Dusting off old 35mm slides and scanning

[Chuck Nacke]

20 hours ago, Chuck Nacke said:

Michael,

 

Found a set of actions in PS that appear to solve my magenta scan problem,  make a layer, blur, invert, overlay, blend, adjust opacity.

According to my 14 year-old daughter, who has much better color perception than I, it works perfectly.  The sliders have never solved 

the problem.  Have a new scan going up in the morning, from RDP, and the color should be fine?

 

I do appreciate you making it known to me to find a solution to my magenta problem with scans.

 

Thank you again,

 

Chuck

Got to go skiing.  See ya all.

 

Chuck

[M.Chapman]

22 hours ago, wiskerke said:

 

For that I use 2 mirrors. One is a 99 cts small beauty mirror that happens to fit in a 67mm filter ring. The other an Ikea tile. The round beauty mirror (non enlarging) has a small part that's scratched blank exactly in the middle. The filter ring is painted white for better judging the alignment. The tile is where the negative or slide would be. This is how in the olden days camera technicians aligned process cameras in red (printers) darkrooms. Or how one would align the view camera. 

 

http://www.montagu-pollock.co.uk/blog/files/_gp39829s.jpg

Out of alignment.

 

http://www.montagu-pollock.co.uk/blog/files/_gp39826s.jpg

In alignment. (via)

 

Neat, I hadn't thought of using two mirrors for increased accuracy. I just use one, a perspex mirror that is 50mm x 50mm (https://www.ebay.co.uk/itm/362429329957) which fits onto the 3 pin slide mount I made on top of my lightbox (iPad + Diffuser). I've got 2 mirrors so I'll drill a hole in the middle of the second one (or maybe buy a round one https://www.ebay.co.uk/itm/282292974306) and give the two mirror method a try.

 

Thanks.

 

Mark

Edited March 19, 2022 by M.Chapman

[spacecadet]

14 hours ago, Harry Harrison said:

it stays in the cassette and it needs a special mechanism to feed it out.

So it does, forgot about that. We find the odd APS cassette in the outer limits of the rooms the children left 20 years ago, they owed no allegiance to 35mm. of course.

Remember how you had to lever cassettes of Tri-X open to process it, instead of just whacking it on the bench like Ilford? I daresay you could do that. But I think these were unexposed films (according to the nifty barberpole-type code in the mechanism) and fresh (ok, stale) APS film is as much use as a bowler hat with sleeves now. They're probably in the freezer now. For the archaeologists.

[wiskerke]

2 hours ago, M.Chapman said:

Neat, I hadn't thought of using two mirrors for increased accuracy. I just use one, a perspex mirror that is 50mm x 50mm (https://www.ebay.co.uk/itm/362429329957) which fits onto the 3 pin slide mount I made on top of my lightbox (iPad + Diffuser). I've got 2 mirrors so I'll drill a hole in the middle of the second one (or maybe buy a round one https://www.ebay.co.uk/itm/282292974306) and give the two mirror method a try.

 

Thanks.

 

Mark

 

If you take an ordinary glass mirror, just scratch off a small part of the backside. I tape a washer to the back, exactly in the middle, scraping away the center. I scratch off the backing using a rust removing pencil from the car stuff department. But a hobby knife or a razor blade with not much force will work also. If you happen to have a mirror that's reflective on both sides, rare nowadays, use the side where the reflective stuff sits as the front side. You then have a first surface mirror or front surface mirror; FS or FSM for short.  The difference is very marginal at best, but it feels more professional 😁. (When using a mirror as an imaging device this does make all the difference, but not here.)

For clarity in use, I paint the rim of the filter holder white. Any bright color will do. In some cases a black rim works better, like when reproducing art works in a very bright environment.

I would use a glass mirror for the big one too. Even glass mirrors bend when you put any strain on them, like with the use of rubber bands, like many instructions (on the internet) will have you do. Which was how I have learned it as well, until I noticed that my results were skewed (pun intended). Btw this is why filters ought to sit loose in their fittings.

 

wim

 

edit: Scratching will not work on perspex. Or far too good actually: you'll end up with a matte circle in the middle.

