CF Systems ColorNeg

I wouldn't regard CMYK (i.e., CC filter equivalents) as an improvement to the usual RGB work flow in Photoshop. I doubt there is a simple correlation to the filter stack on a conventional enlarger, which must compensate for the enlarger itself in addition to the negative. Frankly, Photoshop has all of the tools I need in this regard, along with calibration, to make transportable image files.

Edward - I'm looking for opinions from people who have actually used CM Systems ColorNeg software. As it is, you are posting an opinion with no data or experience to back your assertion.

I downloaded the demo and used it with my Minolta MultiScan Pro. The Minolta software is useless for scanning negs hence my interest. If you have a lot of different types of negatives then it could save a lot of time. As for me I standardized on just one type of neg (Fuji Reala) so my Photoshop action gave me the same results as his plugin. So I didn't buy it.

Jim

Thanks, Jim. I'll be scanning a lot of my father's old negs. Many film types across 50 years of shooting.

I'm sort of in the same boat as Jim Hein: just downloaded the demo and am trying it out. My main motivation was less than satisfactory highlight detail, when I try to scan as a slide and invert scans myself. CFSystem describes Photoshop invert as a linear process, and says this is not what you need to do.

I'm experimenting with some Fuji Reala, shot in daylight with a grey card in most of the shots. It seems to be giving me very good skin highlight detail, and the grey cards are coming out neutral. Kodak Gold25 seems a nice choice for profile, regardless of the film being Fuji.

FYI, I'm getting my "linear" scans via Minolta Scan Utility with my Scan Elite 5400, following the procedure CFSystems describes for acquiring the file with my scanner:

Scan as slide, auto exposure on, and output 16 bit linear.

The one alarming thing is the spikes at both ends of the histogram. I've tried it with black and white neg scans, and not sure if I like the results, looks to be a lot of posterizing in the highlights, could be my settings though.

I've been using it a lot lately, and it's by FAR the best way to scan/correct negative film that I've found so far, though admittedly I haven't tried everything. They just did a major revision to it as well.

It still takes some time/skill to get the hang of it. My tip is to pick a neutral/colorless area carefully and then apply that setting to the whole roll. But it works. Cheap enough...

Mendel - have you checked your Gamma setting? It should be set to 1.0 to get a linear scan.

Could someone post a sample NepPos processed scan and compare it to another faster method they've been using.

I have yet after all of these years hearing about that site seen anyone post a sample.

No samples? See what I mean.

Here's a sample crop of 35mm Fuji Pro 800Z scanned on a Nikon 9000. I'll post others soon.

One more 35mm with Fuji Pro 160S.

One more for the archive - this German web site has a nice demonstration. Just click on the blue text to toggle the photo between the linear scan and edits made in Photoshop or ColorNeg. It works very well!

http://www.colorneg.de/oldneg.html?lang=en

Great shots, Robert.

Now could you just post versions of these shots rendered with the default convert to positive settings in the Nikon scanner driver without using ColorNeg so we can compare the results between the two?

Got your email BTW reminding me of your updated postings.

I'll do more scanning later this week.

I really like the work flow using ColorNeg. A 16-bit linear scan captures all of the information that's available, so I shouldn't ever need to rescan (unless I made an error). I just plow through the scans and leave all of the editing until later. I can always go back to the original scan and re-edit if needed.

Just a screenshot of the Nikon preview within the scanner software would suffice. You don't have to follow through on the scan.

Forgot to add just make sure you assign your monitor profile to the screenshot and convert to sRGB and post.

I'll do the scan too, since digging out the neg takes the most time.

I've rescanned the second image as a negative using Nikon Scan. I couldn't get the colors even close.

I added a lot of red to the Nikon Scan image. I even tried to set his shirt as neutral, which is what I did with ColorNeg. I did want an overall warmth to the photo, ans I don't think I achieved it with Nikon Scan. Perhaps someone with greater proficiency could do better.

Yep! Typical neg scans I get on my Epson flatbed using Epson Scan. That indescribable sickly cyan tint permeating throughout the entire image like some kind of undercolor. It's like a bad stain you can't get completely rid of even using Bon Ami type edits. Screws up the image's color table.

You've sold me on ColorNeg with this one. Thanks, Robert, for taking the time to post these results. Much appreciated.

Glad to help, Tim. The plug-ins are very inexpensive - particularly when you balance the cost against the output.

Sorry that it took me so long!

Tim - make sure that you scan as a positive and set gamma = 1.

ColorNeg corrections behave just like the color filters that I used when printing in a darkroom. I haven't seen anything that works as well.

Can anyone compare this to the VueScan workflow described here? Is it noticeably better? (link)

I tried the work flow that you linked to. It's a pain in the *ss, and the results are dreadful!

Just wanted to add the generating previews from my Nikon 9000 took forever with Vuescan. I consider the software to be only marginally usable with my scanner.

Robert,

Wonder if using Silverfast SE's HDR setting would give me a truer/better? linear scan over using Epson Scan?

Thanks for the response. I guess this is the best solution.

Just tried the ColorNeg trial and wow, night and day. Now I just need to decided to give up food or film this week.

