First scans with Epson V850

[Steven Ruttenberg]

Thanks for the feedback.

I agree it is bluish. It was a lot worse. It was very cool color temp that evening. The trees loom as I remember, not exactly white. I could remove more blue but it would take some work to not rob it from other areas. I tried the colorneg plug in for ps and it made it look ridiculously warm. As for shadows, couldn't tell you they or aren't black cked up. The lens I used doesnt quite cover 4x5 at wider apertures.

I like the image and it looks good on my monitor, but agree too it can be better. I already think compressing it from 1.79gb to 10 mb and then uploading to another site might have something to do with it. Again, thanks for comments. If you want the fiddle with file I can upload original somewhere for download. Always looking to better my images.

[Pere Casals]

Thanks for the feedback.

I agree it is bluish.

Steven, that is not bluish, the image I post is bluish

But the original Velvia slide was as bluish as the scan. I could use a warming filter... but blue can be nice to depict reality.

The Epsons are IT8 calibrated so color fidelity has been calibrated and that fidelity and can be easy checked by scanning an IT8 traget. Thanks to god IT8 targets were made with kodak or fuji film, this helps to avoid spectral missmatchings.

It is true that expensive scanners have better image enhancing software that is more productive when wanting nice images for customers wanting a final image. Thta's amazing for Pro service.

IMHO best choice for a film photographer that wants full control for the job is having the scan as raw as possible, with no sharpening or color enhacements, etc, nothing, and managing all that with Photoshop or the like.

For awesome creative color management there is for example 3D LUT creator, https://3dlutcreator.com/ if one wants full color control beyond Ps sliders a 3D LUT editor it's the most powerful tool. Some may say that an IT8 clibrated scanner delivers nicer colors that another IT8 scanner, but at the end with a 3D LUT editor one can obtain exactly what one wants, single challenge is having a wonderful artistic/technical criterion to steer the result to the intended direction.

I'd like to add a personal feeling, as time passes I'm less a fan of digital adjustments, digital flexibility it's powerful, fast and very Pro, but IMHO nothing compares to a 810 Velvia slide on a light table. It's raw, it's authentic, and it may contain such an amount of beauty that it may make a hard man cry.

[Ted Baker]

Here is an example of a 1.79GB file processed in Camera Raw and PS. Scanned on Epson V850 at 4000dpi, AA turned off, 16 passes and multi-exposure.

Doing the inversion that way by hand is quite difficult. Apologise if you already understand this, for example by just using the RGB adjustments alone you will never be able to match the sort of result you can would get from an optical print. These are 1D transformations, ideally you need a 3D LUT/ transformation, or some other approximation. Transparencies are much easier in this respect, as you are trying to match the original or something close.

[Steven Ruttenberg]

I will look at the 3D lut. I say bluish in the sense that as pointed out the trees are not as white as in daylight. But the image itself has quite a blue cast after scan and conversion using vuescan. The color neggoes the complete opposite and makes it too warm and saturated by default requiring a lot of work as well in ps and colorneg.

I am by no means even remotely close to being an expert on this stuff and I realize that I don't know most of this as it has been almost 20 years since I stopped using film. Any advice on work flow etc isalways welcomed.

One question is how to best invert a color neg?

[Ted Baker]

One question is how to best invert a color neg?

IMHO If you aiming to reproduce something like a traditional colour print, then unless you a very skilled and know what your doing then you need some help in the form of software. Some people speak very highly of color perfect, some like silverfast etc. You may indeed just like epson scan as a starting point. If you goal is to make something that is pleasing and your very own, then you can use whatever you like, but you will need more than straight RGB adjustments.

Sorry I did not mean to be deflating, but It's often more helpful to understand the difficulties that to just pretend that it is simple when its not.

[Steven Ruttenberg]

IMHO If you aiming to reproduce something like a traditional colour print, then unless you a very skilled and know what your doing then you need some help in the form of software. Some people speak very highly of color perfect, some like silverfast etc. You may indeed just like epson scan as a starting point. If you goal is to make something that is pleasing and your very own, then you can use whatever you like, but you will need more than straight RGB adjustments.

Sorry I did not mean to be deflating, but It's often more helpful to understand the difficulties that to just pretend that it is simple when its not.

I agree that this is not simple. In fact, part of the reason I went to 4x5 is precisely that. You also get images that are more unique and different than digital. To me, digital is easier. I'm not saying a digital file can't be screwed up, but with film you have to know what you are doing from the time you unpack the camera to printing or scanning in your file and post processing in some software.

What I am looking for is a method so to speak to get a a color negative "properly" inverted so I have a good starting file to work with like in digital, if you have taken the picture correctly you have a good starting point. I realize that my vision will not be most peoples and what I find looks good others may find garish or terrible. As and example, I'm not a particular fan of Adams, but I do recognize his talent and skill.

