Can we do some comparison negative scan conversion comparisons?

[Sangoma]

Here's one done very quickly on Affinity Photo 2




All of the examples are of the same scans processed on various software.

All of them would probably need some adjustment to contrast/brightness/saturation/white balance and maybe for any tint/cast

All of them on all the software could be tweaked to look exactly the same, and for that reason it would seem that the different examples are more of an indication of personal tastes than software capabilities.

I think the advantages of the various programmes would come in when they are being used for the whole process, ie scanning and any correction.
Thinking more of the companies that develop and scan, they would have presets for the different makes and types of film to retain the films characteristics, and to keep a constant result that could then be batch processed for printing or further tweaked.

I agree it is very interesting to see, but to me it hasn't told me anything about the various programmes, as user adjustment has or could of changed the finished look to make it look how any person's preference sees it.

 

[ChrisR]

Did you leave out IrfanView because it is free

No, an error, sorry Steve. Will add back to the list.

 

[Sangoma]

Also on IrfanView you can setup and save batch conversion profiles, which on the scans here can also include cropping out the black section on the left.

I tried it out of interest, and using the exact same settings for the three photos, they look quite reasonable

 

[excalibur2]

Here's one done very quickly on Affinity Photo 2

View attachment 412202


All of the examples are of the same scans processed on various software.

All of them would probably need some adjustment to contrast/brightness/saturation/white balance and maybe for any tint/cast

All of them on all the software could be tweaked to look exactly the same, and for that reason it would seem that the different examples are more of an indication of personal tastes than software capabilities.

I think the advantages of the various programmes would come in when they are being used for the whole process, ie scanning and any correction.
Thinking more of the companies that develop and scan, they would have presets for the different makes and types of film to retain the films characteristics, and to keep a constant result that could then be batch processed for printing or further tweaked.

I agree it is very interesting to see, but to me it hasn't told me anything about the various programmes, as user adjustment has or could of changed the finished look to make it look how any person's preference sees it.

Well I agree with you...the purpose of a scan is to get the maximum detail off the neg/pos and Dmax (max detail from shadows)... from then on it's just personal taste on how you want the finished JPG/tiff (or whatever) to look.......so many apps to do this starting with Photoshop.

 

[ChrisR]

My thanks to all of you. I thought I should put in some effort as well, so I have done some versions using Vuescan from the RAW DNG file, which I'll put up before long.

So for this as mentioned I used the Vuescan RAW DNG files which I scanned to help Nige's NLP effort. I opened each file in Vuescan Pro, set it as colour negative, and hit Preview to get a basic "scan". I went into the "Color" tab, and did some experimenting. Eventually I decided to set the white balance to Landscape. I clicked the little checkbox near the bottom of the tab for pixel colours, which shows you areas which will be clipped white in yellow (IIRC), areas which will be clipped black in blue, and areas out of gamut in cyan.

I then went through all of the film presets (and there are a lot of them), hoping I could find one which would remove both the white and black clipping... and failed to find any. In fact, the film presets seemed to make relatively little difference to the clipped areas, except a few made them worse. So then I went hunting for a film preset which would have no out of gamut areas. I found only one! That was Kodacolor VR200 Gen 5. (Mostly the out of gamut areas were also shadow areas, but one preset, possibly Ektar, was mostly out of gamut, at least for the 3rd image!)

Anyway, on that basis I decided to scan all 3 images as Kodacolor VR Gen 5. Subsequently resized but not otherwise edited.

(1)



(2)



(3)



It's maybe worth saying, I don't like these scans much... but they are in gamut!

By the way, I did the exercise again for one file, but starting from the RAW file saved as TIFF, and could see only the tiniest of difference on multiple comparative viewing.

 

[Sangoma]

I used Affinity and set so there were no clipped highlights, shadows or tones



Initially there were clipped shadows along the base of the fence on the left, and in the shadow at the base of the tree.

Out of interest, when you say "Anyway, on that basis I decided to scan all 3 images as Kodacolor VR Gen 5." and "It's maybe worth saying, I don't like these scans much" are you referring to new scans out of the scanner, or new developments of the original scans?

 

[ChrisR]

I used Affinity and set so there were no clipped highlights, shadows or tones

Out of interest, when you say "Anyway, on that basis I decided to scan all 3 images as Kodacolor VR Gen 5." and "It's maybe worth saying, I don't like these scans much" are you referring to new scans out of the scanner, or new developments of the original scans?

I did a new development from the second scan to RAW (done for Nige @FishyFish ).

Impressive you got no clipping with Affinity, as I said, I couldn't find any film preset that didn't have clipping.

 

[ChrisR]

OK, the Filmdev scans... therse have had an auto-Levels adjustment in C1Pro, and 50 units of clarity (original scans were not sharpened by Filmdev per request). A couple of these had to be re-sized down a bit as the file sizes (just over 500 kb) meant they wouldn't load.

(1) The track



(2) Boots



(3) Upton House garden (I originally had a few more adjustments on this one, which I have removed).



All expired Agfa Vista 200 in my Olympus mju II, devved and scanned by Filmdev. I remember the light being like that... but it's quite possible this is because I've looked at these images and various versions of them so many times now!

The scratch removal capabilities of the Noritsu scanner and software are really impressive. The scans in post #45 were at maximum infrared scan.

 

[trypdal]

All expired Agfa Vista 200 in my Olympus mju II, devved and scanned by Filmdev

Thanks Chris. I didn't realise this was expired film. Perhaps this can explain some of the doubts I was raising in post #29 on the colour dominants we're seeing in most of the inversions (especially in the boots sample).

I'd be interested - if you have time in the future - to participate/contribute to a new experiment looking at the range of results for a few samples from a fresh roll of fresh Gold or Colormax etc.

In my experience, colour fidelity using NLP/Colorperfect etc can be quite all over the place with expired film, but really predictable with fresh film developed in fresh chemistry.

 

[ChrisR]

Thanks Chris. I didn't realise this was expired film. Perhaps this can explain some of the doubts I was raising in post #29 on the colour dominants we're seeing in most of the inversions (especially in the boots sample).

I'd be interested - if you have time in the future - to participate/contribute to a new experiment looking at the range of results for a few samples from a fresh roll of fresh Gold or Colormax etc.

In my experience, colour fidelity using NLP/Colorperfect etc can be quite all over the place with expired film, but really predictable with fresh film developed in fresh chemistry.

I have a feeling that to do a better job on the comparison, it might be better for all the work to be done by one person, and for that person perhaps to be more of a power user of Vuescan than I am! The initial scan is clearly significant, not sure if I was the cause of the clipping, for example.

I remember some years ago having trouble with some scans of slides from NW Scotland. I eventually sent the slide film strip to 3 different labs, and the results that came back were all significantly different! As long as we can get our images something close to how we'd like them (not necessarily total fidelity, IMHO, just how we'd like them), then that's a win. For me, with C41, that'll be Filmdev for now...