Tim - There are directions on the CF-Systems web site on getting good linear scans with the Epson software. No directions for Silverfast, but I'm sure that you could find similar settings options.

http://www.c-f-systems.com/Scanners.html

Kyle - Give up food or film? Why? ColorNeg only costs about $70. Best $70 I ever spent on scanning!

Robert,

What I find on that link is settings for Vuescan. There's nothing for getting a linear scan using Epson Scan. But I think Silverfast HDR will work just as well and it's free.

What puzzles me about the makers of ColorNeg is, if it's this good of a program for scanning negatives, why aren't they tooting their own horn offering sample comparisons like you just did in this thread. It just seems like a strange way to market a new and powerful application seeing all the trouble most folks have scanning negatives.

mmh, anybody installed ColorNeg on the Mac OS X platform? Any problems?

The reason I ask is I just read the installation guide and it requires auxiliary data files to be dropped in the folder structure within the Package Contents of PS 7 and CS for it to work.

And for CS2 and CS3 you have to create a clumsy workaround editing the data path of an instruction file within the root folder to get ColorNeg to write to the correct folder. Which root folder the instructions are talking about and how to edit data paths is not mentioned.

Sounds like the ColorNeg makers don't know how to write for the Mac OS X/Photoshop platform. I think I'll pass on this until I get a confirmation from another Mac OS X user that it works without corrupting things.

I'm sure this is a good and useful program, but those install instructions scare the crap out of me.

The Mac version is fairly new (I'm using a PC). You should drop them an email - David has been very helpful in answering my questions.

Tim - you're right. I thought there was info on setting up Epson Scan, but there isn't. I'd suggest that you just go into the software, set gamma = 1, scan as a positive, and turn off all auto corrections (ICE is fine to run). And scan in 16 bit mode.

Why isn't ColorNeg promoted better? Well, David is an engineer, and he wrote this program because he likes photography and wasn't getting scan that he liked. He's definitely not a marketeer, and he isn't inclined to blow his own horn or to try to win over folks who are happy with the results they are getting from other programs. Really nice guy, by the way. He'll respond to your questions and he's very generous with his time.

Then there's me. I tried his program and I really liked the results. So I'm sharing my experiences with anyone who might be interested.

Robert, from your results above I'd say you neglected to set white balance as described in the 'Using the Histogram' section of the linked article.

I set his shirt to neutral with the dropper tool. But perhaps someone who uses Vuescan regularly could do much better. Would you like me to send you a Vuescan RAW scan for editing?

I've used Color Neg and find it works much better than the software which comes with most scanners. The problem I have is that it still doesn't have wired in what's neutral- you still have to manually tweak scans.

If I have to tweak anyway I prefer to just scan flat (but in gamma 2.2, color uncorrected) and use generic levels presets created for each type of film to get in the ballpark.

This is still the method I use when I scan film (but now I process film through Lightroom the rare times I use it). http://www.erik-krause.de/tutorial/part1.htm

I used to have to tweak color prints that I made in the darkroom, too. So, for me, a few adjustmens don't create a problem. I tried making adjustments with the tools in Photoshop, but I have found that I get better results with ColorNeg.

Robert

I've been testing it a little, and find that it gets me "nearly there" with less clicks than my manual methods

My method is divided into two sets:

1. Scan as colour negative in software
   • set the black and white points manually for each channel as the software is usually too conservative
   • apply curves to the image to get the look I like
2. scan as colour positive (linear gamma)
   • invert in photoshop
   • set black and white point cliping levels for each channel
   • apply curves to the image

I do all of this in layers on a smaller res scan, then apply the layers to a higher res scan (so I don't go mad waiting). I find that if I don't like the first method I then go to the hassle of the second. Sometimes it works better one way, other times another.

I'm thinking of spending the money on the ColorNeg plugin, but I stay my hand because it seems to clip the shadows more than I'd personally like.

The amount of clipping is adjustable. Launch ColorNeg and select the 'Tails' radio button. You'll see a box for H% and S%. This is where you set the percentage of shadow and highlight pixels that are clipped. You can't set to 'o', but set to 0.0001 and you have effectively no clipping. THe other settings let you define where the s-curves begin.

Another feature is the ability to make independent adjustments on selections. This enables separate color adjustments when needed and makes it possible to dodge and burn. These are functions that I'm really just starting to explore.

Just be sure that you get all of the info available in your negs when you make your linear scan as a positive.

Robert

thanks for the tip. I've not 'read the manual' yest (I'm an IT guy so "who reads them" right?). I found that adjuster and tried it but could not get (visibly) the results in the preview.

I do try to get all that I can in my scans from the start. I'm discovering that there is more density in some colour negatives than in my black and white negatives. I suspect this results from the spectral absorption of the dyes rather than true 'density due to turbidity' as in silver based films.

Here's some of my digging into the limits of my scanner and into the colour negative. If you spot any mistakes or errors (noone's perfect right?) please drop me a line.

:-)

Chris,

Since you used a stepwedge to establish linear characterization of the Epson scanner, why not just use a basic tone curve to force the scanner to reproduce the densities exactly as it is in the stepwedge and save the setting as the optimized linear response of the scanner?

As long as the ends aren't clipped scanning this target, you've got plenty of data to work with. It just needs to be redistributed to for optimal separation of tone.

Tim, I think that's basically good advice.

