ColorNeg

[jag2x]

Hi all,

Just stumbled on a software package called colorneg
http://www.c-f-systems.com/Phototips.html
I was wondering if anyone had used this product? You can take a positive image and invert it to make it look like a negative based on a type of film.
Using this inverted image you can then use it as digital negative?
Have to test this out...
Jacek

 

[Eirik Berger]

I use ColorNeg the other way around, as it is intended to be used. I scan negatives as "positives" (16-bit linear scans) and use ColorNeg to invert the result to a positive image. Works very well with both color and BW neagtives.

 

[Robert Budding]

I use ColorNeg for all of my negative scans. I never tried to invert a positive with it, though. A discussion that I started is here:

http://photo.net/digital-darkroom-forum/00MRuZ

 

[Robert Budding]

I've had good success with the CF System COlorNeg plug-in. But I also have color darkroom experience, so using CC filters is second nature.

I've started a Flickr user group to share tips and tricks here:

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

 

[jd callow]

I do it manually and it takes about 5-10 seconds to get the colour, density and contrast where I want it.

 

[erikhatt]

I do it manually and it takes about 5-10 seconds to get the colour, density and contrast where I want it.

Will you give a few hints on your manually process?

 

[pellicle]

Well I'm not JD, but I also do it manually (and get within ball park in about 20 seconds unless its a big file).

I have a few points on my blog (search on epson) but:

1. scan neg as positve (this causes hassles with my Nikon SA-21 finding the locations)
2. assign (NOT convert) profile to scanners appropriate profile (*)
3. invert
4. set each of black and white end of R G and B channels independently
5. perhaps tweak a curve of one or other if shot in the wrong colour temperature light
6. done

the profile will be the critical step, and depending on scanner will vary. For example I have my Nikonscan set to be bruce RGB, I have my Epson set for doing nothing and so I assign the Epson 4990 TPU profile (if using my 4990)

I have always found that the auto settings clip more that is really signal than I like and I prefer to use curves to settle this than hard clipping. Unless you like that look....

 

[jd callow]

I didn't mean to sound like a jerk. My point was that I tended to be able to get good density and colour balance very quickly from a raw scan. Like pellicle I use the scanner profile when scanning and convert as needed.

My work flow is:
1) Scan Pos
2) invert
3) find white and black edge of histogram
4) adjust density in curves by dragging composite curve from center (sometimes I'll do this in LAB on the 'L' channel)
5) adjust colour balance in curves by dragging channel curve from center
6) readjust density if colour balance sent the image darker or lighter than I liked
7) Somethimnes I'll adjust for global or local contrast in curves by either making a large 'S' for over all contrast or smaller 'S's in the appropriate section of the curve -- the former far more often than the latter.
8) very rarely I'll make localized colour adjustments (such as warm shadows)

The final piece of the puzzle are masks for localized adjustments on the image -- darkening skys, toning down hotspots, or to remove casts from mixed lighting. I use masks with curves for this. I know others who use dodge and burn.

Some of what I do is on adjustment layers and some directly to the image.
and that's how I do it.

YMMV

 

[pellicle]

Hi

I use masks with curves for this. I know others who use dodge and burn.

do you just draw in by hand and then feather the selection or how do you make your mask (just interested ... hopin to learn)

thanks

 

[jd callow]

It all depends. Some times i'll create a selection with a 2 or 3px feather and then add or subtract with much higher feather (or vice versa). Sometimes I'll create a new channel copy and paste the image into the channel and adjust the copied image to create a selection that I then use as a mask.

 

[erikhatt]

8bit

I use ColorNeg the other way around, as it is intended to be used. I scan negatives as "positives" (16-bit linear scans) and use ColorNeg to invert the result to a positive image. Works very well with both color and BW neagtives.

Sounds interesting to test. But my ICG 365 is only equipped with ScanExact 9.5 and deliver 8Bit Tiffs. Too bad i cant use tis, since it demands 16bit.

Can someone please give me av very simple and good definition of 16 bit LINEAR scan?