Scan Slides with Kodak Carousel Projector (Hack)

[ColSebastianMoran]

Update... The Carousel projector has a fan and another motor running continuously. Vibration is an issue, never noticed while projecting, but can see it in trying for the best Camera-scans. I'm re-thinking about the bulb wattage; the projector bulb will result is high shutter speeds. And, also thinking about clamping down everything (projector and camera) to eliminate more vibration sources. More to come, but it'll be a while.

 

[ColSebastianMoran]

Back at it. There's a resolution loss in my rig.

Pretty sure it's vibration from the fan. Using an 8-watt LED, I think I can just turn the fan off (cut the wire). But, I still want faster shutter speeds, was exposing at 1/10. So, I changed the diffuser, first to a single sheet of drafting mylar.

Also, removed the "heat absorbing glass." This glass take about two stops and shifts color.

Shutter speeds now 1/500 at f/5.6 ISO 100. This mylar diffuser seemed to be somewhat visible in my image files. So, I switched to a reflective diffuser. A single sheet of good white paper across the mirror. (Photo coming.)

Still 1/125 to 1/250 and my resolution tests exactly equal to my alternative very solid rig with the Nikon bellows and slide copier (PB-4 and PB-5).

Testing:
- Check focusing; ten tries, all same good resolution.
- And, the resolution is equal to what I get with the Nikon bellows rig.
- With the fan on, I lose just a bit of resolution, so I will cut the wire.

So, I think I'm now ready for production at 24MPx. I think this will be all I need for most of my chromes.

 

[ColSebastianMoran]

More about resolution.

In my version 1, with 105mm f/2.8 Micro Nikkor AF, using auto-focus, with version 1 diffuser, I was shooting at 1/10 or so. With my USAF 1951 test slide, I measure that this was resolving 30 cycles/mm. And, we saw that it didn't resolve the dye clouds.

With version 2, described above, I'm shooting at 1/125 to 1/250 and I'm using the APO Rodagon D 1:1 copy lens, Nikon PB4 bellows on a tripod, and a full-frame Sony A7. Measuring the test slide, I now get 64 to 72 cycles/mm, I suspect limited by my sensor pitch. I measure the same in my Carousel projector and with an alternate rock-steady rig.

What's the source of the difference? Lens? Vibration causing blur at longer shutter speeds? A bit hard to tell, but I suspect the lens. I doubt that the Nikkor at 1:1.4 on AF is as good as the Rodagon D at 1:1 carefully focused.

 

[ColSebastianMoran]

Here's a comparison at 2:1 in Lightroom.
- LEFT: Rodagon D, Sony A7, 1/200th, ISO 100, new diffuser
- RIGHT: 105mm Micro Nikkor, D7200, 1/10th, ISO 100, old plastic diffuser

Note difference in sky (dye clouds) and building (windows resolved).

A couple of conclusions:
- 24MPx resolves enough of the dye clouds; I doubt additional pixels would add photographic info for this film (Kodachrome 64).
- Post processing needed. Blur the sky? This combination resolves a lot more dirt and fine fibers. Clean slides before camera-scan?
- Carousel projector scan is very fast. 140 slides in under ten minutes.
- Might be grain aliasing in the left image??

Thanks, @dmr, for pushing me to look more closely at resolution.

 

[ColSebastianMoran]

Another comparison, here is a 2:1 screen shot of the sky in the Tetons image, upper right in full image:
- LEFT: Rodagon D, carefully focused, as in previous item
- RIGHT: 105mm Micro Nikkor AF

 

[ColSebastianMoran]

And here's the Tetons image from the new Carousel rig. I have blurred the sky, so don't look for dye clouds here. Kodachrome 64, Nikon gear, hand-held. I'm sure this will make a good print. Screen resolution here:

And, here's a link to the full 24MPx image (12 MB).

I'm satisfied that I can make good camera-scans of chromes with this rig.

 

[predicolous]

This is awesome! The new results really outperform the older lens.
Love the speed of digitizing you mentioned!!!

 

[ColSebastianMoran]

This is awesome! The new results really outperform the older lens.
Love the speed of digitizing you mentioned!!!

Yes, the throughput is fabulous. Then, there's a bit of post...

Curious about the 105 Nikkor, it's supposed to be better.

 

[ColSebastianMoran]

Here is the inside of the rig, showing the position of the reflector, the removed heat-absorbing glass, and the condensor (left in place):

Bulb is a SORAA 3000°K 95CRI "Vivid Light" MR-16 12v bulb. Bulb options are changing fast; I expect to switch to 4000°K at next iteration.

 

[ColSebastianMoran]

More learning:
- With higher resolution rig, easy to get grainy, gritty images. Solution: turn-down or turn-off capture sharpening in LR, increase NR, and start at Clarity=0.
- Slides kept in a carousel tray and it's cardboard storage box for 20 years have myriad dust and microscopic fibers; clean first or do a lot of spotting in post. I'm suspicious that cardboard mounts are worse than plastic.
- Focus once, shoot many slides, the projector positions slides accurately for focus.
- However, slides may fall slightly differently left-right. Either move in a bit and lose edges of all slides, or include edges and crop in post.

 

[ColSebastianMoran]

Back to the Teton's image from #66 above.

I just had it printed at 20x30" at Costco, and I'm very happy with the print.

This print from camera scan is far better than any I ever got from mid-level labs back in the day.

Folowing cues from Bruce Fraser:
- Camera Scan with 24MPx Sony A7, no capture sharpening
- Noise reduction first
- Then Focus Magic (Richardson-Lucy deconvolution), masked to non-sky
- Gaussian Blur in sky
- Further adjustments in LR
- Output sharpening at LR defaults

The final