November 21, 2017, 12:44:36 PM

Author Topic: How to scan slides  (Read 18136 times)

Mikehit

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 2156
Re: How to scan slides
« Reply #30 on: August 02, 2016, 05:39:28 AM »
I am not sure how the software works. When it is scanning I guess it is identifying dust particles or scratches by detecting a discrepancy in transmission (?) of light through the slide substrate. But I was wondering if you could apply the programme to a scan of a slide and 'remove' dust or scratches after the fact. If so, then you could photograph a dirty/scratched slide and run VueScan over it.
If it measuring the slide with a combination of visible light for the scan and IR to detect dust/scratches I guess it is not possible to apply it afterwards.

AFAIK the scanner with "hardware" dust/scratch removal uses the different data they get from visible and IR light to determine what is dust and what is not. It doesn't work with B/W film because the silver particles doesn't behave like the dyes in colour film (IIRC silver is removed from colour film during the bleach/fix phase).

This should work better than analysing the image pixels and trying to detect what is a scratch/dust, because it is based on different physical properties, but the software needs the data from the scanner which are not saved within the final image. Thereby, without the data IMHO the software can only try to detect scratch/dust by analysing pixels and their surroundings - which may work less reliably.

That makes sense. Thank you again.

I may move to a two-stage process. Photograph them first off and any with dust/scratch problem I then scan them.

canon rumors FORUM

Re: How to scan slides
« Reply #30 on: August 02, 2016, 05:39:28 AM »

kaihp

  • EOS 5DS R
  • ******
  • Posts: 782
Re: How to scan slides
« Reply #31 on: August 03, 2016, 04:20:51 PM »
I went to the forum thread that pwp linked to and saw the Canon 9000F flatbed scanner mentioned. I have one of those, and want to warn against it: just don't waste your time (and money) on it for scanning slides/negatives.

The fundamental problem with the 9000 (and likely others too; do your own research) is that the scanner is focused on the glass, but the negative/slide holder has a finite thickness which means that the negatives/slides are 0.5mm displaced from the scanners focus plane. The resulting IQ is pathetic and even when scanning at the highest resolution (3600dpi or so), the IQ is significantly worse than in the scans I had made (~1200dpi) when I developed the film originally in 2001.

Other flatbed scanners may or may not have this problem, but if I ever start scanning the old family photos, I'll be using my 5D3 and a macro lens for the capture process.

Nonsense... I have both the Epson V850 and Canon 9000II both work quite well. One cost 160$ the other almost 900$ One from each (both low res)... Which one is from the Epson and Canon? The only major differences I can tell you is that the 850 will scan large format, the 9000 only up to medium format and the 850 will wet scan with less mess than the 9000. Besides that not much but cost.

The biggest hurdle to scanning is knowing how to scan to get the best results...

Pookie,

All I'm doing is reporting my experience (as limited as it is). I tried both the Canon and the SilverScan software, both times stepping away (very) disappointedly. I'm sure that if I spent the proverbial 10.000 hours on learning how to scan slides and negatives optimally I'd get a much better result, but with the initial results being so appalling as they were for me ... well, then I won't be encouraged to do so.

You sound like you have a lot of experience in scanning. How much of a learning curve was it to get to an acceptable/reasonable level for you?

If it's OK with you, I'd like to get back to this topic and post an example in 1-2 weeks time, as my schedule is getting pretty crazy just now.

dcm

  • EOS 5D Mark IV
  • ******
  • Posts: 660
  • It's not the gear.
Re: How to scan slides
« Reply #32 on: August 20, 2016, 01:30:55 PM »
Had a chance to compare scanning versus photographing slides recently to get higher resolution images.  Here's 6 samples using:
  • Plustek 7600i at 7200dpi (9624x6504)
  • Epson V750 at 6400dpi (8560x5795)
  • Canon EOS 6D/100L Macro (5472x3648)
  • Canon EOS 1DX2/100L Macro (5472x3648)
  • Canon EOS M3/28 Macro
  • Canon EOS M/28 Macro

I used this setup with the 100L.  I had to add a 67-52 step down with 35mm of extensions to the Polaroid slide copier to reach MFD on the 100L.  I removed the long tube to reach MFD with the 28mm on the Ms.  I used a Huion LED Light Pad as the light source and shot at f8 to minimize vignetting.