I do it using a QP card + black target (black, white, gray and then inside of a lens cap, etc for true black). I also set R=B=G for the QP Card points to deliver basically neutral results.

I do find that depending on the scene I will want to further optimize the tonal relationships as the negative can record a density range a lot more intense than paper white to true black.

Roger

Sounds like a good method, Roger.

It seems it's all about getting past the limitations of the medium and capture hardware to get what you want, not what theory dictates. Density ranges in particular 35mm negatives are so hard to determine viewed with the naked eye even with a light table. If the Epson scanner can distinguish that many levels as illustrated in the stepwedge, I don't see why it can't grab all the data in a negative.

Something else about Chris's other blog concerning the inadequacies of the blue channel in negatives got me to thinking about scanning negatives in general.

Have you noticed the lack of contrast when viewing a negative either physically or through the computer display within a scan as positive preview? No matter how light or dark the negative is overall whether from under or overexposure through the scanner or from processing, when you invert it or turn it into a positive by any means, the lack of contrast still prevails just darker and lighter versions. Saturation level is very much controlled by this as well.

The histograms of each channel reflects this lack of contrast by representing them with somewhat narrow, conical shapes. Just because the blue channel is shoved all the way to the left doesn't mean there's not enough blue data (tonal variances) in the image. It just means contrast needs to be applied. Saturation levels of the scan as positive negative will also shove the blue channel too much to the left because the color orange from the orange mask requires a lot of yellow and the more saturated the orange the more the blue channel gets shoved to the left in an RGB encoded capture.

Other than that more importantly I would think being able to capture the AMOUNT of tonal variances from dark to light in each channel without clipping each channel would be the primary goal. Histograms will only tell you the clipping points and hue relationship but not the amount of tonal variances. Using this step wedge approach to get the scanner to represent these tonal variances by force using an optimizing correction curve might assure this though.

Hopes this makes sense?

Bill

heaps of points there, but I'll just focus on this area at the moment. The problem is that while it is true that the tonal levels exist there in the blue it is at once a narrow range and in the darkest area of the scanner. This presents a problem in the linearity of the response (as shown in the stepwedge study) but its worse as the quantization levels are very close together. Its significant just how close the data was between those steps in the dark region. This translates to more noise in the blue area (as seen in the segment below). Note that the left hand side is the view of the red and the right hand side blue channel (although the histogram in both remains for the blue).

Note the noise that appears in the pale blue sky?

Perhaps it would be useful to publish a zoomed in portion of the stepwedge, as I had a devil of a job distinguishing data in those areas and needed to make 'educated guesses'. If I was very thorough I would have published the 'errors' in a separate column.

I have been ruminating on this the last few days and I'm now working on an idea which I'll put to the test today.

I'ld have to see the final working scan as positive of your negative to figure out exactly why you're getting these types of histograms. I never got this scanning Kodak UC 400 on my Epson 4870 scanning as positive with either Color Control, ICM/Colorsync or No Color Correction selected in the Configuration tab. The scans were never as dark as yours. Viewed on a light table the orange mask was never that saturated either. Kodak UC 400's mask was a bit fucia/magenta-ish orange and about midrange in density. It wasn't dark and rich looking as the one on your blog link.

I'm confused by the red and blue visuals mixed with the lightened blue channel shown in the Levels depiction. I can see you've lightened the blue channel to show the noise because it's so dark, but don't understand why the red is so dark.

well ... its up there now

http://cjeastwd.blogspot.com/2008/10/epson-3200-more-optimal-scans.html

hope others find it as useful as me. I hope also it helps to show why the blue levels are what I was saying them to be (and how it is that I'm seeing what I see).

:-)

Interesting. I'm surprised there's that much scanner noise in the dense area of negatives with that scanner. I don't really have problems with my film scanner (Canon FS4000US) except for quite extreme examples (highlights of overexposed images). If there is a bit of sky noise it's easy to correct using selections and Noise Ninja. I don't want my sky noisy or grainy.

Have you tried increasing hardware exposure slightly even if it "blows out" some of your shadows?

Have you tried multiscanning or multiexposure to see if that changes your signal to noise ratio?

Tim

"That indescribable sickly cyan tint permeating throughout the entire image like some kind of undercolor. It's like a bad stain you can't get completely rid of even using Bon Ami type edits. Screws up the image's color table."

If you're using photoshop, have you tried using the Hue control? Select Cyan and then perhaps move it slightly positive (try +20 or so).

I tried to get the colors right in Vuescan - without success. Even if I could get the colors right, I still wouldn't use Vuescan because it takes forever to generate previews. The Nikon software is much better, particularly when i scan as positive and do all inverts and color correction with ColorNeg. ColorNeg is fast and easy, and I like the results.

Chris,

I get that cyan cast when inverting in PS a scan as positive negative after neutralizing using the levels eyedropper on the mask area along the borders. And sometimes I get it letting the Epson scan do the inverting using the Negative Film setting with brightly lit shots of wide gamut flowers with Kodak UC 400.

It's not like it's a pain to correct but I was under the impression using ColorNeg would be an easier one button click solution for setting up a more efficient workflow.

In fact I took Robert's first cyanish Nikon scan version and applied Auto Color with a bit of a tweak to the blue channel in levels and got more realistic color rendering in about 2 minutes time over his ColorNeg version, but that procedure may not work with the rest of the images on the roll because of variances in neutrality contained in each image.

