Fix the interpretation of film negatives scanned images

[Porpho]

When scanning B&W photo negatives and importing those images in Photo, the resulting image is too bright with contrast that is much too low. While this can be corrected by adding a Curves layer and applying a rather drastic curve, the same exact file loads in Adobe Camera Raw, Lightroom, and Photoshop correctly, which leads to believe that there is room for improvement in Affinity Photo. 

I attached examples below. You can see the original negative as it renders unretouched, the version in ACR with a simple inversion curve, and the same in AP, again, with a simple inversion curve.

The original film scan file is available if you need it. 

Thank you.

[carl123]

8 minutes ago, Paolo Portraits said:

The original film scan file is available if you need it. 

Can you upload that to the forum?

[Porpho]

Sure, the file is 350MB, will that be OK or should I provide a Dropbox link?

[carl123]

Dropbox will be fine

Thanks

[Porpho]

My pleasure. Here it is:

https://www.dropbox.com/t/gu6RvPceb6J7IisR

Thanks!

[walt.farrell]

There are significant differences between how the Apple Raw engine interprets that DNG file, and how the Serif Raw engine interprets it.

Apple:

 

Serif:

 

As a workaround for now, can you scan to TIFF instead? 

 

[Ldina]

54 minutes ago, walt.farrell said:

As a workaround for now, can you scan to TIFF instead? 

I saw the same thing as Walt, and agree that TIFF will be the way to go. I don't see much of an advantage to saving a scan as a DNG file (though I have never tried and may be missing something). As long as it is a decent scan and covers the dynamic range of your original, the results should be identical or very close.

I'm not sure why, but some DNG files (seems they are often from smart phones, unsupported cameras, or in your case, Silverfast scanning software) do not work well with the Serif RAW engine (or the underlying LIB-RAW module). Maybe these devices are encoding things differently, or perhaps not in conformance with the DNG standard? I was previously under the impression that DNG format was a fairly standardized, universal, open-source format (though not everyone supports it). I hope someone with greater knowledge about this subject will post some clarifications. 

[Porpho]

Yes, I can, @walt.farrell but I have a lot of previously scanned negatives that I might want to process in AP and it is indeed useful to have the option to scan in RAW. 

[Porpho]

1 hour ago, Ldina said:

I saw the same thing as Walt, and agree that TIFF will be the way to go. I don't see much of an advantage to saving a scan as a DNG file (though I have never tried and may be missing something). As long as it is a decent scan and covers the dynamic range of your original, the results should be identical or very close.

I'm not sure why, but some DNG files (seems they are often from smart phones, unsupported cameras, or in your case, Silverfast scanning software) do not work well with the Serif RAW engine (or the underlying LIB-RAW module). Maybe these devices are encoding things differently, or perhaps not in conformance with the DNG standard? I was previously under the impression that DNG format was a fairly standardized, universal, open-source format (though not everyone supports it). I hope someone with greater knowledge about this subject will post some clarifications. 

Since we shoot film, there can be situations when the exposure of the film is not optimal. When printing in the darkroom it is possible to compensate for those issues by, for example, exposing the paper for a longer time, to work around excessive density in the negative. When scanning, that is not always possible and many, many people prefer to scan to DNG for a multitude of reasons. When using that option, you can also ask the software, Vuescan or Silverfast, to take multiple samples of the are scanned. The Epson V750 and later scanners can do that and provide a richer sample of data that then can be tweaker in ACR or other software. The DNG file provides a lot of flexibility in adjusting the contrast and tones of the image.

Given that ACR can indeed interpret the DNG file correctly, whether that is generated by Vuescan or Silverfast, it would benefit Affinity Photo to be able to claim the same. AP is becoming the alternative of choice for many who are looking to move away from PS and it's best to have a solution that is, as much as possible, a drop-in replacement. Just MHO. I spent two days dealing with these issues from using DNG files in AP and I know that a lot of other photographers would easily give up thinking that the software is simply not working. 

Just pointing out the laws of the market, I'm not passing judgment. I love AP and I believe that in many ways it's a better software than PS but being able to import DNG file without having to tweak them for two days is a key issue.

Thanks everybody for the suggestions! 😀

[walt.farrell]

7 minutes ago, Porpho said:

When using that option, you can also ask the software, Vuescan or Silverfast, to take multiple samples of the are scanned. The Epson V750 and later scanners can do that and provide a richer sample of data

Did you do that for this one?

What does it actually do, to provide that "richer sample"?

[Ldina]

@Porpho Thanks for the explanation. I have an ancient, lower-end scanner, so I have never had the ability to scan to RAW format. I never felt I was missing anything, but not having tried it, I wouldn't know. 

 

[Porpho]

3 minutes ago, Ldina said:

@Porpho Thanks for the explanation. I have an ancient, lower-end scanner, so I have never had the ability to scan to RAW format. I never felt I was missing anything, but not having tried it, I wouldn't know. 

 

You're very welcome. If it's of any consolation, pretty much all flatbed film scanners are ancient 😀 We are in dire need of some really updated scanners with decently designed film holders.

 

[Porpho]

6 minutes ago, walt.farrell said:

Did you do that for this one?

What does it actually do, to provide that "richer sample"?

I didn't do it for this specific image, since the exposure was correct and the negative is developed right. When the software runs a single sample it makes a single read of the value and that can be OK or be slightly off. When running multiple samples the hardware/software combo gathers more data that can be then averaged, and it can possibly expand the tonal range. It can also reduce noise in dark areas.

Cheers.

[Ldina]

31 minutes ago, Porpho said:

We are in dire need of some really updated scanners with decently designed film holders.

I don't do much film scanning, and when I do, I use my DSLR mounted on a stand, a 100mm macro lens, an LED backlight, a diffusion filter (to eliminate pixelation from the backlight), and shoot RAW. APhoto does fine with my DNG files from my DLSR. Anyway, that's my 'film scanner!' If I did a tremendous amount of film scanning, it might be a different story. 

[Lee D]

@Porpho Thanks for the report.

I've created a report with the developers regarding the difference between the two RAW engines opening the DNG file.

[Porpho]

47 minutes ago, Lee D said:

@Porpho Thanks for the report.

I've created a report with the developers regarding the difference between the two RAW engines opening the DNG file.

My pleasure, thank you for looking into it.