Canon slide copying set-up

[sooth]

.

Edited October 19, 2023 by sooth

[Harry Harrison]

19 minutes ago, sooth said:

DIY copy stand (vertical, more expensive setup)

 

https://forums.negativelabpro.com/t/lets-see-your-dslr-film-scanning-setup/27/36

 

Good thread, I was racking my brain to work out what the 'NLP' was that they seemed to like so much, it wasn't going to be Neuro Linguistic Programming. 

 

I should have looked at the URL - Negative Lab Pro - for inverting colour negatives, do you know anything about it?

 

 

[Chuck Nacke]

After pages and pages of text, I feel the need to write that I have not seen

a single 35mm slide copied with a DSLR that equals the quality that can

be done by using a quality desktop film scanner (4,000 dpi or above).  

Mark has shown the best results that I have seen.  Keep in mind that

my goal is to only scan a chrome, slide, once and have a 100+ mb file

in 16bit color depth to work from today or ten years from now.

 

On another note, I do not believe that scanning at above 5,000 dpi gains much?

 

Scanning 35mm chromes is a lot of work, I spend over four hours scanning and

spotting each chrome.  Captioning and keywording also takes a lot of work.

 

Chuck

[M.Chapman]

Playing with Topaz again using the Denoise AI option (the default).

 

I found I wasn't getting the quality of results I was seeing this morning. After some experimenting I find that if I use Denoise AI on my image after it's downsized to 3000 x 2000 image, it's pretty good (left hand image). But if I apply the same denoise to the 2795 x 4193 original image the results are significantly worse (right hand image). A sort of mottling effect (the original film grain?) extending inwards from the edge of the frame and in other places. So I tried a 2000 x 3000 crop from the original and that also shows the problem, so it's not to do with the size of the image, but more to do with the the scale of the "grain" it's working on. If the grain spacing / amplitude goes above a certain limit it's retained as a "feature". No surprises there I suppose, but it makes the results harder to predict. It's also very slow to process the image, making it difficult to use. It's a shame because under the right conditions it can work really well.

 

 

So I tried the AI clear option and that struggles too.

I wonder how well DxO Prime Denoise does?

 

Artificial Intelligence  

 

Mark

Edited February 9, 2020 by M.Chapman

[MDM]

29 minutes ago, M.Chapman said:

 

Artificial Intelligence  

 

Mark

Oxymoron  It’s the big buzzword. I gave up on Topaz for the moment and went for a Nepalese with my wife. 

 

 

2 hours ago, Harry Harrison said:

 

Good thread, I was racking my brain to work out what the 'NLP' was that they seemed to like so much, it wasn't going to be Neuro Linguistic Programming. 

 

I should have looked at the URL - Negative Lab Pro - for inverting colour negatives, do you know anything about it?

 

 

 

No but there are some serious setups by some serious guys on there. Looks interesting. I might join in when I start on my negs. The more I see the happier I am with my setup. 

 

 

Edited February 9, 2020 by MDM

[KevinS]

4 hours ago, M.Chapman said:

I wonder how well DxO Prime Denoise does?

I'm not well experienced with Noise Reduction, if fact I didn't use it at all until I started using DxO software. The regular NR on PhotoLab 3 is fine for most images, and is the default setting. For a small percentage of images I do apply Prime, by toggling it on and off and comparing. I believe it is automatically applied to image areas as needed and I certainly don't select areas of the image. Pretty sure it works on RAW files only. Some info here;

https://www.dxo.com/dxo-photolab/prime-denoising-technology/

Edited February 10, 2020 by KevinS

[spacecadet]

17 hours ago, Harry Harrison said:

 

 

 

I should have looked at the URL - Negative Lab Pro - for inverting colour negatives, do you know anything about it?

 

 

There are supposed to be all sorts of complications with curves and whatnot, but I usually get good results like this.

To invert, drag the bottom left of the RGB tone curve to the top left, then top right to bottom left. (I have this saved as an import preset, together with some contrast adjustment- a simple S-curve).

Eyedropper a midtone for colour balance.

That's it.

Under-exposed negs may need some work on the curves as well.

 

Edited February 10, 2020 by spacecadet

[Harry Harrison]

12 hours ago, Chuck Nacke said:

I feel the need to write that I have not seen

a single 35mm slide copied with a DSLR that equals the quality that can

be done by using a quality desktop film scanner (4,000 dpi or above).  

Mark has shown the best results that I have seen.