Edited March 19, 2022 by wiskerke

[M.Chapman]

I think small (50mm) diameter surface coated plastic mirrors will be flat and rigid enough (they are laser cut from large flat 3mm thick sheet  not injection moulded). If I was copying large format or above, I'd go with glass. Using perspex will also allow me to drill the hole. I must do the sums and see what the alignment tolerance is to keep circle of confusion below 1 pixel on a 20MP MFT sensor using a 45mm lens at f/5.6 and filling the frame with a 35mm slide.

 

Mark

Edited March 19, 2022 by M.Chapman

[Martyn]

4 hours ago, spacecadet said:

So it does, forgot about that. We find the odd APS cassette in the outer limits of the rooms the children left 20 years ago, they owed no allegiance to 35mm. of course.

Remember how you had to lever cassettes of Tri-X open to process it, instead of just whacking it on the bench like Ilford? I daresay you could do that. But I think these were unexposed films (according to the nifty barberpole-type code in the mechanism) and fresh (ok, stale) APS film is as much use as a bowler hat with sleeves now. They're probably in the freezer now. For the archaeologists.

 

I have digitised some old APS negatives for personal use, not for Alamy ... insert a small thin jewellers screwdriver into one of the recessed holes on the cassette edge either end of the light trap, depress the screwdriver whilst putting another one into the hole in the centre of the cassette end, turn and wind out the film !

[M.Chapman]

18 hours ago, M.Chapman said:

I think small (50mm) diameter surface coated plastic mirrors will be flat and rigid enough (they are laser cut from large flat 3mm thick sheet  not injection moulded). If I was copying large format or above, I'd go with glass. Using perspex will also allow me to drill the hole. I must do the sums and see what the alignment tolerance is to keep circle of confusion below 1 pixel on a 20MP MFT sensor using a 45mm lens at f/5.6 and filling the frame with a 35mm slide.

 

Mark

On the subject of alignment - I did some basic sums using simple lens theory.

 

I'm using a 20MP MFT sensor which has a pixel size around 3.3um to digitise 35mm Fuji Velvia slides.

I'm also using a macro lens with a focal length of 45mm used at f/5.6 (its sweet spot).

The magnification of the setup is 0.5x (MFT sensor is 1/2 FF)

 

Wikipedia states Depth of Focus = 2 x Aperture x CofC x (1 + magnification)

 

If I set the acceptable Circle of Confusion (CofC) to 1 sensor pixel width (3.3um) the depth of focus at the sensor is about 0.056mm and the depth of field at the slide film plane is 0.22mm. This means that the slide film has to be parallel to the film sensor within 0.22mm over the 43mm slide diagonal which is about 0.3 degrees.

 

But... this is potentially overkill. Fuji Velvia 50 is reported* as having a maximum resolving capability of 80 line pairs / mm So each line pair is 12.5um wide and the width of a line (or a space) is 6.25um.

 

If I set the acceptable CofC to 6.25um the depth of focus at the sensor is about 0.105mm and the depth of field at the slide film plane is 0.42mm. This means that the slide film has to be parallel to the film sensor within 0.42mm over the 43mm slide diagonal which is about 0.56 degrees**

 

*Film resolutions are given in Table 3 here https://cool.culturalheritage.org/videopreservation/library/film_grain_resolution_and_perception_v24.pdf

 

**Note the above assumes no other sources of error. If the film is not flat or the lens shows field curvature, these will eat into the available tolerances.

 

Mark (hoping I've got the maths right and not made any typos) 

 

Edited March 20, 2022 by M.Chapman

[John Mitchell]

On 18/03/2022 at 14:53, Chuck Nacke said:

Got to go skiing.  See ya all.

 

Chuck

 

Don't break a leg, Chuck. My ageing Super Coolscan 4000 ED also now produces scans with a strong magenta cast. I posted about it on the forum awhile back. Here's a link to the thread:

 

https://discussion.alamy.com/topic/11409-nikon-scanner-question/#comment-207680

 

P.S. I haven't gotten around to cleaning the mirror yet as some suggested in the thread. However, I plan to do so.

 

 

 

[spacecadet]

11 hours ago, John Mitchell said:

 

Don't break a leg, Chuck. My ageing Super Coolscan 4000 ED also now produces scans with a strong magenta cast. I posted about it on the forum awhile back. Here's a link to the thread:

 

https://discussion.alamy.com/topic/11409-nikon-scanner-question/#comment-207680

 

P.S. I haven't gotten around to cleaning the mirror yet as some suggested in the thread. However, I plan to do so.