For sure I'm no expert, not even close and willing to learn and try and develop into the best possible photographer that I can be.

So no, you're not deflating, I like honesty and good information so I appreciate all the comments.

[Pere Casals]

Doing the inversion that way by hand is quite difficult. Apologise if you already understand this, for example by just using the RGB adjustments alone you will never be able to match the sort of result you can would get from an optical print. These are 1D transformations, ideally you need a 3D LUT/ transformation, or some other approximation. Transparencies are much easier in this respect, as you are trying to match the original or something close.

I ask, (because I've seen you know a lot more than me about that) perhaps a good way to obtain a 3D LUT would be scanning a color print, scanning the negative that resulted in that print and calibrating a 3D LUT that would take the negative scan and it would deliver the print scan.

Of course that print/negative may not contain all color space, but if color spaces are deformed to make the thing match then the LUT conversion would be near from what we want. Of course the print could had some color balance, contrast, etc applied, but these are minor adjustments.

...but anyway perhaps we can take a calibration target, shot it in daylight, obtaining an "standard" RA-4 print, and using the negative and the print for the calibration. This can be done for diferent color films.

It is clear that negatives have more DR than prints, and the print may have blocked part of the negative's DR, so it would not be a direct result, but the core of it...

[Ted Baker]

I ask, (because I've seen you know a lot more than me about that) perhaps a good way to obtain a 3D LUT would be scanning a color print, scanning the negative that resulted in that print and calibrating a 3D LUT that would take the negative scan and it would deliver the print scan.

Of course that print/negative may not contain all color space, but if color spaces are deformed to make the thing match then the LUT conversion would be near from what we want. Of course the print could had some color balance, contrast, etc applied, but these are minor adjustments.

...but anyway perhaps we can take a calibration target, shot it in daylight, obtaining an "standard" RA-4 print, and using the negative and the print for the calibration. This can be done for diferent color films.

It is clear that negatives have more DR than prints, and the print may have blocked part of the negative's DR, so it would not be a direct result, but the core of it...

Up to very recently I did not understand any of this, even though I have been photographer shooting negative stock and also a programmer for more than a little while. I was inspired by some of the open source tools for the pure digital photographer, but a little disappointed with what was offered for those who still want to shoot film. So I naively thought I could write some of my own tools to fit between the scanner and post production tools. To that end I spent a lot of time researching the topic, of which I probably understand about half

Thought It perhaps should not come a surprise that a lot of the best software techniques/calibration/data etc is kept secret...

Anyway what you describe is very valid technique, and IMHO conceptually the correct way to approach the problem. It is essentially the last step in the chain of Kodak's DPX/Cineon digital intermediate system when targeting digital projection (this is the system which is used in the motion picture industry for the last 20 years).

I finished my first prototype, that I a currently using with some success (its not properly calibrated though), that I intend to rewrite and then release as an Open Source set of tools. If anyone wants to send me a few negative scans I will invert them. Linear scans from a scanner with part of the film rebate or DSLR scans saved as RAW, just send me a PM, to work out the details.

[Steven Ruttenberg]

Reworked image of Aspens from earlier.

This has a faded grid box on it because I am trying out a software to convert negatives to positive instead of relying on photoshop invert or the scanner software to do it. So far I like it. This is a test image. I didn't delete any dust yet as I am still playing with it. I don't like the trees in the center that seem to have a fog over them. Not sure if that is result of scan, post-processing or an actual part of the image. It has promise. The software is Color Perfect. So far I like it, so most likely will get it and that will remove the grid from images. Anyway, I can find plenty to fault on this image, at the same time, it looks pretty good to me for having only been working with film again for a short period of time. The learning curve is much steeper that digital imo. But getting there with everyone's help and input.

The one thing I don't like is the fog stuff over some of the trees in the center of the image. Not sure if it is from scanning incorrectly, the process in converting to a positive with Color Perfect or just what the image is. I didn't remove and of the dirt/dust since this is a practice image so when I buy the plug in I will work this image some more and remove the dust. Also, I have a second image taken right before this one wher the sun was just slightly above the trees and a slightly different exposure. Needless to say, shooting into the sunset and strongly back-lighting the Aspens just make this a pretty technical shot to get correct for later processing/printing. Compared to the previous one I thought I liked, this one is much better I think.

Interested in thoughts on scanning, using Color Perfect and what needs work, including the shot itself.

[Ted Baker]

Interested in thoughts on scanning, using Color Perfect and what needs work, including the shot itself.

Looks much closer to what I imagine Ektar looks like. (Never having shot it).

It may also help to shoot a roll of Ektar if you have a smaller camera and have it scanned with a noritsu or fuji frontier scanner or have a few optical prints made (Portraits are good for this). Then scan it yourself and see how close you can match it and if you can make it better or different etc etc. No need to spend a fortune.

Or even some old negs that you have prints from from yesteryear can be good for this.