q60k3target-5 by dvmtthws, on Flickr

I used a standard Kodak Kodachrome target for comparison.  I captured raw images from all 6 and converted to JPGs in LR with no processing.  Can you tell which is which without following the links to the full size originals?


q60k3target-1 by dvmtthws, on Flickr


q60k3target-2 by dvmtthws, on Flickr


q60k3target-3 by dvmtthws, on Flickr


q60k3target-4 by dvmtthws, on Flickr


q60k3target-6 by dvmtthws, on Flickr


q60k3target-7 by dvmtthws, on Flickr

For scanning slides, the Plustek is slower overall than the Epson since it requires manual intervention between each slide and it doesn't handle anything but 35mm.  For low volume this is a workable solution.  There are newer options that can handle a variety of 35mm or smaller formats but I don't have any experience with them.

For photographing slides a FF with 100L is a good approach and the M's with the 28 also looks good.  A crop DSLR the EF-S 60 Macro might be a viable alternative to achieve the same effect, I just don't have one to test.
« Last Edit: August 20, 2016, 07:10:05 PM by dcm »
Shooting with Canon since 1982 - 1DX2, M5, some lenses

LDS

  • EOS 5DS R
  • ******
  • Posts: 1041
Re: How to scan slides
« Reply #33 on: August 21, 2016, 09:59:57 AM »
Had a chance to compare scanning versus photographing slides recently to get higher resolution images. 

The issue is these targets are mostly designed for colour calibration - they don't tell much about the resolving capabilities of the scanner (but maybe in the face picture, still designed to assess skin tones, not resolution). And in fact the images show the different colour rendition of the technology used - and in some images is evident the bluish cast Kodachrome may show in some scanners.

You would need something alike the 1951 USAF resolution test chart to measure resolving power - as far as I know, most scanners are evaluated using something alike.

Pookie

  • EOS 5DS R
  • ******
  • Posts: 854
  • You talk, I show... who do you believe?
Re: How to scan slides
« Reply #34 on: August 21, 2016, 12:17:38 PM »
I went to the forum thread that pwp linked to and saw the Canon 9000F flatbed scanner mentioned. I have one of those, and want to warn against it: just don't waste your time (and money) on it for scanning slides/negatives.

The fundamental problem with the 9000 (and likely others too; do your own research) is that the scanner is focused on the glass, but the negative/slide holder has a finite thickness which means that the negatives/slides are 0.5mm displaced from the scanners focus plane. The resulting IQ is pathetic and even when scanning at the highest resolution (3600dpi or so), the IQ is significantly worse than in the scans I had made (~1200dpi) when I developed the film originally in 2001.

Other flatbed scanners may or may not have this problem, but if I ever start scanning the old family photos, I'll be using my 5D3 and a macro lens for the capture process.

Nonsense... I have both the Epson V850 and Canon 9000II both work quite well. One cost 160$ the other almost 900$ One from each (both low res)... Which one is from the Epson and Canon? The only major differences I can tell you is that the 850 will scan large format, the 9000 only up to medium format and the 850 will wet scan with less mess than the 9000. Besides that not much but cost.

The biggest hurdle to scanning is knowing how to scan to get the best results...

Pookie,

All I'm doing is reporting my experience (as limited as it is). I tried both the Canon and the SilverScan software, both times stepping away (very) disappointedly. I'm sure that if I spent the proverbial 10.000 hours on learning how to scan slides and negatives optimally I'd get a much better result, but with the initial results being so appalling as they were for me ... well, then I won't be encouraged to do so.

You sound like you have a lot of experience in scanning. How much of a learning curve was it to get to an acceptable/reasonable level for you?

If it's OK with you, I'd like to get back to this topic and post an example in 1-2 weeks time, as my schedule is getting pretty crazy just now.

The problem with that comment is your telling others that that equipment is garbage and not to waste their time. You make a completely false comment about the way slides are scanned like it is a "fact". Only to backtrack and explain that in reality you haven't actually spent much time trying to figure it out with the qualifier "as limited as [experience] is".

