Larry Rexley wrote:
I have a lot of my Dad's old Kodachrome slides. I transferred about 10 boxes of slides so far. I used the M6 mark ii and Canon EF-S 55-250 IS STM with an extension tube, in a dark room the slides about 3-4 feet away from the lens, behind the slides was a piece of white paper lit by 2 3000k LED work lights.
This worked very well for me --- as I could use my standard workflow using DxO Photolab to process the results quickly. DxO excels at white balancing, adjusting colors and tones, contrast --- but it's best attribute is its ability to sharpen and de-noise. I was able to bring the grain to the level where you just see a hint of it.
I have tried quite a few film scanners and like you found them FAR too slow to make doing more than just a handful of slides or negatives very painful.
I just got the EF-M 28mm macro a month ago --- based on your results I am going to try that next! I think I will try putting a bunch of empty filter rings, and then some 'step-up' rings (I have a full set stepping up from 49mm to 82mm with many rings) to see if 35mm slides will 'fit' in one of the rings.... I'll try to make a make-shift slide adapter at the right distance to fill the frame with a slide.
I'll report back how it goes.
I spent some time making a slide duplicator for the EF-M 28mm macro lens with several step-up rings, and a bunch of old filters with the glass removed. I found that the setup worked best when the slide film plane was about 32mm from the plane of the front of the ES-22 adapter (43mm filter thread) mounted on the EF-M 28mm lens like so:
From the front of the ES-22 filter adapter is the following: a 43->55mm step-up ring (4mm thick) - a stack of empty 55mm filters, including an empty polarizer filter ring which allows you to rotate the slide to 'level it' before shooting (the stack is 26mm thick from the front of the 43->55 step-up ring to the rear of the 55->67mm step up ring), and a 55->67mm step-up ring (4mm thick: the slide sits inside this about halfway, or about 2mm from the front of the filter stack).
This would also work with a step-up ring 43mm->52mm, stack of empty 52mm filters 26mm thick, and a step-up ring 52->67mm. That would be even better, as the 55->67mm step up ring allows a tiny amount of light around the edges of the slide, and a 52->67mm ring wouldn't. But I had all the right stuff to do a 55mm stack and the step-up rings for my infrared work, so I used them. it would also work with a 49mm stack and the appropriate step-up rings, but anything smaller than 49mm would block light from the edges of the slide.
Square slides are slightly too large to fit into the 67mm step-up ring, but I 'nip' about 1mm off each corner with a pair of scissors, so that the slides will sit snugly and securely in the ring without moving or falling out. Some of the slides processed by Kodak have rounded corners already and fit perfectly into the 55-67mm step-up ring.
To get a good light source, I used two work lights with 60-watt 3000k LED light bulbs placed 3 inches away from a stack of white printer paper. It is best if they are high quality LED lights with a Color Rendering Index of at least 90 to produce better full-spectrum color source, and I pointed the duplicator at a perfectly uniform part of the white paper, with the slide just a couple inches from the paper. I shot the images using Auto-White-balance with White Priority, but also would white balance in post using the eyedropper on something that was gray or white, when possible.
Exposures seem to be between 1/8 second and 1/30 second with this setup at ISO 100. It is important to shoot at ISO 100 to get the best dynamic range and most highilght and shadow detail to work with - critical for slide work.
If the filter stack is tight and snug, you can use slower exposures like 1/4s as the slide will not move relative to the camera sensor --- but you have to make sure to turn lens image stabilization off through the camera menus!! Otherwise the lens stabilizer will try to compensate for your hand-held motion (if you are hand-holding) which you do NOT want as the slide will move with the sensor and motion of the whole setup should not cause any blurring of the slide. I've gotten tack-sharp duplicates at 1/4 second (see last image below).
The problem with duplicating slides with most setups is that slides have an extremely high contrast range --- higher than a lot of duplicators can output. But the Canon M6ii, M50, and M200 have a great dynamic range that should be enough for most slides. Often you'll have to cut the contrast way back in post, or bring up the shadows pretty dramatically, and/or cut the highlights, especially if the original slide was shot in bright sunlight.
DxO PhotoLab is great at handling wide dynamic range, and as I've mentioned before, has great sharpening and de-noise abilities as well. Experimenting, I've also found that the micro-contrast feature of DxO works well for reducing the excessive grain that older slides often have, by dialing back the micro-contrast below the zero point, almost to the minimum value of the range.
Here's a typical example using one of my Dad's Kodachrome slides originally taken and processed in 1961 (61 years ago) with a vintage Kodak Pony 135 camera, a rangefinder 35mm camera. My dad used a hand-held light meter to determine exposure.
Here's the result right out of the duplicator --- in DxO Photolab 5, I have only white-balanced the image and downsampled it to 2160 pixels high:
1. 1961 Kodachrome slide duplicated with the Canon M6ii & EF-M 28mm f3.5 lens at f5.6, 1/25s, ISO 100
Here's the result after cropping the image and adjusting exposure, downsizing to 2160 pixels high with no other processing:
2. Same image after cropping and adjusting exposure, downsampling to 2160 pixels high
And here's a 'final' image after doing the following:
- Custom tone curve to bring out shadow and midrange detail
- Turn on Deep Prime de-noise, set to +95 to reduce severe grain
- Increase global sharpness to +50 to bring out detail (and balance de-noise sharpness loss)
- Apply unsharp mask of radius 91, Threshold 0, intensity 140 to bring out detail and balance de-noise
- Decrease micro-contrast to negative (-) 95 to further de-emphasize grain and dust
- Increase contrast to +20 to offset loss of contrast due to decreasing microcontrast
- Increase overall color vibrancy and saturation to +10, increase blue channel saturation to +10 to bring out blue in the sky and ocean, reduce Orange channel saturation to -30 to balance the skin tones
- Use the repair tool to 'clone out' most of the obvious dust on the slide (as R2 says it is best to blow or clean the slides before shooting to avoid this as much as possible!). I left a few so that you can see how reducing micro-contrast and increasing de-noise also reduced the impact of the dust fairly effectively
- Downsize image to 2160 pixels high (that resolution still captures all the detail in the image) - downsizing helps reduce the grain further. 'Bicubic Sharper' interpolation was used to preserve detail, and export image quality 85 was used to keep image size reasonable while maintaining very good image quality
3. Final image after processing for tone curve, de-noise, sharpness, reducing grain, repairing dust
Finally, here's another slide duplicated.... this was a Kodachrome slide from the late 1990's that I took with a Canon T70 and a sharp Canon lens.
This one needed a lot of color adjustment, but no custom tone curve.... only bringing up the shadows a bit. Similar de-noise and sharpening to the image above --- but I did not need to dial the microcontrast or contrast at all, they were both left at zero.
The train is unsharp as it was moving and the shutter speed wasn't very fast, however the bridge trestle is extremely sharp and shows no grain. This one was downsample to 2160 pixels wide, so is the same scale as the previous image.
4. Burlington-Northern train seen from a trestle on a hike, somewhere out West. Original was a late 1990's Kodachrome slide, duplicated with Canon M6ii, EF-M 28mm f3.5 lens at f5.6, 1/4s (hand-held in home-made duplicator), ISO 500
