘Nikon Coolscan’ and, at last, ‘Imacon’. Those who grew accustomed to high quality, look with distrust to flatbed scanners equipped with slide-module. What can we expect from cardboard box? PHOTOESCAPE will try to sort out advantages and disadvantages of camera-based negative digitizing.

What is a key difference between film-scanner and DSLR? Any digital camera equipped with rectangular CCD or CMOS light-sensitive chip, captures the whole image simultaneously. Film-scanners have a similar setup with flatbed scanners: one-pixel wide CCD-line, which moves opposite picture, captures its parts successively. Color scanning requires three CCD-lines covered by red, green and blue filters. This way of digitizing is old like IBM PC XT 286 computer and suitable for slow image processing, only available to old PCs. Such way allows getting a great resolution without extremely expensive large rectangular CCDs like 20 years ago. Now we have 15-20 Megapixel CMOS in most of DSLR, and even in cheapest consumer ‘soapboxes’. Why not to try using it for negatives digitizing? The main problem is high quality optical transfer of negative image to camera sensor. First of all, we need good macro lens to create 1:1 scale picture, in case of full-frame camera chip. BTW, any film-scanner is equipped with similar lens. The second thing is to create completely uniform negative lighting without hotspots and unwanted lens illumination. Let’s recall the experience of enlarger designers to understand how to illuminate negative properly.

There are two types of design: with reflected diffusive lighting and condenser-based with directed light. We’ll try the first way for beginning because condenser is too heavy and difficult to mount. Diffusive light can be made of simple cardboard box of shoe packaging. The greatest mistake of web DIY ‘inventors’ is using passing light diffused by paper sheet. Better way is to use the same sheet as a reflector moved from negative as far as possible to prevent its sharp imaging. Office paper reflects much more light then passes: this raises luminous efficiency dramatically. For most uniform light we need two lamps, located symmetrically in relation to film gate. The best decision is to use modern LED-lamps because of low heat release. One more important thing is to prevent direct negative lighting by light-shielding screens, mounted inside of our shoe-box. These screens must protect lamp light from negative surface, but let the light reach the reflecting screen. Paper tube doesn’t allow banning the outer light from the face side of negative because any paper, even deep black, reflects too much light, which decreases image contrast. The best way is to use our setup in dark room like in good old times. Those who have a lot of negatives don’t need to get accustomed.

Here we used two CFLs, but LEDs would be better due to higher efficient.

Now we have the box, the camera with macro lens on a tripod and the negatives. In our setup we used Canon EOS-5D Mk II camera with Canon EF 2,8/100 Macro USM lens. To get the highest possible quality of digitizing we have to lock all of this at the solid tripod and observe camera sensor and negative parallelism strictly. The box must be installed at the solid table surface. For more reliability we should put to our ‘lantern’ bottom a pair of used brake pads from a car. You can use any similar scrap metal, but not very rusty. These are basic things, needed for perfect result, but not primary. To get a good scan we must use definite camera settings. Most important is F-number, which can fall within f/8—f/16. The main reason is to provide maximum resolution without diffraction limit reaching. Shutter speed must be not shorter than 1/60 sec. because of LED-lamps flicker. In this way, ISO-setting have to be not higher then 100, and we must switch on mirror pre-lock. Shutter realizing must be done by wire remote or through self-timer. All these precautions let us prevent camera from shaking and get the sharpest possible image. Now we can try transfering some negs into digital by this setup. What can we finally get? To make things clear we’ll compare files, scanned from the same negative by film-scanner ‘Acer ScanWit 2720S’ and DSLR. For the average b&w 35-mm negative camera result is not very impressive: if to focus at film grain, film-scanner has better resolution. But it’s not so critical, as most pros think. Camera scanning has some advantages, which all of us can discover by using this method.

Overall view of setup taken with outdoor lighting for better visibility

First of all, we can scan ANY negative size by the same setup. Film gate can be changed as easy as our cardboard box was made. The second advantage is: film digitizing by camera is MORE quickly than by film-scanner. Those who used ‘Nikon Coolscan’ or other device types several years know that you can drink a lot of coffee cups while you scanner ‘irons’ negatives at highest quality settings. Camera needs a few seconds to make a number of different samples. All another advantages must be very interesting for experienced photographers. Good news for newspaper shooters who have a lot of overexposed negatives, processed in daily rush with hot developer. Now you can scan ANY negative in the same way as you printed them on photographic paper. Delicate lamp of film-scanners is useful only for properly exposed and developed films, but can’t ‘break through’ jet black negs. Camera doesn’t have such a problem because you can set any shutter speed. One more advantage is good for proof-framing lovers: you can scan not only effective area, but the whole film with perforations.