That's fair enough Chuck, I respect your opinion. However without perhaps singling out any particular images perhaps you could say what in your view lets down the examples of DSLR scanning that you've seen on this forum. Is it perhaps highlight or shadow detail, contrast, colour balance, grain or just plain detail?

[MDM]

21 minutes ago, Harry Harrison said:

That's fair enough Chuck, I respect your opinion. However without perhaps singling out any particular images perhaps you could say what in your view lets down the examples of DSLR scanning that you've seen on this forum. Is it perhaps highlight or shadow detail, contrast, colour balance, grain or just plain detail?

 

I think Chuck's assessment is not based on evidence, just personal preference (or qualitative bias if you prefer). Looking at one image as displayed on the Alamy website and comparing it to other images displayed elsewhere is not an objective assessment of the various processes. 

 

The only way anyone can make realistic and meaningful judgements is to try the different methods on the same slides oneself and even then there is a very high degree of subjectivity - what does "equals the quality" actually mean - sharpness, absence of noise, colour fidelity and so on. Most importantly the images need to be viewed at full size in a colour managed program downloaded to a computer.

 

This is not worth arguing about really as there is no way of proving one is right. All I can say with certainty is that I get significantly superior results using the criteria that I consider valid with my D810 setup than I ever got from my reasonably high quality LS4000 scanner - superior in terms of dynamic range, sharpness, highlight and shadow detail, noise control, white balance and so on. This is partly because I am shooting raw and also because the camera and lens I am using gives superior results to the LS4000 in any case (using the same criteria).

Edited February 10, 2020 by MDM

[M.Chapman]

 

20 hours ago, Harry Harrison said:

That's fair enough Chuck, I respect your opinion. However without perhaps singling out any particular images perhaps you could say what in your view lets down the examples of DSLR scanning that you've seen on this forum. Is it perhaps highlight or shadow detail, contrast, colour balance, grain or just plain detail?

 

And.. to be fair, I think the images that Chuck has complimented me on were DSLR digitised 120 medium format transparencies that I posted much earlier in this thread.

 

 

The quality I'm getting from my 35mm slides is nowhere near as good.

 

Mark

Edited February 11, 2020 by M.Chapman

[geogphotos]

I would suggest that 99%+ of an image's quality is in the content of the image. For me the point of digitising these old slides is only really to do it when the content makes it worthwhile.

 

There is a complete oversupply of present day images of just about every subject under the sun. But until somebody invents a time machine there will always be value in historic images which can't be reproduced ( excluding landscapes, seascapes, trees, plants that haven't changed in appearance etc). So these old images have value even if shot on substandard lens/camera/film etc. Even if they are just snapshots showing people doing things, wearing old fashions, standing by old cars etc. They may have low technical quality but high content quality. This is what Alamy's archive route is there for.

 

Back to the practicalities. Using a camera is clearly a lot faster than a film scanner, but a film scanner has dust removal and can be left to whir away while you do other things for a few minutes and then load in the next ones. Slower but once set up it can just run with fairly minimal input. Is there any equivalent dust removal system for camera shot digital images? I am not necessarily thinking about the removal of absolutely everything but the main, most obvious dust spots and blemishes?

 

Do you save time on the copying but then lose it in the post-processing if digitising via camera?

Edited February 10, 2020 by geogphotos

[Harry Harrison]

1 hour ago, MDM said:

Looking at one image as displayed on the Alamy website and comparing it to other images displayed elsewhere is not an objective assessment of the various processes. 

Quite so, and having tried all the different methods discussed (high end scanner, standard 4000 dpi film scanner, DSLR Scanning) I'm happy with the conclusions that I've drawn from my own results but I'm still interested to know why Chuck fails to see what he's looking for in the sometimes high resolution examples that have been offered for discussion on here.

Edited February 10, 2020 by Harry Harrison

[MDM]

1 minute ago, Harry Harrison said:

Quite so, and having tried all the different methods discussed (high end scanner, standard 4000 dpi film scanner, DSLR Scanning) I'm happy with the conclusions that I've drawn from my own results but I'm still interested to know why Chuck fails to see what he's looking for in the sometimes high resolution examples that have been offered for discussion on here.

 

Could you summarise your conclusions please Harry, especially in relation to using a high end scanner versus camera - the thread has gone on so long that it is difficult to find anything now.

[Harry Harrison]

2 hours ago, spacecadet said:

There are supposed to be all sorts of complications with curves and whatnot, but I usually get good results like this.