 

 

 

No hurry...........only 3 years ago.

[wiskerke]

15 hours ago, John Mitchell said:

 

Don't break a leg, Chuck. My ageing Super Coolscan 4000 ED also now produces scans with a strong magenta cast. I posted about it on the forum awhile back. Here's a link to the thread:

 

https://discussion.alamy.com/topic/11409-nikon-scanner-question/#comment-207680

 

P.S. I haven't gotten around to cleaning the mirror yet as some suggested in the thread. However, I plan to do so.

 

That link to the routine is still there and still valid:

http://www.shtengel.com/gleb/Nikon_5000_mirror_cleaning.htm

 

wim

[John Mitchell]

7 hours ago, spacecadet said:

No hurry...........only 3 years ago.

 

My how time flies. Guess I could blame it on COVID. Now with WW III looming, I might have a new excuse to procrastinate.

[Chuck Nacke]

23 hours ago, John Mitchell said:

 

Don't break a leg, Chuck. My ageing Super Coolscan 4000 ED also now produces scans with a strong magenta cast. I posted about it on the forum awhile back. Here's a link to the thread:

 

https://discussion.alamy.com/topic/11409-nikon-scanner-question/#comment-207680

 

P.S. I haven't gotten around to cleaning the mirror yet as some suggested in the thread. However, I plan to do so.

 

 

 

John,

 

Found this and it appears to work for me : 

 It appears to work quite well.  Last image that I posted, original was shot on 35mm EPP.  I have not tried it on a K-14 scan yet.

 

I wish I had a Nikon 5000 scanner, but.

 

P.S. Skiing sucked this weekend,  The season in New Hampshire, USA Is over.  Time to think about 2022/23 season, if we are all still 

here?

 

Best,

 

Chuck

 

[MDM]

10 hours ago, Chuck Nacke said:

John,

 

Found this and it appears to work for me : 

It appears to work quite well.  Last image that I posted, original was shot on 35mm EPP.  I have not tried it on a K-14 scan yet.

 

I wish I had a Nikon 5000 scanner, but.

 

P.S. Skiing sucked this weekend,  The season in New Hampshire, USA Is over.  Time to think about 2022/23 season, if we are all still 

here?

 

Best,

 

Chuck

 

 

That is an interesting technique Chuck but it will only work in cases where there is a uniform cast across the image which is often the exception rather than the norm. Even then there is no real control without using WB sliders or similar.

 

In the case of your last image of the soldiers being transported,  there is a very strong orange cast affecting the soldiers inside the jeep and a blue cast on the area outside of the jeep. This is typical of pictures shot indoors where there is a yellow or orange cast from tungsten lighting and the outdoor lighting is bluish. There is even a difference in the cast on the soldiers low down in comparison to those sitting up high presumably because they are getting more daylight. In this case there is no one single, simple correction that can be applied and clearly the soldiers are far more important than the countryside. Keeping it simple and focusing on the two soldiers low down, the skin tones need a very strong correction for the orange cast (in LR WB slider terms, a strong negative Temp and negative Tint). If I was processing this image, I would also reduce the saturation a lot (or selectively reduce the yellow and red saturation).

[spacecadet]

Whilst that's technically true I have to say that image reads perfectly well to my eyes.

Add snow (maybe tanks) and, unfortunately, it could have been taken yesterday. 

Topical or what. Thanks, Chuck.

Edited March 21, 2022 by spacecadet

[MDM]

12 hours ago, spacecadet said:

Whilst that's technically true I have to say that image reads perfectly well to my eyes.

Add snow (maybe tanks) and, unfortunately, it could have been taken yesterday. 

Topical or what. Thanks, Chuck.

There is a completely objective way to check skin tones which I will post later when  I get a chance. I don’t know if you learned about the skin tone line using vectorscopes from your time studying cinematography as that was before the digital age. It is a remarkable fact that human skin tones fall in a very narrow area of hue regardless of lightness and saturation (so irrespective of race). Modern video editing apps usually have vectorscopes but unfortunately these have not yet made it to Lightroom or Photoshop. If you have an app with a vectorscope, check it out. 

