Recommendations for Scanning Slides

Have a large number of 35mm slides and negs to digitize; wonder what works best?  Looking at Kodak digitizer for around $150.  Want simple, fast, accurate images that could be printed up to 8 x 10.

Flatbed scanner with a backlight, like the Epson V600.

Of course my biggest tip is to carefully and hyper-critically cull the selection before the tedious process of prepping and cleaning the slides for scanning.

Thanks!  I just solved my problem ( I think).  I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2  mounted slide.  I can just carefully slide the slide in under an edge of the tape.  I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface.  Two second delay after a shutter press and the image is captured.  So far I am really impressed with the detail still preserved in these vintage film images.  I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance.  So far, so good and have done about 100 slides.

Glad to hear it.

paul wassermann wrote:

Thanks! I just solved my problem ( I think). I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2 mounted slide. I can just carefully slide the slide in under an edge of the tape. I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface. Two second delay after a shutter press and the image is captured. So far I am really impressed with the detail still preserved in these vintage film images. I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance. So far, so good and have done about 100 slides.

Pretty close to what I was about to suggest (Duplihood ).

However, try putting your A9 into multi-shot auto HDR mode. Slides commonly have about 2 stops of extra DR that is encoded with wildly wrong (too dark) tonality, making it invisible to humans viewing the projected slide, but the huge DR of the auto HDR mode and tone mapping it applies often can pull those shades back into useful scene detail.

ProfHankD wrote:

paul wassermann wrote:

Thanks! I just solved my problem ( I think). I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2 mounted slide. I can just carefully slide the slide in under an edge of the tape. I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface. Two second delay after a shutter press and the image is captured. So far I am really impressed with the detail still preserved in these vintage film images. I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance. So far, so good and have done about 100 slides.

Pretty close to what I was about to suggest (Duplihood ).

However, try putting your A9 into multi-shot auto HDR mode. Slides commonly have about 2 stops of extra DR that is encoded with wildly wrong (too dark) tonality, making it invisible to humans viewing the projected slide, but the huge DR of the auto HDR mode and tone mapping it applies often can pull those shades back into useful scene detail.

Thank you for the suggestion...I'll give it a try tomorrow!

I used an Epson scanner with slide and negative scanning attachments to scan slides and negatives. Took a good amount of time, but now I have digital copies of all my slides and negatives.

Biggest issue: dust and other debris. I understand that this can be largely fixed in software for color film/slides if you have a scanner that has an infrared channel. So I'd highly recommend that.

iljitsch wrote:

I used an Epson scanner with slide and negative scanning attachments to scan slides and negatives. Took a good amount of time, but now I have digital copies of all my slides and negatives.

Biggest issue: dust and other debris. I understand that this can be largely fixed in software for color film/slides if you have a scanner that has an infrared channel. So I'd highly recommend that.

Flatbed scanners tend to have more of a dust issue than other methods of slide scanning -- after all, they have a piece of glass fixed horizontally near the focus plane and it can get dusty.

The dust/srcatch removal is generally done by capturing an NIR image to use as a mask. The dark pixels in NIR are simply interpolated over. A better fix would be to do true inpainting to fill the gaps. I was thinking I could write code to automatically recognize and fix the dust spots without NIR, but it's been done: Automatic Sensor Dust Removal . Yes, that's essentially the same problem as dust on scanned film. Anyway, they link to python OpenCV code that finds the spots & inpaints to remove 'em: Python-Automatic-Sensor-Dust-Removal .

ProfHankD wrote:

iljitsch wrote:

I used an Epson scanner with slide and negative scanning attachments to scan slides and negatives. Took a good amount of time, but now I have digital copies of all my slides and negatives.

Biggest issue: dust and other debris. I understand that this can be largely fixed in software for color film/slides if you have a scanner that has an infrared channel. So I'd highly recommend that.

Flatbed scanners tend to have more of a dust issue than other methods of slide scanning -- after all, they have a piece of glass fixed horizontally near the focus plane and it can get dusty.

The dust/srcatch removal is generally done by capturing an NIR image to use as a mask. The dark pixels in NIR are simply interpolated over. A better fix would be to do true inpainting to fill the gaps. I was thinking I could write code to automatically recognize and fix the dust spots without NIR, but it's been done: Automatic Sensor Dust Removal . Yes, that's essentially the same problem as dust on scanned film. Anyway, they link to python OpenCV code that finds the spots & inpaints to remove 'em: Python-Automatic-Sensor-Dust-Removal .

I was able to digitize all my slides by photographing them with my Sony 90mm macro lens using a backlit drawing tablet.  I used masking tape to make a top and bottom little ledge that would hold the slide.  Blew compressed air on the slides and wiped with microfiber cloth and then photographed.  They came out great...surprised how much detail was present in the underexposed areas of the slides.