I'm looking for a quick way to get fabulous looking color out of negatives with the least amount of time fiddling around edting each frame? I haven't found it yet.

One last try with the Vuescan image. I took it into Photoshop and set his shirt to neutral (same area that I used when making the ColorNeg adjustments). Not good. It looks as if the color integrity of the image was compromised when I used Vuescan. It could be user error.

If interested applied Auto Color and Fade To set to Color Blend to your original posted Nikon Scan As Negative and got this:

Here's the sickly cyan image corrected in Photoshop. It's totally correctable- try putting the gray dropper on the metal bars near the yellow thing in the background.

Here's how to do what I did: (link)

Roger - The cyan has been removed, but the skin tones still look wrong. That's what I meant when I said the color integrity seems compromised. I'm not a Vuescan expert and, from what I've seen, it's not worth the effort. It takes forever to generate previews, so I don't see that it's worth the time and effort to really learn to use it.

Took Robert's sickly cyan Vuescan version and had a heck of a time getting it to look as below.

1. Applied Snap to Neutral/Find Dark & Light channel and increased Red gamma in Levels.

2. Applied S-curve whose shape and slope was similar to Epson Scan to each channel. Too complex and took time finding the shapes.

3. Increased Saturation +34 and hue -3.

4. Pulled back the green in the middle of the curve and moved 0 point up 5 and moved the 0 point up 10 for blue but locked down the rest of the blue curve.

This will not work for an efficient workflow, but it's do-able. The color table of this image is wrecked and I can't figure out how unless it's because of being linear requiring a color profile to be assigned to it.

Tim - the Vuescan scan wasn't linear. I used gamma = 2.2 and scanned it as a negative. All corrections were done in Vuescan using the work flow that Kyle linked to above. I may have made an error, but I think I followed the directions as posted. Anyway, the results aren't great. But your corrections do get it closer.

Found an easier way with different results fixing the Robert's Vuescan version.

1. In Levels Red Hilite-246, Green Hilite-239, Blue gamma-.75/Hilite-218.

2. In Hue/Sat...Hue -5, Sat +35

And got these results...

Clearly there is no push button formula getting color right with negatives from all the color renditions of the same negative displayed here. I've had the same problems with my Epson using Epson Scan, Silverfast and scan as positive techniques. It's hard to establish a consistency in all this frame by frame.

Robert,

Does it really matter what technique you use? I mean you're still left hunting for the color by how you remember the scene. There's no reference. There's no pattern for predictability examining histograms to tell you what to adjust to get there since the initial preview generaly sucks and you don't know why. You can see what this causes by all the color versions produced in this thread. It's like color babble. Basically you're conducting photo restoration on each scan.

Some of my Epson scans turned out fine and some just looked horrible like the Vuescan versions posted here and there's nothing to indicate why. And even then when I go back some of these scans I THOUGHT looked fine still look off and then I'm back to tweaking. Maybe I'm just picky. I have neg scans where I have as many varied color versions as seen in this thread.

That's why I got a digital camera. But every now and then I get to thinking someone has come up with a secret way to get consistent results with negatives but then I'm always disappointed by the results when I try to apply it to my own negatives.

There's got to be a way to study a negative and derive some kind of understanding and method to getting good and consistent results so you don't have to slave on every frame.

"Tim : Does it really matter what technique you use?"

I would also repeat the same question because is something is paying well for me than I don't need to switch on another technology. I tried a very little the ColorNeg, but while the combo of Vuescan and Photoshop is doing enough for me than I left using ColorNEG. Yes it may be that I couldn't expertise that software so I didn't getting the desired results. That is useless for me at all while Vuescan is far better solution for me till now.

Tim

"I'm looking for a quick way to get fabulous looking color out of negatives with the least amount of time fiddling around edting each frame? I haven't found it yet."

About 20 years ago I spent a year using only black and white (deliberately to avoid colour). After photographing that way for the year I came back to using colour (hoping it would have encouraged me to compose differently) and I was shocked at how annoying lighting temperature was to deal with. Indoors with mixed lighting which were tonal breezes in Black and White became nightmares in Colour.

So I believe that there is in fact no easy way (certainly nothing automated which implies mechanical repeatability) which we can work with Colour. I think that digital cameras work wonders, but when comparing how much smack in the mouth punch I can get with colour negative I think its like trying to tame a wild and powerful beast.

I took some side by sides with what I cosider testing stuff for colour films. This one was done with the 10D the in camera JPG was blown, but I was able to pull something without washouts of the reds and yellows from the RAW (manually, and I'm still not 'satisfied')

and then this one (same day different place) with Fuji Pro160S in 120 roll.
At first its skys look 'washy' but then if you look at the digital above (in between the trees near the horizon) its actually the same there as here.

I just can't see how you could get software to automate this sort of thing and take into account issues like film stock variations colour temperatures, scanner and emulsion sensitivity changes with varying exposures/film densitys. If things behave linearly then we're in a better position to make adjustments, but if things don't (behave linearly) then ... well its by hand and eye :-)

I do like CF systems stuff by the way, and think it does a good job 'most of the time' ... if only that suited the way I like to get the most out of my films.

Wow, Chris.

That's a damn good shot with that 10D. Amazing color detail. I like it better than the bottom image.

