Scanning Negatives—What Am I Doing Wrong?

[erl]

Advertisement (gone after registration)

If you believe that a RAW file from a camera contains the maximum captured data, which can be later processed to personal desires, then you will accept the following, I hope.

 

Scanning with the widest range available to your particular scanner (ie. black & white points wide apart) will capture the most data, but inherently produces a 'flat' looking scan. Now you will not want to use all that data, but select the portion you want according to taste. By archiving your original scan from which you now create either a tiff or jpeg, you can process your tones to how you want them. Yes, this will throw away data, but that is better than throwing it away at the scanning stage, never to be retrieved. This way, the 'lost' data is not really lost because it is still in the archived scan, available to be processed differently when you 'change your mind' on what you like.

 

Adjustment to the scanning stage are similar to jpegs in cameras. They are processed (data loss) before you ever get access to them.

 

All this pre-supposes that you are interested in maximum quality. For speed, yes process at the scanning stage and lose the dumped data and for some purposes that is fine. It really comes back to 'user satisfaction'.

 

I hope that is enlightening to some degree.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Julian Thompson]

I understand what you mean about my example not being perfect - let me try to qualify it slightly.

 

First off, working with a positive jpeg like this does not bring out the best in colorperfect, so if you did have the negative it is likely that a lot of the reddish tinge would be removed by use of the 'lock film base colour' control in vuescan and then by colorperfect.

 

Secondly, I have not attempted to do any actual post processing of levels or anything else - some of the examples shown (whilst they are aesthetically MUCH nicer than the colorperfect one) actually lose detail in the shadows through post and have a lot of 'definition' added to them.

 

What I was trying to show was just a significant improvement of a badly scanned positive using a one touch click in CP - you can then process it as you wish for whatever it is you're trying to do with it in PS or aperture or whatever.

 

I'll have a go at one from there for fun :-)

[Julian Thompson]

Actually - you know what - I've just had a go at pp 'ing this to get it where I would want to be and I've realised what's happening.

 

A little while back I was using Vuescan without using color perfect after on the negatives - rather like you are here - and I was having this problem where by no matter what I did the shots always came out with very odd red channel issues.

 

And that's where you are now - the scan has a weird red channel - even when it's been run though colorperfect.

 

If you mess with the red channel it can be fixed but as has been pointed out - it's a bad scan.

 

Can you scan it as a negative for me with vuescan with color balance set to NONE in the menu?

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Julian Thompson]

PS That one above here was with one click just in case I was confusing - look at the adjustment!!!

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Julian Thompson]

(I meant NOT with one click of course - sorry!)

[skunkworks]

Scanning with the widest range available to your particular scanner (ie. black & white points wide apart) will capture the most data, but inherently produces a 'flat' looking scan. Now you will not want to use all that data, but select the portion you want according to taste. By archiving your original scan from which you now create either a tiff or jpeg, you can process your tones to how you want them.

 

OK, I understand your position and it makes sense if what you are doing is making an archival scan. If you want the most data, though, you might as well scan as an HDR file. Following the same logic, you would then want to scan as large a file as possible since you can always remove data for a smaller print; even though the file is cumbersome you'll never have to scan again. I can see why this has it's advantages, but there are tradeoffs to consider.

 

Personally, I consider the negative to be the archive: although subject to chemical aging, if well stored it will last longer than my computer's hard drive and software (and perhaps the file type and any ability to read it). And as I said previously, if you need a specific "look" you can just scan again--perhaps with a far superior future scanner (which has happened to me in a big way). Also, since the film itself already is biased, there's little practical reason to consider a neutral scan as archival. If there is any weakness in the negative, whether with exposure, developing, age, even focus, you are already probably going to manipulate the scan to compensate as it would be a total waste to go for a perfectly neutral scan of an underexposed image.

 

When I scan, I try for the highest quality digital file; it never looks flat and requires only minimal post-processing. The result is a file I can easily store and handle. Size matters to me: I deal with tens of thousands of files and I'm not rich enough to afford massive storage space. I rarely print larger than 11 x 14" so my scans are optimized for that size.

 

However, if what you are after is to scan once and dispose of the negative, then I concede your points. It just isn't practical for me.

[erl]

Advertisement (gone after registration)

Skunkworks, I accept that you have devolved a workflow to suit your requirements. That is paramount, and I would not suggest changing if you are satisfied.

 

However my intended point, maybe I didn't express it well, is to obtain the most data from the film during scanning. This is regardless of how large or small your intended output is. I don't believe any scanner will extract ALL the data on the film, but the intent is to digitize as much as possible for your initial file (I call this the archive file) from which you can make any number of subsequent manipulated files. To achieve this end you must not clip any data. Hence, widening the spread between the B&W pts.

 

I never recommend disposing of the film, that is close to suicide as far as I'm concerned. Currently I have in excess of 30,000 films on file, and growing. It is painless to archive them, and as I think you intimated, more 'secure' than digital archiving. However, accessing digi files is usually more efficient than film files, but both are valuable and worth keeping.

[janki]

I use Nikon Scan 4 software on my seven years old, but well calibrated iMac.

 

I always manage to print very good color correct pictures of the scanned TIFF files without any post work.