Strongly overexposed negative, scanned by film-scanner (left) and by camera (right)

Let us check an overexposed negative, which can be in the cupboard of any self-respected photographer. We’ll try to scan strongly overexposed negative of 1987 that could be easily printed at photographic paper, even with enlargement until 20×25 inches, but now is not useful to scan by film-scanner. It was shot by ‘Zenit” camera with shortest shutter speed 1/500 sec through catadioptric lens ‘MC ZM-5SA’ with constant aperture f/8. Such negative density was the only one available for shooting shiny water surface at available settings. ‘Acer ScanWit’ can only make poor-quality scan, when we bend characteristic curve to the extremely high amplification of highlights. And even in this case, water light spots are displayed very bad and show the defects of CCD-line at highest gain.

Enlarged highlight spots of picture above. Scanned by film-scanner (left) and by camera (right)

The same negative is digitized by our setup, showing clear highlights with proper shadows. For more flexibility you can use RAW to fine tune image halftones. Now try to switch on RAW in any film-scanner! 35-mm films can be digitized by camera with slightly lower quality, then by film-scanner. We don’t have possibility to make the same comparison for mid-format negs, but we can associate our technology to flatbed scanner. In this case we’ll discover a big surprise. Comparing enlarged frame piece, scanned by ‘Epson Perfection 4990’ with the same square, cropped from Canon camera shot of 6×6 negative, we can find out that flatbed scanner’s grain is smeared. Camera scan has enough quality. There is only one disadvantage: you need to invert negative image and adjust curves at Photoshop manually while flatbed scanner does it automatically. But for experienced pro this is not a problem. For good scan quality hand-made tuning is absolutely necessary.

But all these notes were just the introduction for most interesting part of our research. It’s not a coincidence that PHOTOESCAPE wrote about two manners of negative illumination in the beginning of the article. Now we’re going to try the second way with directed light, focused by condenser. And we’ll use most extreme type of lamp – the spot source of light without any diffuser. This type of enlarger light was very popular in b&w art photography during 1980th. We don’t know about Wild West, but those years it was extremely fashionable at least among Soviet pro shooters. By using fine-grain films photographer could make extremely sharp pictures with rich details in shadows, which was invisible with traditional bulb closed by frosted glass. There was a whole school of photography technique for ‘spot light’ with special soft developers, sharping image edges. The only disadvantage was a large number of dust and dirt on negative, which appears under such lighting. In fact, the enlarger with spot light source turns to defectoscope, stressing any dust, any film scratches, and even condencer bubbles like X-ray. Most delicate negative manipulation and dust safety did not save from painstaking retouch of every print. But for the digital files this is not the problem, that’s why PHOTOESCAPE was eager to use the same technology at digital: is that possible or not?

Cardboard and cutter are not useful in this case because glass condenser is too heavy to secure it without metal frame. Lamp position needs precise adjustment by special mechanism, making supposed device more complex. The best way is to modify usual negative enlarger, combining it with DSLR. The old Poland ‘Krokus 4SL’ became the victim of PHOTOESCAPE, turning from enlarger to scanning setup. It is equipped with large glass condenser, suitable for illuminating negatives up to 6×9 centimeters. Our first designer’s decision is to ‘drop’ enlarger optical axis horizontally for more comfortable camera operation. To lock the camera we’ll just modify or change enlarger basement. ‘Krokus’ is equipped with wooden board for placement to the working table like most other enlargers and we can cut off a half or use the new board. The new basement is more preferable because enlarger still stays useful for direct purpose, which is to print on photographic paper. New board size must be quite right for camera and enlarger condenser optical axis alignment. For camera mounting we can use metallic angled profile with drilled holes. Two holes are for screwing to board, and the last one is for tripod socket. After profile mounted to board one more hole must be drilled in the wooden basement for enlarger rod bracing. To find placement precisely it would be better to try on enlarger head to fitted camera with lens.

Overall view of scanning device with spot light source.

The second modification depends on changing the lamp. Instead of usual electric bulb we chose a miniature LED-lamp for G4 socket. In our device we used LEDARE lamp by IKEA with smallest radiating area from available devices of the same type. 30 years ago all of us used car lamps with tiny filament, but now this is real anachronism with too large heat release. One more thing is the necessity of lamp adjustment rod extension for about 2—3 inches. The reason is too short original rod, dedicated for bulky bulbs. We can’t adjust our tiny lamp correctly without rod extension. After basement modification and lamp changing we can mount our camera to the setup and begin to use it. Enlarger lens must be detached, of course. Operation algorithm is the same as for usual darkroom printing. First of all, we should choose the image scale by moving enlarger head closer or farther from camera. Our target is to fill the viewfinder by negative image completely. Because of close distance, lens focusing affects the image size equally to enlargement, that’s why we have to focus and change the scale simultaneously. After negative fills the frame, we can move up enlarger’s lensboard closer to camera lens to cut off outside light. This will be an additional camera stabilizing that reduces any shake.