 

However, it is possible to use the colour numbers in Lightroom or Photoshop with the eye dropper tool to check skin tones or even an app in the OS. As a general rule, red is always greater than green which is always greater than blue as for example R85G70B60. In dark-skinned people the numbers might be R55G40B30 but the differences between the channels are similar right across the lightness range. In the case of Chuck’s image, on the soldiers’ faces, the red is much, much greater than the green giving an orange cast. In his images with amagenta cast, the blue is often greater than the green in the faces. 

 

This is not subjective and dependent on monitor used. It is very evident to my eye in any case but the bottom line is that it is simple to check. In addition the saturation is too high in the skin tones but that does not affect the hue. 

Edited March 21, 2022 by MDM

[spacecadet]

No vectorscopes, no. Not in 16mm Grading on the Hazeltine was a job for the labs.

To my mind, in a single image, not a film where shot-to-shot colour balance is important, it's quite subjective.

It's possible to over-think these things. Wood for the trees and all that.

[MDM]

33 minutes ago, spacecadet said:

No vectorscopes, no. Not in 16mm Grading on the Hazeltine was a job for the labs.

To my mind, in a single image, not a film where shot-to-shot colour balance is important, it's quite subjective.

It's possible to over-think these things. Wood for the trees and all that.

Actually it is not subjective at all. If you measured the colours directly (not photos) of a series of faces in the same light using a spectrophotometer designed for the job you would find that they all have a very narrow range of hue. Skin tones are one of the so-called memory colours which means that most people (people with some form of colour blindness aside) will notice quite quickly that there is something off  if they are not within the objectively measured range. Whether that is of importance to an individual photographer or not is a matter of choice of course. For me, if I am shooting portraits then I do try to ensure the skin tones are correct - using a good quality grey card in the same light as the subject and a calibrated monitor greatly help in this regard. Each to his own but the bottom line is that, while  perception is subjective, there is an absolute objectivity involved here. 

Edited March 21, 2022 by MDM

[spacecadet]

Great image, Chuck.

Subjectively

[John Mitchell]

17 hours ago, Chuck Nacke said:

John,

 

Found this and it appears to work for me : 

 It appears to work quite well.  Last image that I posted, original was shot on 35mm EPP.  I have not tried it on a K-14 scan yet.

 

I wish I had a Nikon 5000 scanner, but.

 

P.S. Skiing sucked this weekend,  The season in New Hampshire, USA Is over.  Time to think about 2022/23 season, if we are all still 

here?

 

Best,

 

Chuck

 

 

I'll check that out. When I lived in Montreal, I used to go skiing in Vermont (usually Jay Peak). Never made it much farther south than that. Yes, hopefully we'll all be here next fall. It looks to be a bit of a toss-up at the moment. 😟

[MDM]

26 minutes ago, spacecadet said:

Great image, Chuck.

Subjectively

 

I agree. As I said way back I think Chuck’s war images are amazing and the scan quality is entirely secondary to the content. However, Chuck did thank me for my input in relation to the magenta casts in some of his other images which is why I commented on his most recent image. The fact is that there is a strong colour cast on this image clearly visible in the skin tones.  Whether that matters to anyone or not is another issue. I am happy to stop commenting at any point but this is an area of image-making where I do have significant expertise and I am happy to help anyone who wants my input. 

Edited March 21, 2022 by MDM

[wiskerke]

24 minutes ago, MDM said:

using a good quality grey card in the same light as the subject

 

Even better:

Always in my kit bag. Size of a credit card.

Here but somehow not in production anymore and being transferred to here, where it's not yet in production?

There's a bigger one (SG 140) with many skin tones added.

 

For grey/gray card I use these slightly bigger patches, the QPCard:

wim

 

[MDM]

8 minutes ago, wiskerke said:

 

Even better:

 

Always in my kit bag. Size of a credit card.

Here but somehow not in production anymore and being transferred to here, where it's not yet in production?

There's a bigger one (SG 140) with many skin tones added.

 

For grey/gray card I use these slightly bigger patches, the QPCard:

 

wim

 

Yes. Agree. I use a Color Checker Passport which also has a grey card but basically the same thing. It is easier to meter from a larger card than the little squares on the CCP. 