 

I produce far better pictures from the film on my Nikon scanner then the lab.

 

@bumac mentioned the neutral grey pipette. It is the key to correct colors in the picture. I use the neutral gray pipette in the Nikon Scan 4 software on e.g. sand, concrete, stone, mud, care wheels, and zinc coated iron, i.e. any brown/grey surface.

 

I do the adjustment several times to get an idea of what direction I have to move/adjust the picture’s color curves.

 

Of course you must have knowledge of the color circle.

 

Jan Erik

[piblondin]

Unfortunately, I'm still having problems with the red. Here's a new scan and my settings are attached below. http://dl.dropbox.com/u/856708/Scan-100719-0006.jpg

 

Any help is much appreciated! Thanks again.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[erl]

Wow! That red is bad.

 

My first reaction to it is that the film is badly processed. If you have another image, from a completely different film and even preferably from a different lab, try scanning that. It may simply be the problem.

 

In the days before I processed my own films, I would never submit my films for processing before lunch on Mondays or after lunch on Fridays because some labs notoriously had their processes out of balance at those times. Particularly Friday afternoons they tended not to replenish the chemistry knowing it was going to sit over the W/E and 'go off.'

[piblondin]

I installed Parallels and Nikon Scan for Windows on my Mac. The resulting scans there are much better. So, I think it's an issue with VueScan. Any ideas? I would love to be able to scan using a native Mac application!

[erl]

Another suggestion, if you want to keep using VueScan, is to do a pre-scan, then change your selection of film settings and observe the changes. Do that with other variables in the program to see if you can pin down what is delivering the red cast.

 

I am reasonably certain it must be either the film processing or the film profile you have selected in the menu. I would also recommend trying an image from a different film to eliminate that as the cause.

[piblondin]

I upgraded my version of VueScan and that helped considerably. See attached. Now I just need to figure out the post to get something close to the lab print.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[erl]

OK. That is greatly improved. It now looks you only need to 'squeeze' the black point and white point together to get your proper contrast happening. May also need mid tones slider adjusting. Only know that after the former two. Colour may need tweaking. Depends on how the original lighting was. I suspect it should be warmer, but only you were there.

[piblondin]

Thanks, Erl! I guess there was something wrong with VueScan. It seems a lot of their incremental updates have to do with handling color across different scanners. After all that, I am contemplating a new scanner. If anyone happens to read this and has opinions on my current Nikon CoolScan IV ED vs. an Epson V600 or V700 for scanning 35mm negatives and slides, let me know.

[erl]

I have the Coolscan V ED and it is fine. The only difference between yours and mine that I am aware of is that the V is faster.

 

VueScan does upgrade very frequently, but I have used all upgrades for years and the real reason for upgrades is to ensure compatability with new scanners. Rarely does the processing get upgraded for colour.

 

I am not familiar with the other two scanners you are contemplating, but I suspect they are not dedicated film scanners. If that is correct, you will be making a backward quality step. Other may have more and better advice on that. Whatever scanner, I would recommend sticking with VueScan, even if you use another as well.

[samwells]

Re scanners - I'd agree with Erl, the Epson V600 or V700 won't give you the quality of the Nikon dedicated film scanners. I have both the Nikon 5000ED and the Epson V750 Pro; the Epson is excellent for scanning prints, and for scanning MF negs, but it just isn't quite there for 35mm negs. There's a comparison review of the Epsons at the website photo-i.

 

cheers:

Sam

[Lucklik]

The scan from the lab is bad.

I use Vuescan for 6 month now and never went back to the Nikon scan software on my Coolscan 4000ED.

Since I set the exposure my scans have improved a lot.

I found this on the internet try it.

VueScan Exposure Tutorial - a set on Flickr

 

Succes

Luc

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[plasticman]

I also agree that the original lab scan is a poor guide. Typically the lab has applied a steep s-curve to accentuate contrast and make the whole image look 'punchy'. While you may want that as your final result (that's up to you), they've sacrificed a lot of detail and nuance in your image, and I'm confident you can do a better version if you follow your own interpretation rather than try to reproduce theirs.

 

Following on from Julian, I also want to strongly recommend CFSystems ColorPerfect. If you capture a linear negative scan from your negs, even the default settings from ColorPerfect will give you better results (imo) than hours of fiddling with curves in post.

 

Use the settings outlined here to get the optimum original, then convert to positive in Ps with ColorPerfect.

WARNING: ColorPerfect is probably the least intuitive and most impenetrable, non-standard application you will ever see - but it seems to me this is because the developer is a genius with color and not with the way that normal mortals interact with computer programs. As I said, even the default settings will outshine the version made by Vuescan or Photoshop.

 

Here's a quick comparison screen-grab of the way that an orange toy car in evening light (an incidental background detail in a randomly selected shot, incidentally) looks when output (left to right):

 

- Ps invert from negative

- Vuescan's own positive output

- ColorPerfect at default settings

 

ColorPerfect's color is by far the closest to reality, and has more detail and contrast too.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[andym911]

had to ahve a go.(LR )..here my result that looks ok to me but there again I don't know what the scene was like.

 

best

andty

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!