Yes, fair enough, but they do still look like scans as you yourself agree, the helicopter and the car particularly. At some point I'll try the free trial of Negative Lab Pro to see how good it is in comparison because I never got on with Colorperfect. My aim would be to, without too much effort, make scans form colour neg not look as if they are scans from colour neg.

[MDM]

33 minutes ago, geogphotos said:

 

 

Back to the practicalities. Using a camera is clearly a lot faster than a film scanner, but a film scanner has dust removal and can be left to whir away while you do other things for a few minutes and then load in the next ones. Slower but once set up it can just run with fairly minimal input. Is there any equivalent dust removal system for camera shot digital images? I am not necessarily thinking about the removal of absolutely everything but the main, most obvious dust spots and blemishes?

 

Do you save time on the copying but then lose it in the post-processing if digitising via camera?

 

 

How long is a piece of string? It depends on how dirty your slides are, how experienced you are with post-processing and what steps you take. I estimate about half an hour per slide to bring it up to Alamy QC quality. It also depends on camera setup - the ES-1 is very fast as alignment and focus are very easy. If you use a tripod and manually align, then it is likely to be very slow and frustrating. It also depends on what scanner you are using - many don't have hardware dust removal and the efficiency will no doubt vary between scanners. I can't imagine how you could have a hardware dust removal setup with a camera but maybe somebody else can - the next big thing perhaps.

[Harry Harrison]

4 minutes ago, MDM said:

Could you summarise your conclusions please Harry, especially in relation to using a high end scanner versus camera - the thread has gone on so long that it is difficult to find anything now.

Cripes. I basically agree with you, I'm very happy with DSLR scanning, and not just because it's quick, I may even get an ES-1 one day but really the Illumitran (great name isn't it, like something from a sixties sci-fi movie) means I don't need to. 

 

However, as I've said somewhere before, I'm marking my own homework so I will scan a sharp detailed slide with lots of shadow & highlight detail by all three different methods that I'm able to use, upload them and see what others think. Just not this week!

[NYCat]

I came across this on another forum... "For anyone wanting to know how to convert color negatives, yes, you can do it in Lightroom or Photoshop but just because you can, doesn't mean...there's not a better way. Get Negative Lab Pro (Lightroom plugin) and life will be better. They have a very active group on Facebook. "

 

I never thought I would have anything to contribute to this thread!

 

Paulette

[spacecadet]

31 minutes ago, Harry Harrison said:

Yes, fair enough, but they do still look like scans as you yourself agree, the helicopter and the car particularly. At some point I'll try the free trial of Negative Lab Pro to see how good it is in comparison because I never got on with Colorperfect. My aim would be to, without too much effort, make scans form colour neg not look as if they are scans from colour neg.

Eh?

I don't agree and I don't know what you mean. They look like photographs to me. 

I'm sure they could be improved but I spent as much time as I thought sensible- I don't see any merit in trying to make scans from film look like digital images.

I don't know about the software you've mentioned because I haven't used any.

Edited February 10, 2020 by spacecadet

[MDM]

7 minutes ago, NYCat said:

I came across this on another forum... "For anyone wanting to know how to convert color negatives, yes, you can do it in Lightroom or Photoshop but just because you can, doesn't mean...there's not a better way. Get Negative Lab Pro (Lightroom plugin) and life will be better. They have a very active group on Facebook. "

 

I never thought I would have anything to contribute to this thread!

 

Paulette

 

Well thanks Paulette and your input is much appreciated and welcome 😀. Unfortunately you are about 16 hours too late. I had never heard of it until last night but there is a post at the top of this page which has a link to it. I had a look at it last night and it seems like a very interesting app. 

[Harry Harrison]

9 minutes ago, NYCat said:

I never thought I would have anything to contribute to this thread!

Thanks Paulette, this absolutely isn't a 'Guys Only' thread I hope. I have a rough & ready way in Lightroom, you have to manually set up an 'Invert' curve first, which is useful anyway, and then the eye-dropper on the film rebate gets close but you still need a huge amount of contrast added which is difficult to do without accentuating the noise and grain.

[Harry Harrison]

3 minutes ago, spacecadet said:

I don't see any merit in trying to make scans from film look like digital images

It's just that the conventional wisdom on the forum seems to be that they do need to if they are going to pass regular QC rather than go up via the archive route as I think your images did. When I shot colour negative I wasn't intending that they would look 'historical' or that they might eventually be described as possibly having imperfections. I often printed them myself and they portrayed the scene as I saw it. Still, I may not even upload any but I will be scanning them for myself so if 'NLP' does a good job I may well consider that the $99 + vat cover price is worth it.