[Chuck Nacke]

1 hour ago, MDM said:

Yes. Agree. I use a Color Checker Passport which also has a grey card but basically the same thing. It is easier to meter from a larger card than the little squares on the CCP. 

Michael & Wim,

 

I also carry a color index card and use it on corporate shoots, will be using mine on Friday.  I also appreciate you pointing out my color scan problem.  It does seem to vary with the

film emulsion. 

 

Michael,

 

What do you think or have you had a look at the link I posted for correcting the color cast,  I have tried it and it appears to work well?

 

Chuck

[MDM]

28 minutes ago, Chuck Nacke said:

Michael & Wim,

 

I also carry a color index card and use it on corporate shoots, will be using mine on Friday.  I also appreciate you pointing out my color scan problem.  It does seem to vary with the

film emulsion. 

 

Michael,

 

What do you think or have you had a look at the link I posted for correcting the color cast,  I have tried it and it appears to work well?

 

Chuck

I did check out the link Chuck but have not had time to try it out yet. I will do but I think it will only work on images where the colour of the light is uniform over the whole image as it seems to work by taking an average over the entire image - similar to taking a picture in mixed lighting with say a subject lit by tungsten lighting near a window with daylight coming through. You have to meter the subject and use a colour checker in the same light as the subject to get a correct exposure of the subject. You don’t want an average between the tungsten and the daylight as neither will be correct. 
 

In the case of the Russian soldiers, they are in different light to the landscape outside (not sure if you used flash or just lighting within the vehicle) so I don’t think that method works there. As I said above, the faces of the two guys low down have a strong orange cast. Similarly in the case of the gunman image, his face is mostly in shade and has a strong magenta cast, whereas the scene behind is lit, so the colour of the light is different in the two areas and I don’t think that technique works here either. I think his face is more important than the colour of the snow so you could just correct the whole image for him which will make the snow go greenish or it could be done selectively by correcting for his face separately from the rest of the scene. 
 

Again I would emphasise that the contents of your images are far more important than the colour casts. 
 

 

Edited March 21, 2022 by MDM

[Chuck Nacke]

1 hour ago, MDM said:

I did check out the link Chuck but have not had time to try it out yet. I will do but I think it will only work on images where the colour of the light is uniform over the whole image as it seems to work by taking an average over the entire image - similar to taking a picture in mixed lighting with say a subject lit by tungsten lighting near a window with daylight coming through. You have to meter the subject and use a colour checker in the same light as the subject to get a correct exposure of the subject. You don’t want an average between the tungsten and the daylight as neither will be correct. 
 

In the case of the Russian soldiers, they are in different light to the landscape outside (not sure if you used flash or just lighting within the vehicle) so I don’t think that method works there. As I said above, the faces of the two guys low down have a strong orange cast. Similarly in the case of the gunman image, his face is mostly in shade and has a strong magenta cast, whereas the scene behind is lit, so the colour of the light is different in the two areas and I don’t think that technique works here either. I think his face is more important than the colour of the snow so you could just correct the whole image for him which will make the snow go greenish or it could be done selectively by correcting for his face separately from the rest of the scene. 
 

Again I would emphasise that the contents of your images are far more important than the colour casts. 
 

 

Michael,

 

I do appreciate your comments and that you took the time to talk about this.

 

Back when I was shooting film in the studio we would do emulsion tests and

order a specific emulsion of one film to do the job.  That was then and.

 

War of combat images are shot under whatever light or as we use to call it

"available darkness" their was.

 

I was operating under the assumption that the color cast was an artifact of the

old scanner and software I am using.  I do not have access to a target for the

CanoScan FS 4000 or targets for test emulsions of PKL-200, EPP, RD, RDP.

So I was looking for a way to offset what I believed to be a scanning artifact.

While I do work on a calibrated monitor, my color vision is not great, typical

male Red/Green issues.

 

John,

 

We have been at Loon Mountain every weekend since the end of November,

my daughter is on the Loon Race Team.  Never skied Eastern Canada.  I do miss

north Lake Tahoe in California.  I think ski season is over in New Hampshire for

this season.  It was in the 40's on Sunday.

 

Chuck