Wonder how long it took you to get those results compared to the Fuji Pro160s version.

You have a bigger version of that 10D shot to post?

Bill

the 10D image is good isn't it, I was quite happy with it too. Bigger? Well I guess you noticed that it was a link to flickr? If you'd like something bigger than that one (1024 x something) email me direct and I'll send you a tiff in proPhoto.

In terms of time, well the preparation and the scan takes about as long as anything. The rough adjustemnts up there are perhaps 5 minutes work? I took the shot as a test though, so I wasn't thinking of making my reputation out of it ;-)

Hi,

I was just wondering if anyone here using a CS 9000 and Silverfast AI to scan 35mm & 6X6 color transparencies and B&W negs?

Need some advice on best practice.

Thanks

My point is that ColorNeg produces results that are very close to what I've been able to produce in the darkroom. At it's fast and easy. The tweaked images posted here do not look the same.

For the question of repeatibility, I'd reread the 12:49 and 3:19 posts.

I think this method can get you close: http://www.erik-krause.de/tutorial/part1.htm

Digital is easier to do consistently but film scanning gets easier with some practice.

I'm getting consistently good results making linear positive scans with Nikon Scan and inverting and color correcting with ColorNeg. I was simply trying to answer some of the questions that others posed, such as a comparison with Vuescan.

Chris,

I take it the "Bill" you're referring to is me, Tim. I didn't realize your images were flickr embeds. Didn't think to click on the image. The size is large enough, though. When you say 5 minutes of rough adjustments are you referring to the second image below the 10D image? Is that the average time you spend to get results of similar quality on each frame of an entire roll of negatives?

Just curious. Anyone want to submit their average they spend scanning and editing to final outcome an entire roll of negatives?

So if it is the nature of negatives to be this difficult in getting good and consistent results then why doesn't the movie industry just sell whatever software they use to get the great results they get 24 frames a second?

I can't believe they have to futz with each frame the same way that's been demonstrated here and in my own scanning experience. I'ld gladly pay around $200-300 for their app if it would make this experience easier.

And why hasn't Adobe come up with a better solution? With all the complex features they keep adding to Photoshop that hardly anyone ever uses with each upgrade you'ld think they'ld come up with something to make tackling negatives easier. Maybe a negative analyzer that reads the data intelligently and derives calculations taking into consideration the mask and color/density linearity. How hard can that be?

Tim - The soccer scan and the first shot of of my Pro 160S scan each took about 15 - 20 seconds to correct using ColorNeg. I used the ball to set the neutral point and then tweaked the light level in the soccer shot, and the shoulder area of the shirt for neutral in the second.. No fancy adjustments or jumping through hoops. That's why I like ColorNeg.

The linear scan itself was done in a batch of 12 and took less than 10 minutes (single pass, D-ICE).

Thanks for the detailed feedback on your scan times.

I take it you're satisfied with the final color rendering and that's all that matters.

So all 12 frames gave consistent results similar to the two posted here? If so, that's pretty fast turnaround.

"And why hasn't Adobe come up with a better solution?"

I would try loading your scans into Lightroom. Scan "flat" (don't clip white or black point) and you can do batch Lightroom edits that dramatically cut down your processing time. I'm not going to claim every image will look good in 30 seconds, but I wouldn't claim that for processing RAW files, either.

I've found that, if the lighting was constant during the shoot, that I can get very good results applying the same CC filter pack to the entire shoot. Then it's just a matter of very small tweaks if I want slightly different look for individual shots. Perhaps I could get to the same point with other tools. But I won't bother because ColorNeg is working very well for me. I am still learning to use it with selections, so I'll soon be able to make separate color adjustments when needed (mixed lighting really can be a pain).

Tim

I agree with Robert above "I've found that, if the lighting was constant during the shoot, that I can get very good results applying the same CC filter pack to the entire shoot."

and it also answers your question about the movie industry, they carefully (and have you ever seen how carefully) expose and correct for colour in the shoot.

Still, they have whole segments with which to apply colour corrections to.

I said that I spent 5 minutes to get it to there. That wasn't something I'd print though. I might spend 20min or so on something I'm really interested in.

But (unfortunately) I spend far more time spotting dust off (I don't have ICE on my 4x5 scanner)

Tim

sorry I got your name mixed up ...

I used to spend a lot of time correcting colors when I printed color in a darkroom. So even spending 20 minutes on an image that I want to be perfect seems like a very short time. Family snaps, on the other hand, now are much better than the colors that I used to get from any mini lab, and I only spend seconds making the corrections.

Robert

"Family snaps, on the other hand, now are much better than the colors that I used to get from any mini lab, and I only spend seconds making the corrections."

yep .. automatic stuff (even from the Epson scan) is pretty darn good, and certainly better than many prints I got in the 80's and before (even from the same bits of neg). For these things I find that ColorNeg does a one click (well pick the film base ;-) wonder.

its only the tricky stuff where you get out of the middle and into shoulder and toe where I find I need to do it myself.

Yes, Chris, there will always be some shots that take a lot of work. It's a good thing that I consider all this to be fun!

Some comments...