DoctorCamel wrote:

https://vimeo.com/12405612

Thanks!  My system (excluding the tripod, camera, and macro lens) cost about $50.  I set exposure to manual, f11 aperture priority, ISO 100 and AWB.  Used two second delay with remote trigger.  I got the light pad on Ebay used and it has really met my needs.

ProfHankD wrote:

paul wassermann wrote:

Thanks! I just solved my problem ( I think). I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2 mounted slide. I can just carefully slide the slide in under an edge of the tape. I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface. Two second delay after a shutter press and the image is captured. So far I am really impressed with the detail still preserved in these vintage film images. I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance. So far, so good and have done about 100 slides.

Pretty close to what I was about to suggest (Duplihood ).

However, try putting your A9 into multi-shot auto HDR mode. Slides commonly have about 2 stops of extra DR that is encoded with wildly wrong (too dark) tonality, making it invisible to humans viewing the projected slide, but the huge DR of the auto HDR mode and tone mapping it applies often can pull those shades back into useful scene detail.

Alternatively, just bracket raw and feed the results to hdrmerge - https://github.com/jcelaya/hdrmerge - rather than trust whatever the camera does internally as far as tonemapping.

Entropy512 wrote:

ProfHankD wrote:

paul wassermann wrote:

Thanks! I just solved my problem ( I think). I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2 mounted slide. I can just carefully slide the slide in under an edge of the tape. I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface. Two second delay after a shutter press and the image is captured. So far I am really impressed with the detail still preserved in these vintage film images. I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance. So far, so good and have done about 100 slides.

Pretty close to what I was about to suggest (Duplihood ).

However, try putting your A9 into multi-shot auto HDR mode. Slides commonly have about 2 stops of extra DR that is encoded with wildly wrong (too dark) tonality, making it invisible to humans viewing the projected slide, but the huge DR of the auto HDR mode and tone mapping it applies often can pull those shades back into useful scene detail.

Alternatively, just bracket raw and feed the results to hdrmerge - https://github.com/jcelaya/hdrmerge - rather than trust whatever the camera does internally as far as tonemapping.

Lots of tools can do HDR processing, but the Sony in-camera version actually does surprisingly better than most in this case, which is an odd case because the extra info isn't with appropriate brightness to scale normally. The tone mapping just happens to do well with it.

The other issue is that you might need as much as 6 stop HDR coverage, and bracketing on many cameras doesn't handle that well. It also means a pile of raws because it's best done with more than two shots. Still, you are absolutely right that you can do a little better with raw captures and really careful tweaking of the tone mapping.... 

ProfHankD wrote:

Entropy512 wrote:

ProfHankD wrote:

paul wassermann wrote:

Thanks! I just solved my problem ( I think). I have a light tablet, have made a painter's tape rectangle just slightly smaller than a 2 x 2 mounted slide. I can just carefully slide the slide in under an edge of the tape. I set up my Sony a9 with 90mm macro lens on a small tripod with the camera carefully positioned square with the slide surface. Two second delay after a shutter press and the image is captured. So far I am really impressed with the detail still preserved in these vintage film images. I am able to clean, load, and photograph a slide in about 30 seconds; the camera does a wonderful job selecting exposure and white balance. So far, so good and have done about 100 slides.

Pretty close to what I was about to suggest (Duplihood ).

However, try putting your A9 into multi-shot auto HDR mode. Slides commonly have about 2 stops of extra DR that is encoded with wildly wrong (too dark) tonality, making it invisible to humans viewing the projected slide, but the huge DR of the auto HDR mode and tone mapping it applies often can pull those shades back into useful scene detail.

Alternatively, just bracket raw and feed the results to hdrmerge - https://github.com/jcelaya/hdrmerge - rather than trust whatever the camera does internally as far as tonemapping.

Lots of tools can do HDR processing, but the Sony in-camera version actually does surprisingly better than most in this case, which is an odd case because the extra info isn't with appropriate brightness to scale normally. The tone mapping just happens to do well with it.

The other issue is that you might need as much as 6 stop HDR coverage, and bracketing on many cameras doesn't handle that well. It also means a pile of raws because it's best done with more than two shots. Still, you are absolutely right that you can do a little better with raw captures and really careful tweaking of the tone mapping....

6 stop not handled well by most cameras?

I'm not sure of any camera that doesn't offer 5-shot 2EV steps (8EV between maximum and minimum), or 3-shot 3EV (exactly 6EV total)

I guess maybe I'm spoiled by Sonys, even really old Sony bodies (like the A6000) did 3-shot with 3EV steps per shot?

paul wassermann wrote:

Have a large number of 35mm slides and negs to digitize; wonder what works best? Looking at Kodak digitizer for around $150. Want simple, fast, accurate images that could be printed up to 8 x 10.

When I scanned a couple thousand slides I used a slightly older (now) flatbed that can hold a number of slides (I don't remember number) at a time and the software worked pretty well.

I did consider having it done by a service but that was pretty costly at the time, maybe a decade ago. Back in win 7 days

I also did a lot of negative scans too and again worked pretty well but it was a time commitment