My point is, they work fine. I use them often and it doesn't take much time to figure it out. I won't even get into the "I'll just shoot with a 100L and get good enough copies...". If you scan often and know how to use the equipment you'll find you can get much much more resolution than any camera will and huge 1.5 GB files that make pictures of slide look like an absolutely ridiculous compromise.
« Last Edit: August 21, 2016, 12:50:55 PM by Pookie »
I'm limping by with my current equipment... once I get that new lens with IS and blue goo... then I'll finally be able to go out and take my first decent picture...

dcm

  • EOS 5D Mark IV
  • ******
  • Posts: 660
  • It's not the gear.
Re: How to scan slides
« Reply #35 on: August 25, 2016, 12:08:56 AM »
Had a chance to compare scanning versus photographing slides recently to get higher resolution images. 

The issue is these targets are mostly designed for colour calibration - they don't tell much about the resolving capabilities of the scanner (but maybe in the face picture, still designed to assess skin tones, not resolution). And in fact the images show the different colour rendition of the technology used - and in some images is evident the bluish cast Kodachrome may show in some scanners.

You would need something alike the 1951 USAF resolution test chart to measure resolving power - as far as I know, most scanners are evaluated using something alike.

Agreed, but it depends on what you are looking for.  I didn't have the USAF 1951 slide handy at the time, but I managed to test again with a LaserSoft 35mm version of the test chart.  I've seen suggestions that it doesn't have the quality of the glass test slides, but it was good enough for this comparison.  I took the best of the images using live view and different AF settings for each device.  Here's a crop of the central areas for each camera - you'll need to visit the original to view the 1:1 images. 


epson 6400 by dvmtthws, on Flickr


plustek 7200 by dvmtthws, on Flickr


M by dvmtthws, on Flickr


M3 by dvmtthws, on Flickr


M5 by dvmtthws, on Flickr (added 2017-08-16)


6D by dvmtthws, on Flickr


1DX2 by dvmtthws, on Flickr

These jpgs were created by LR from the original raw files.  Only adjustments were a -1.1 exposure on the Plustek since it serverely overexposed the slide.  I took a look at the original DNG and CR2 files in ACR and used a chart LaserSoft provides showing the actual usable resolution of the device in dpi based on the smallest group and element you can clearly discern.  Here's what I observed (YMMV) along with the usable/actual ratio for comparison:

Epson 6400 - group 5 / element 4 - 2299 dpi - 36%
Epson 3200 - group 5 / element 3 - 2048 dpi - 64%
Epson 1600 - group 4 / element 6 - 1448 dpi - 91%
Plustek 7200 - group 6 / element 2 - 3649 dpi - 51%
Plustek 3600 - group 5 / element 6 - 2896 dpi - 72%
Plustek 2400 - group 5 / element 1 - 1626 dpi - 67%
Plustek 1800 - group 4 / element 6 - 1448 dpi - 80%
M 5904*- group 6 / element 1 - 3251 dpi - 55%
M3 6834* - group 6 / element 3 - 4096 dpi - 60%
M5 6834* - group 6 / element 3 - 4096 dpi - 60%  (added 2017-08-16)
6D 3860* - group 5 / element 6 - 2896 dpi - 75%
1DX2 3860* - group 5 / element 6 - 2896 dpi - 75%

* approximate dpi of the sensor

If you look at the other scans for the Epson and Plustek in the album you will note that there is little difference at lower resolutions.  The Epson 1600 resolves as well as the Plustek 1800.  You quickly get diminishing returns as as the file size grows much faster than the usable resolution. 

The cameras were all tested at 1:1 so the APS-C cameras cropped the slide.  Still, I'm impressed with the 28mm Macro on the M's.  Didn't expect much difference between the 6D and 1DX2 since they have the same sensor resolution and were using the same lens.

Not sure there is a clear winner when everything (ease of use, efficiency, etc.) is considered or if this helps anyone considering whether to scan or make photo copies, but it's food for thought.

