Problems scanning color negatives (Fuji Reala and Epson Perfection V750)

[nhabedi]

Hi,

I bought an Epson Perfection V750 Pro scanner some months ago, mainly to scan my old black-and-white negatives. That went fine so far and I've been happy with it.

Now, after shooting black-and-white exclusively with my M4-P for more than 15 years, I finally started doing a bit of color photography after buying an M8.2 end of last year. Last month, I became even more adventurous and shot two rolls of Fuji Superia Reala 100. This might have been the first rolls of color film the M4-P ever saw, although I'm not totally sure as I bought it used in the beginning of the 90s...

However, when I then scanned these two rolls, I was more than disappointed by the results. On both rolls, I had a couple of frames left at the end and used those for a few "comparison shots" with my M8. Now, I used different lenses because of the M8's crop factor and I certainly didn't apply enough care to make this a real test or comparison that could be published in a magazine. But that wasn't what I was after anyway. I just wanted to see how the colors of the analog photos looked compared to the M8.

What I can say is that the M8 more or less got the colors "right". (I shot stuff I have at home and can compare.) The scans I can't seem to get right, though. If I use Epson's own scan software, the scans all have a somewhat blue-ish cast. If I use the highly-praised SilverFast software which came with the scanner using the correct "Negafix" profile for the Fuji film, the colors are even worse and wander off towards red or purple. I also downloaded a trial version of VueScan and used its Reala profile, but the results also weren't great.

For one film I also had scans done by a scan service. While their scans weren't totally perfect either (a bit too punchy and contrasty for my taste), the colors were much better in the sense of not having an obvious color cast. This seems to demonstrate that it is possible to get the colors more or less right and even in an automated way. However, I don't seem to be able to achieve the same results here at home.

I've created a temporary webpage with some examples which I think is more convenient than having to look at lots of photos uploaded to this forum. It's here:

Fuji Reala Test Scans

Any idea what could be wrong? Am I missing something? Any hints?

Thanks in advance for your help, Edi.


Some notes:

• The two films were developed in two different labs, so I think the problems are likely not due to a lab ruining the film during development.
  
• All three scan applications provide various means of fiddling with the scan to adjust the colors and if I play with this long enough I can eventually come to a point that is acceptable. However, I couldn't figure out a setting that works well for all photos on a roll and I don't want to believe that you have to adjust each individual photo for several minutes before you can scan it.
  
• In case that wasn't clear, I'm only talking about the color here. For my purposes, I'm generally happy with the resolution of the scanner although the samples on the website are obviously all very low-res.
  
• I'm using a calibrated Eizo monitor, so that shouldn't be the reason for wrong colors. Besides, even on my laptop the differences a quite obvious.
  
• SilverFast provides an option to calibrate the scanner using an "IT8" target, but this is only used for slides and the profile is turned off when scanning negatives. With VueScan one can seemingly calibrate the scanner for a specific type of negative film, but I haven't done that yet.
  
• I understand that it is probably easier to get accurate results with slides, but I specifically want to shoot color negatives.

[andybarton]

Scanning colour negatives has always been a hit and miss affair in my experience. One of the big variables is the processing, I reckon. Finding a reliable, reasonably priced C41 processor is part of the solution, but how you do that these days, is another question.

I use Vuescan and a Nikon V. Sometimes they come out fine, other times, they are all over the place.

This is one of the reasons why I very rarely use colour negative film these days - life is too short. I generally use Astia for color film work now, and I am very happy with my results.

[joakimsundkvist]

For Vuescan and color negs I recommend the calibration and work flow explained in:
Better colour neg scanning with VueScan. | Urban Motion

/Joakim

[andybarton]

That's very helpful Joakim. I look forwards to trying this out later.

I have never understood why negative colour film needs a pink base. I am sure someone here will be able to tell me

[elmarman]

The 'pink' base on colour negative films is because the dyes are not perfect in their light absorption-- for example, the magenta layer should absorb all the green light but it does not,having an 'over-lap' in other parts of the spectrum, so the mask is added, same as in colour papers, to get better colour absoption. When I started colour negative work in about 1968 I used Agfa CN17 and Gevaert'Scientia Color' which were un-masked films. Masked films have 'purer and vibrant colours'

[masjah]

Originally Posted by elmarman

The 'pink' base on colour negative films is because the dyes are not perfect in their light absorption-- for example, the magenta layer should absorb all the green light but it does not,having an 'over-lap' in other parts of the spectrum, so the mask is added, same as in colour papers, to get better colour absoption. When I started colour negative work in about 1968 I used Agfa CN17 and Gevaert'Scientia Color' which were un-masked films. Masked films have 'purer and vibrant colours'

Agreed. A good explanation is found in Osterloh's old book "Applied Leica Technique":

"A problem with colour negative films is the fact that each of the three light-sensitive layers does not cover its respective segment of the colour spectrum accurately. In other words, although the green-sensitive layer has its greatest sensitivity in its designated spectral range, it is nevertheless also slightly sensitive to the two other ranges, blue and red.

The slight overlap in colour sensitivities is also present in the blue- and red-sensitive layers. In actual use, this would result in unwanted side-effects, such as colour shifts, lower saturation and reduced brightness. In order to minimise the sensitivity of each layer to the other two ranges, i.e. in order to avoid undesired densities, so-called coloured masks are coated between the colour-sensitive layers. These masks act like filters by absorbing the undesired colour components. Their presence is easy for anyone to see because of the overall orange colouring of the processed film."

[nhabedi]

Originally Posted by joakimsundkvist

For Vuescan and color negs I recommend the calibration and work flow explained in:
Better colour neg scanning with VueScan. | Urban Motion

Joakim, thanks a lot, that was a very helpful hint! It obviously helps to RTFM, but this gem is really well-hidden. Seems to be a feature (measuring the film base yourself) that only VueScan offers.