[spacecadet]

5 minutes ago, Harry Harrison said:

It's just that the conventional wisdom on the forum seems to be that they do need to if they are going to pass regular QC rather than go up via the archive route as I think your images did. When I shot colour negative I wasn't intending that they would look 'historical' or that they might eventually be described as possibly having imperfections. I often printed them myself and they portrayed the scene as I saw it. Still, I may not even upload any but I will be scanning them for myself so if 'NLP' does a good job I may well consider that the $99 + vat cover price is worth it.

We'll have to agree to disagree, then, because I don't have any problem with my film scans looking as they do and neither do the buyers. It's moot whether they would pass QC as the job's done. Of course I don't personally have to worry about QC anymore but that's by the by.

Incidentally I don't find that increasing the contrast on the neg scans increases the grain as such. I don't have to do much of it as I have an import curve. and I don't have any noise as I scan at 100ISO. Indeed some NR can suppress the grain a bit, but I'm not sure I've ever done it for an upload.

NLP is just a plugin for LR, so it doesn't do anything you couldn't work out yourself. Emulating a particular film isn't something I'm trying to do.

[MDM]

23 minutes ago, Harry Harrison said:

It's just that the conventional wisdom on the forum seems to be that they do need to if they are going to pass regular QC rather than go up via the archive route as I think your images did. When I shot colour negative I wasn't intending that they would look 'historical' or that they might eventually be described as possibly having imperfections. I often printed them myself and they portrayed the scene as I saw it. 

 

Same here exactly. I bought my first Nikon LS2 slide scanner in 1997 and started printing on an Epson 400 at the same time. The prints are awful when I look at them now and rubbish  compared to what I could do in a wet darkroom at that time but it was cutting edge prosumer stuff back then. If I am digitising old film images now I don't them to look like re-photographed enprints with no colour correction or other post-processing. I just want to get them looking as good as I can, as if I was printing them myself in a darkroom when I took them. 

Edited February 10, 2020 by MDM

[spacecadet]

7 minutes ago, MDM said:

If I am digitising images now I don't (want) them to look like re-photographed enprints with no colour correction or other post-processing.

Neither do I, and IMHO they don't.

I've removed my examples. I thought I was being helpful. I didn't put them up to be judged in this way.

Edited February 10, 2020 by spacecadet

[Chuck Nacke]

2 hours ago, MDM said:

 

I think Chuck's assessment is not based on evidence, just personal preference (or qualitative bias if you prefer). Looking at one image as displayed on the Alamy website and comparing it to other images displayed elsewhere is not an objective assessment of the various processes. 

 

The only way anyone can make realistic and meaningful judgements is to try the different methods on the same slides oneself and even then there is a very high degree of subjectivity - what does "equals the quality" actually mean - sharpness, absence of noise, colour fidelity and so on. Most importantly the images need to be viewed at full size in a colour managed program downloaded to a computer.

 

This is not worth arguing about really as there is no way of proving one is right. All I can say with certainty is that I get significantly superior results using the criteria that I consider valid with my D810 setup than I ever got from my reasonably high quality LS4000 scanner - superior in terms of dynamic range, sharpness, highlight and shadow detail, noise control, white balance and so on. This is partly because I am shooting raw and also because the camera and lens I am using gives superior results to the LS4000 in any case (using the same criteria).

Michael,

 

The only example of an "almost" in my opinion, acceptable copy of a 35mm chrome made by photographing a chrome was posted by Mark: 35mmVelvia50.jpeg and it was only a 3000 by.  No I am not basing anything on evidence, I have other work to do...  I make my own judgment first on the image, then on the context and lastly on how it appears on the vehicle that I am seeing it on.  That is life in the digital world.

 

In response to Harry,  If you or anyone reading this forum wants to see examples of my scans from 35mm chromes you are welcome to look at my most recent ten images, don't know the exact number, images I have online.  There are a few that are not perfect, but at least I do have the original 16bit TIFF's at over 5500 by.

 

As I have written before on this thread, "I am not looking for the quickest way to digitize a film frame, I am looking for at permanent copy of a chrome at as close to the original as I can get."

or as a dear friend of mine says," Speed, Quality, Cost, choose any two of the three."

 

A problem, I am guilty as well, on this forum is that people do not read all of what is written or they do not understand what is being said.  I for one am trying to make sure that I have read what is written and understand what has been written.

 

Chuck

Edited February 10, 2020 by Chuck Nacke
addition