For background, I've been scanning films for years now, using a variety of scanners, and have untold thousands of scans under my belt. In other words, I do have more than a bit of experience. Current tools include a Nikon V and 9000, NikonScan, Vuescan and Silverfast AI studio

I'd tried the demo version of ColorNeg some time back and hadn't much cared for the tool, but having seen the interest in this thread, I thought I'd try it again. Unfortunately, I'm no more impressed now than I was the first time.

I'd love to find the "holy grail" of negative processing software - a tool which makes dealing with color a trivial process, but not believing in either the Easter Bunny or Santa Claus, I'm pretty realistic in my expectations when dealing with any package like this. Some measure of subjective user intervention will always be required. So what's my beef?

As expected, images are never quite right with the initial, unmodified application of the filter. They're OK, but nothing to get excited about. OK, no surprises there. So now, for me it comes down to a simple question: How good is the tool for making necessary adjustments and then evaluating those adjustments? And this is where this software falls flat on its face for me. The author may well be right in his contentions that his algorithms are superior to PS - but I'll never know, simply because the application AS A COLOR ADJUSTMENT TOOL just doesn't work for me.

The single biggest problem (other than rather obscure documentation which requires simple folk like me to do a bit of trial and error to figure out what some stuff really does) is that the image preview just isn't of high enough quality for me to evaluate whether subtle changes are really doing what I want. It's far too low in resolution, and when using high resolution files (21Mpixel in my case) the down-sampling algorithm used has such egregious aliasing artifacts that grain, even in the finest-grained C41 film, is accentuated to an extent that I can't really see what the tool is doing short of making grossly exaggerated changes. You can't make subtle color or contrast changes when you can't see subtleties. I can't evaluate what I've done until I hit OK, look at the changes in PS and then go back and iterate again (and again and again...).

I could probably get used to ColorNeg's way of looking at the world (though I doubt I could ever live with the preview quality), but I see no reason to try when there are better, easier ways to get where I want to go. I'm used to doing subtle color tweaks in PS, and get far better initial scans to tweak from using either the Erik Krause Vuescan method as pointed to by Roger, or even the latest revision of Nega-Fix in Silverfast AI.

Give this a better, more untuitive AI, a better quality preview and more intuitive, standard terminology (tails? - huh?) and there might just be a good engine lurking under the hood of this beast. But as is, I'll pass.

As always, YMMV, but there are my thoughts on the subject after two separate evaluations.

Scott

Thanks, Scott.

I was hoping a professional with your level of experience would chime in here and you've confirmed what I couldn't and saved me a lot of hassle and possibly further searching on the subject of scanning negatives in general.

What color is suppose to look like from a negative is subjective and so is the valuation of the software doing the conversion as well.

A couple of clarifications...

I said "Give this a better, more untiutive AI..." - obviously, this should read "intuitive UI". Good grief.

As for the experience I referenced, please keep in mind that my evaluation is still just an opinion. I mention the experience level mostly so that I don't get responses like "..are you sure you got a good 16-bit linear scan?" and so forth.

I really appreciate the fact that folks like CF Systems make their efforts available for evaluation, because trying it yourself is still the best way to evaluate.

And finally, I thought I'd attach a screen capture to illustrate some of what I talked about. This is Fuji Superia Reala, scanned on a Nikon V using Vuescan RAW. At this point, all that I've done is bring up the plug-in, select the film type, and autocolor.

As you can see, we've got a long way to go here, and as you can also see, the quality of the preview isn't going to help much. When I convert using these settings and see the image in PS, I see that I'm even further off than is obvious from this preview. Color is way off, and contrast is even worse. The image appears somewhat posterized as well.

In other words, there's much to do. Selecting a gray point (there are several potential candidates) gets me closer, but again - bring it in to PS, and I see that I'm still not that close. More than just minor tweaking is still required, in both color and contrast. And so on.

The handful of other images I tried gave similar results. Some were closer to start, some further away, but all required work, and trying to really see what I was getting was still impossible without executing the filter and evaluating the results in PS.

Scott

Grrr...

Messed up the attachment - trying again:

Double GRRR!

Wrong capture. I sure wish PN allowed editing. This (I hope) is the right one.

Scott

Scott - you should look at the CF Systems web site. There are versions that provide larger preview sizes. Overall, I do find the invert and adjustments closer to my darkroom experiences printing. That is my standard. And I've had better results with ColorNeg than with other tools. YMMV.

Scott

if you don't mind me asking, are you exposing for the shadows in those images? I haven't got access to a coolscan here so I can't try this, but I'm wondering about your method for scans of negs. Do you use the Nikonscan as negative or positive and invert yourself?

I have been considering getting a 8000 or 9000 for the 120 roll I do. For example, with my Epson (which is limited I know ... but I do 4x5 too) I get my blue information rather further down in the data range than I'd like. Now this is when I do a preview.

I get the feeling (based on scan times) that there is something Epson software is doing something with analog gains (either longer exposure for the Blue or something). If I make the adjustments as per the left hand side of the above image (rather than doing a full linear scan then fixing levels / curves in PS layers) I actually get rather significantly different results from the scanner.

Do you get neat noiseless scans of edge horizon sky (and clouds) like in the image above from your Nikon in the blue (with a linear scan)? Do you tailor your scan settings in Nikonscan?

thanks for your time

Scott - I looked at your web site. Nice shots, but I think we're each after something different in colors.