EDIT:  Added dpi estimate for cameras and a usable/actual ratio for comparison
EDIT:  Added M5 results (same as M3).
« Last Edit: August 16, 2017, 10:21:00 PM by dcm »
Shooting with Canon since 1982 - 1DX2, M5, some lenses

LDS

  • EOS 5DS R
  • ******
  • Posts: 1041
Re: How to scan slides
« Reply #36 on: August 26, 2016, 04:39:01 PM »
Agreed, but it depends on what you are looking for.  I didn't have the USAF 1951 slide handy at the time, but I managed to test again with a LaserSoft 35mm version of the test chart.  I've seen suggestions that it doesn't have the quality of the glass test slides, but it was good enough for this comparison.  I took the best of the images using live view and different AF settings for each device.  Here's a crop of the central areas for each camera - you'll need to visit the original to view the 1:1 images. 

Very nice work, thank you.

canon rumors FORUM

Re: How to scan slides
« Reply #36 on: August 26, 2016, 04:39:01 PM »

dcm

  • EOS 5D Mark IV
  • ******
  • Posts: 660
  • It's not the gear.
Re: How to scan slides
« Reply #37 on: August 27, 2016, 12:24:39 AM »
Agreed, but it depends on what you are looking for.  I didn't have the USAF 1951 slide handy at the time, but I managed to test again with a LaserSoft 35mm version of the test chart.  I've seen suggestions that it doesn't have the quality of the glass test slides, but it was good enough for this comparison.  I took the best of the images using live view and different AF settings for each device.  Here's a crop of the central areas for each camera - you'll need to visit the original to view the 1:1 images. 

Very nice work, thank you.

Thanks.  I'm no Roger and I lack his equipment but I think the home office test gives a reasonable estimation of what you will see when making digital copies of slides.

Unfortunately I don't have access to a 5Ds or 5DsR.  Based on pixel density I'd expect the 5Ds to perform similar to the M3 and the 5DsR a bit better, but hard to tell how much.  It would be nice to test the hypothesis.
Shooting with Canon since 1982 - 1DX2, M5, some lenses

Pookie

  • EOS 5DS R
  • ******
  • Posts: 854
  • You talk, I show... who do you believe?
Re: How to scan slides
« Reply #38 on: August 16, 2017, 04:58:14 PM »
Quote
Not sure there is a clear winner when everything (ease of use, efficiency, etc.) is considered or if this helps anyone considering whether to scan or make photo copies, but it's food for thought.



I have been doing a lot of scanning this year. I own a V850, Canon 9000 II, and a Pacific Image Primefilm XA. I scan slides, negatives in 4x5, 120, and 35mm. In addition to Kinetronics brushes, air cans, linen gloves... basically all the accesories needed for proper scans. I will often get Frontier drum scans for client work in 120 or 4x5.

I just started using a Canon 5D IV and a 100mmL... I'd honestly say there is little difference in the results but for two major exceptions.

Obviously the Frontier scanner (or Noritsu) is the best for large scans suitable for the highest reproductions and you pay for this service.

In regards to in-home work though you can get decent results with a camera and a sharp lens the problem though is where you want to spend your time if quality is critical...( even with a home scan). A camera will make a perfectly good scan but if you have ANY corrections needed to the source (scratches, dust, hair, etc) you will spend an extraordinarily long time cleaning the image. The scanner on the other hand will auto correct these in color images (not BW) through IR scanning and masking. With BW scanning software can do a great job automating this process although not as good a IR. I use VueScan as a third party program.

You could spend a minute or two setting up your slide and snapping the picture... then spend 5-20 minutes correcting the defects in LR or PS. I often find the scratch and dust detection woefully inadequate and often have to resort to healing brushes. Or you could spend 4-6 minutes scanning both visible and IR... this happens while you wait hand free.

It's really all dependent on your output needs... if you don't care and are just documenting then a Rube Goldberg setup will work perfectly fine... if you want very clean files, scanners are the way to go.

Epson... full scan using VueScan ( 3.5 minute scan ) minimal hands on PP


Canon 5D IV with 100mm L... ( quite  a bit of work to get it clean, about 10 minutes total)
I'm limping by with my current equipment... once I get that new lens with IS and blue goo... then I'll finally be able to go out and take my first decent picture...

canon rumors FORUM

Re: How to scan slides
« Reply #38 on: August 16, 2017, 04:58:14 PM »