I've now played a bit more with VueScan and also calibrated the scanner (which makes a difference for VueScan although it is ignored by SilverFast) and the results look a lot better now. The colors still aren't exactly like on the M8, but they're much more realistic and the M8 isn't 100% accurate either. (No camera is.) What remains is probably at least partly due to the specific characteristics of the Reala film and one of the reasons for shooting film in the first place.

I unfortunately have two issues with VueScan (it crashes pretty often the first time I start it and in some cases it didn't scan exactly the area I had selected via manual cropping) which I'll try to sort out with their support. Otherwise, it seems to be an application I can recommend. SilverFast always gets all the praise, but my experience so far is that with VueScan you can get better results for color negatives.

Thanks again,
Edi.

[suilvenman]

What excellent ideas, advice and links. It's threads like these that prevent many hours of frustration and anguish for a newcomer to digital post-processing. I'm really appreciative of this.

Cheers, Ken.

[tobey bilek]

Scan the film using profesional mode. Adjust the exposure and color balance and save it at the top where it says current settings.

You let the M8 auto color balance and daylight film is fixed color balance. Therefore the test is not valid.

Go outside or use flash. With proper scanning, they will match. Abusing color film under non standard conditions is asking for trouble which you found.

[nhabedi]

Originally Posted by tobey bilek

You let the M8 auto color balance and daylight film is fixed color balance. Therefore the test is not valid.

Go outside or use flash. With proper scanning, they will match. Abusing color film under non standard conditions is asking for trouble which you found.

All examples I've shown were shot with daylight being the only light source. I don't think that I'm "abusing color film under non standard conditions" if I shoot daylight film at daylight. And no amount of fuzzing with the white balance makes for example the M8 photo of brown wooden tiles look like what Silver Fast made of it.

[rulnacco]

I've got the V750 and have generally had good luck scanning various C-41 films with Silverfast, although it can take a bit of tweaking. Then, I generally turn to VueScan as well, but I really find it a lot clunkier to deal with.

If you do have VueScan, though, you might want to check out CF Systems' ColorNeg plugin:

C F Systems Innovations in Sight and Sound - Photoshop plug-ins page

It's not perfect, butI've found it works extremely well with several film types. You can download and demo it for free. You might see if it works well for your preferred films.

[cmatter]

Originally Posted by nhabedi

Hi,

I bought an Epson Perfection V750 Pro scanner some months ago, mainly to scan my old black-and-white negatives. That went fine so far and I've been happy with it.

Now, after shooting black-and-white exclusively with my M4-P for more than 15 years, I finally started doing a bit of color photography after buying an M8.2 end of last year. Last month, I became even more adventurous and shot two rolls of Fuji Superia Reala 100. This might have been the first rolls of color film the M4-P ever saw, although I'm not totally sure as I bought it used in the beginning of the 90s...

However, when I then scanned these two rolls, I was more than disappointed by the results. On both rolls, I had a couple of frames left at the end and used those for a few "comparison shots" with my M8. Now, I used different lenses because of the M8's crop factor and I certainly didn't apply enough care to make this a real test or comparison that could be published in a magazine. But that wasn't what I was after anyway. I just wanted to see how the colors of the analog photos looked compared to the M8.

What I can say is that the M8 more or less got the colors "right". (I shot stuff I have at home and can compare.) The scans I can't seem to get right, though. If I use Epson's own scan software, the scans all have a somewhat blue-ish cast. If I use the highly-praised SilverFast software which came with the scanner using the correct "Negafix" profile for the Fuji film, the colors are even worse and wander off towards red or purple. I also downloaded a trial version of VueScan and used its Reala profile, but the results also weren't great.

For one film I also had scans done by a scan service. While their scans weren't totally perfect either (a bit too punchy and contrasty for my taste), the colors were much better in the sense of not having an obvious color cast. This seems to demonstrate that it is possible to get the colors more or less right and even in an automated way. However, I don't seem to be able to achieve the same results here at home.

I've created a temporary webpage with some examples which I think is more convenient than having to look at lots of photos uploaded to this forum. It's here:

Fuji Reala Test Scans

Any idea what could be wrong? Am I missing something? Any hints?

Thanks in advance for your help, Edi.


Some notes:

• The two films were developed in two different labs, so I think the problems are likely not due to a lab ruining the film during development.
• All three scan applications provide various means of fiddling with the scan to adjust the colors and if I play with this long enough I can eventually come to a point that is acceptable. However, I couldn't figure out a setting that works well for all photos on a roll and I don't want to believe that you have to adjust each individual photo for several minutes before you can scan it.
• In case that wasn't clear, I'm only talking about the color here. For my purposes, I'm generally happy with the resolution of the scanner although the samples on the website are obviously all very low-res.
• I'm using a calibrated Eizo monitor, so that shouldn't be the reason for wrong colors. Besides, even on my laptop the differences a quite obvious.
• SilverFast provides an option to calibrate the scanner using an "IT8" target, but this is only used for slides and the profile is turned off when scanning negatives. With VueScan one can seemingly calibrate the scanner for a specific type of negative film, but I haven't done that yet.
• I understand that it is probably easier to get accurate results with slides, but I specifically want to shoot color negatives.

In my experience the key to scanning color negs (and BW, to an extent) is the scanner software you use. The only two I've ever found to be worth anything for scanning color negs have been Ed Hamrick's Vue Scan (be sure to pay for the Pro version.. +- $100).. for virtually any scanner, and Hasselblad (formerly Imacon)'s Flexcolor (free with an Imacon scanner or Hasselblad digital camera). Nothing else has worked nearly as well.