Robert, I have spent considerable time on the CF Systems website, but the availability of different sized dialog boxes is a well-hidden feature. If I have time, I might try one of the larger ones, but I seriously doubt it will help me much. A relatively minor size increase isn't the issue. The ability to zoom is needed, as well as better quality down-sampling, to allow evaluating the image as a whole with reasonable viewing quality.

As for your comment on us being "after something different in colors" - I hardly know how to respond. Regardless of how you feel about my view of color or I about yours, the fact is that I value tools that are flexible, and allow me to carefully and completely evaluate what I'm trying to do, and this tool simply doesn't get me there. The starting point is still typically way off (as it usually is with most color neg software), and the tools (the TOOLS, not the methodology - I really can't comment about that) to get me closer seem flawed to me.

I seriously doubt we're so different in our goals, even if our subjects and subjective choices may be dissimilar.

I'm glad you've found a tool that you like and that fits your way of working, but frankly, I just don't see the value that you do. The bottom line, however is still as I wrote above: I appreciate vendors like CF Systems who are willing to provide evaluation software, as the best way to decide about things like this is to try them for yourself.

Scott

Chris,

Work calls, but give me a little time, and I'll respond with some details.

Scott

Is the poor CF Systems preview quality related to the fuzzy gridlines they put on all images in the trial version?

Download page for different preview sizes is here:

http://www.c-f-systems.com/PlugInSizes.html

Scott - I agree that a zoom feature would be invaluable for evaluating images and setting a neutral point. ColorNeg isn't perfect, and the UI could be improved. But overall I like the results.

My comment on colors wasn't met to be negative. Your work is excellent. But I see somewhat pastel tones in your images. I've seen this in some of the j. holmes color space variants, too. It really comes down to differences that one might prefer in selecting films. It just depends on what you're after. To see what I mean, compare the grass color in the foreground of the following shot to the color of the grass in my soccer shot above.

http://www.westernsteamphoto.com/HTML-3/487-leaves-osier-ebound.htm

This will save you from scrolling up.

Robert

its nice that you post your images here, but if people were intelligently comparing your images, why would they not just save them (as they can be from the page) and then open them side by side in their editor?

Good point, Chris.

Roger,

I don't see any evidence of the gridlines in the preview, so I don't think that's the issue.

Any time you take an image as large as this one and down-sample this much, and especially if there's a fair bit of fine detail and/or grain involved, you need to be careful about the algorithms used, and you've absolutely got to do something about anti-aliasing. I think that's the issue here.

And while you didn't ask, I'll also mention that I don't find the gridlines in the filtered image at all objectionable for an evaluation version.

Scott

Scott,

I've seen your work posted in a film vs digital discussion while back and noted you as the train guy who creates such memorable captures of trains and the great southwest. Beautiful work. I still have a hard time distinguishing which of your gallery images were shot on film or with a DSLR.

I do see on some of your gallery images that you tend to lean toward green in some of your final renderings, but I don't think this has anything to do with negative conversion software, which can give far worse results than this subtle cast, but more to do with the eye's adaptation to editing brightly lit outdoor scenes. Robert's renderings are on the red side which is generally more pleasing to the viewer.

Do you rely heavily on clicking in areas to get an R=G=B? In my experience with my own negatives I've found this approach virtually impossible to get even close to decent results. I always have to do it by eye using levels to get a more realistic depiction of color temp cast from how I remember the scene. Most of your renderings nail outdoor neutrality without inducing overly bluish shadows, red to purplish magenta midrange neutrals and pinkish highlights typical of what I used to get from the lab.

I definitely bias shots of my kids towards the warm end of the spectrum. But that's because that's how they make me feel. I also don't have a very good monitor, even though it is profiled.

Robert, thank you for posting the link to the different preview sizes. This was one of the biggest complaints I had with ColorNeg. I've been using ColorNeg with my Coolscan 8000 for a while now.

The thing I really liked about ColorNeg was that I only had to scan once. My scanner is noisy so I don't like scanning all that much.

The one problem I had using Nikonscan was that I would take a series of photos in the same environment. After scanning, I would end up with photos where all of them were slightly different colors. If you look at just one photo, you couldn't tell, but if you scrolled through the series of photos, you can really tell. With ColorNeg, I use the CC Filter feature and now, my photos have consistent color to them.

It does take some time to get the color correct using ColorNeg, but now that the large preview window is available, it should make things much simpler.

Glad I could be of help. The plug-in really is very good. My next task is to learn how to effectively use selections in combination with ColorNeg.

Have you seen the German ColorNeg site? Not all pages are in English, but there are some very nice examples:

http://www.colorneg.de/oldneg.html?lang=en

Examples on 'Virtual Grades' for B&W scans is here:

http://www.colorneg.de/sw.html?lang=en

http://www.colorneg.de/warumcolorneg_sw.html?lang=en

http://www.colorneg.de/virtualgrades.html?lang=en

Thank you for the links. I haven't seen the German ColorNeg website. The first link was especially useful. ColorNeg has so many controls that I have only tried a few of them. I don't think I've seen the FilmData slider before. I'll need to check that out.

BTW, do you know of any tutorials that exist for ColorNeg? I wish there was a tutorial or step-by-step instructions on this tool. For the most part, I am very happy with it, but I want to get the most I can out of the tool.