Optical axis of camera and condenser must be aligned precise

There is a big surprise for beginners: lens movement almost doesn’t affect image sharpness, when negative is illuminated by the spot light source. This is because of specific light beams’ structure, different from usual in camera and projectors with apportioned light source. When we use spot lamp, almost all light energy concentrates in narrow ‘channel’ in the centre of aperture, just like iris closed to f/16—f/22. That’s why such light is so curious for us: image builds by only center part of lens, the most free from aberrations. The sharp picture can be captured by mediocre lens, but the good one will ‘draw’ ideally. Image focusing is still needed, even with spot lamp, but in this case light source must be covered by frosted glass for sharp visibility. Under light diffuser we can focus image precisely and even can use AF. After that procedure we should remove frosted glass, and most important part of enlarger adjustment begins. Now we have to adjust the lamp position carefully to get most uniform light. Most comfortable way of adjustment is to switch on Live View mode in a camera with aperture, closed to f/11—f/16. This allows watching light spreading through additional mirror while moving the lamp rod. Otherwise you’ll need very long hands or neck to do all things simultaneously.

Untuned spot lamp image at LCD screen of camera.

The main principle of adjustment is familiar to all photographers who adjusted enlarger lamp at least once. If lamp adjustment is incorrect, image of filament takes only one part of full frame. Proper adjustment means that all image square is illuminated uniformly. Except lamp centering you must move lamp rod closer or farther from condenser to find the position when lamp image is completely defocused and fills the whole frame. In this case, lamp image projects by condenser directly into the entrance pupil of lens, providing even frame lighting. One of key conditions for precise lamp focusing is the lens and condenser focal distances that match. Every condenser is made for definite lens focal length. For example, our Canon EF 2,8/100 Macro USM matches two glass condenser of ‘Krokus’ without additional glass, provided for 50 mm lenses. Improper condenser doesn’t let to adjust lamp correctly.

After light alignment our setup is ready for negative digitizing. But there is a number of differences from previous cardboard setup using. For the spot lamp camera lens’ F-number almost doesn’t affect to exposure for the same reason, which we discussed before. The basic light amount runs through narrow channel inside the beam centre. That’s why we have to shoot at wide open iris, despite adjusting the lamp at closed one. Closing of aperture during lamp adjustment provides appointment of radiating area image directly in the center of entrance pupil for maximal efficiency. This will force us using very fast shutter speeds, even at lowest ISO settings. For 1,1Watt IKEA lamp average speed with normal exposed 35-mm negative is about 1/2000 sec, but for mid-format films speed it increases until 1/8000. So, the low-voltage lamp is the greatest light source because we can use DC-power battery instead of AC-current via step-down transformer. Low LED lag leads to lamp flicker, unacceptable for fast shutter speeds. 12V is a popular voltage for batteries, for example in cars, but we used NP-E3 from Canon EOS-1D. This battery consumption is enough for several hours of working with such lamp.

Enlarged samples of negative, scanned by different ways. From left to right: flatbed scanner, film-scanner, spot-light enlarger.

After all technical adjustments we can take first pictures and compare them with the same, created by film-scanner and our previous technology with cardboard box. First of all, image structure is absolutely unusual and does not look like any other. Film grain is absolutely sharp, excelling ‘shoebox’, flatbed scanner and even film-scanner. Halftone imaging is different too. Except high contrast we’ve got countless tiny details in deep shadows. Highlights are more ‘flat’ than with usual scanning ways, but in the final image this is not so important. All of this is referred to both silver grain b&w and chromogenic color or monochrome films. The expected disadvantage is great amount of dust and dirt, which can be removed with careful retouch. Even the condenser bubbles and scratches became visible at the final image, especially in our ‘Krokus’. The reason is a condenser position, which is too close to negative. The best decision is to move it up for at least 1 inch. Another way to decrease dust quantity is to remove glass plates from negative holder, which are absolutely useless due to wide depth of field. We can change it for metallic or textolite plates with frame windows.

Described technology is too complex for customers who want to quickly and easily digitize their film archives. Fine retouch of any image can take several hours if you have a serious skill. But those photographers who used ‘spot lamp’ for darkroom printing at least once will be definitely intrigued to try this. Especially since enlarger modification almost doesn’t take a price, except new lamp. In some cases this setup is useful for scanning with usual lamp, which is more comfortable than ‘shoebox’. Finally, you can use very good enlarger lens instead of camera ‘macro’ if you have it. But the space between camera and enlarger lens needs thorough light seal. We hope, our ideas will guide you to the best and cheapest film scans ever!

©PHOTOESCAPE. All rights reserved.