I hope I am not going too far off topic, but have you also tried the other CF Systems tools? I am wondering if they are good? I have tried ColorPos for my E6 scans, but I haven't really taken the time to use it.

The German ColorNeg site really is good. Perhaps they should assume world wide marketing since the CF Systems site looks as if it was written by an engineer (it was).

I'm not aware of a tutorial. I've read the manual a few times, and I've payed around with many of the settings. And I've sent a few email to the author - he's been very helpful and generous with his time.

Hello all, I see there are people who are satisfied with ColorNeg. Me not. I cannot get the good results with ColorNeg with 80% of my negatives. I invert negatives manually after scanning as slides. Here is uploaded one of scans: http://www.box.net/shared/sgtqfpin7q ColorNeg process gives greenish look and too highlighted face. I set Fuji Superia Xtra 400 profile. Other profiles also not help. Here is the result I get manually: http://album.foto.ru/photo/311076/ The reason that I want to use ColorNeg is that in 20% it gives better results than I manually(and in no-time in comparison with my work). I want to get more or less good results with ColorNeg. I read ColorNeg manuals and tried modification of parameters in the plugin but could not reach good result. If somebody of you can get better results with my scan (again http://www.box.net/shared/sgtqfpin7q) please write to me how to do it.

I downloaded your scan - it doesn't look as if you've set gamma = 1. Take a look at the following page for directions:

http://www.c-f-systems.com/Scanners.html

I used your scan and processed using ColorNeg set to Fuji Superia XTRA 400. I used the black leg of her pant to set a neutral point, and I adjusted the lightness down to minimize clipping.

Make sure that you click on the radio button 'Color Adjust' before you set the neutral tone.

This was just a 'quick and dirty' adjustment. It still needs a bit of work, but it's much closer than before.

Robert,

Your version is a big improvement with a more realistic rendering with regards to color and tone.

A simple adjustment in Hue/Saturation can fix the woman's overly pink skin.

Was it that easy? If so, ColorNeg looks like a useful tool. Just hope it's consistent applying it to other images on the same roll of film.

Tim - All I did was open the image in ColorNeg Filter (Photoshop Filters menu), select the 'Color Adjust' radio button, and then I clicked on a neutral color (her pants leg). I then clicked on the 'Lightness' radio button and adjusted the slider to minimize clipping. Finer color adjustments are easy, but I just did a quick 'n dirty to get it close. It took only a few seconds.

I've been very happy with ColorNeg.

Hi guys! First of all - the image was scanned with 1.0 Gamma. After I wrote my question here I asked author of ColorNeg about this. He answered immediately and helped me. What I did I lowered lightness and began searching appropriate neutral point by clicking possible areas. The good point I found on head. it was black hair that yield color MUCH better. So now I know what to do. What is rest of correction after ColorNeg filter is orange or magenta cast which is easy to correct. I found that colors are better than made manually. Here is one example I got after http://album.foto.ru/photo/337146/ - with ColorNeg. Things get better now.:) Thanks to all !

George - the scan above looks excellent!

Yep, what I enjoy now is the full tonal range, esp. improved dark tones. Highlights are not the best, I got better highlights on scans from Noritsu(but shadows there leave more to be desired) And skin tones are more pleasant.

Rob, I got your answer by e-mail. By the way, it's me on a photo 3 years ago:)

George - you can make localized adjustments, too, with ColorNeg. I'm still exploring, but my understanding is that I can get the equivalent of dodge & burn.

Robert, I have read manual for ColorNeg, according to it there is a possibility to create a mask in Photoshop. Also Dunthorn mentioned this possibility when I asked him. I found creation of a mask on non-inverted negative not simple and did not try it. If you find how to do it you are welcome to share

I've started a ColorNeg Flickr group for anyone who would like to pose questions or share tips and suggestions. Example photos are encouraged!

Could you provide a link? I tried doing a search over at flickr using your name and got nothing.

My bad. Here's the link:

http://flickr.com/groups/977912@N23/

Here's a recent review of the ColorNeg plug-in.

http://dpanswers.com/content/rev_colorneg.php

Finally a decent thorough review of ColorNeg.

Just wish there was a way of getting rid of the cyan patina as evidenced in this thread and in that review. Other than that ColorNeg is supreme at squeezing as much color detail and depth out of a negative over anything I've seen. Note the two tone color effect in the back hills detail where both the Nikon camera and scanner software renders them with a single monochrome type color effect.

Thanks for posting this link, Robert.

As mentioned, the preview sizes are selectable. I thought I put up a screen shot of the ColorPerfect interface PrefectPCSizes on a 1920 x 1200 screen. For me 1600B works best. The image is a straight conversion without any processing, Fuji 800 ASA - hence a bit grainy - shot with bounce flash METZ 45CL-4 TTL as main light, ambient -1 or -2 stops (can't remember), Minolta 700Si with 50 mm at f1.4. Scanned with Minolta Dimage Elite 5400 at 2700 dpi, 16 bit linear as required.
-- Jan de Lange

A quote from the review:

"I used the Colorneg calibration without adjustments and spent perhaps a minute or so tweaking the colour balance and tone in Lightroom."
It really makes sense to make all color and tone corrections in ColorNeg/ColorPerfect. That would really be the true measure of performance.