Kodak Ektar 100 - ugly colors. Advice needed

[babaluma]

It's very easy actually to get good colours. Do not use VueScan, it's a joke. Dumb film profiles and a load of pointlessly convoluted controls. All you need to do is hit the colour balance option in Epson Scan or whatever basic scanning package.


Colour balance fixes any perceived problems with sky and shadows too, it only takes 30sec - 1min. Though we should stop talking about this here.

I have tried to scan with Epson SW - result was about the same.
So, it looks like as it's not because of scanning software failures

[babaluma]

I have a little experience with the film, and it was very hard to get decent colors with conventional scanning software. There's nothing wrong with your scanner, of course. I advise you to get raw scans from VueScan and then invert them in PS using Curves or ColorNeg.

There are problems with the film after all. Or maybe just features. Color of pale sky usually shifts to cyan. There is blue cast in shadows (I must admit that sometimes it does look great). The blue cast is severe in case of underexposure, so you may try to underrate it a bit.

Good point,
i have also been thinking about working with raw output rather then with Jpegs

[babaluma]

You didn't post anything?

Sry for another delay, - now i did.

[Poisson Du Jour]

*** Dupe post: deleted ***

[Poisson Du Jour]

The overall pallete is quite ghastly, particularly the articulation of red and the almost Kodachrome-like quality of pale sky.

Firstly, its important to reinforce something: scanners are not "OK". All scans of film will require work. My Epson V700 has been found by my lab to have a moderate blue displacement. I use Photoshop Elements 8 for post-scan and print RIP work to TIF files.

How are you setting the scans for viewing — on the web or via hard printing?

You say in your post that it was evening and a bit red, and you point the finger at the film that, probably by dint of its design, "lifts" the R/ channel higher (like Velvia does to the G/ and B/ channels), but it still has a very peculiar blue shift. So in effect, this lift is perhaps not satisactory to you?

Colour is always set for the destination device (e.g. web or printer) when working on scans (via Photoshop or some other software). Additionally, the rendering intent must be correct (absolute, relative or perceptual), as must the profile, all again to match the destination device and desired use. If not, the colours you see at your end will not be the same as the printer or screen (not all screens display colour correctly) — it's not rocket science to line everything up, but it does require a bit of training.

The negatives are not beyound redemption. Straighten, correct colour, crop and reassess.

[babaluma]

The overall pallete is quite ghastly, particularly the articulation of red and the almost Kodachrome-like quality of pale sky.

Firstly, its important to reinforce something: scanners are not "OK". All scans of film will require work. My Epson V700 has been found by my lab to have a moderate blue displacement. I use Photoshop Elements 8 for post-scan and print RIP work to TIF files.

How are you setting the scans for viewing — on the web or via hard printing?

You say in your post that it was evening and a bit red, and you point the finger at the film that, probably by dint of its design, "lifts" the R/ channel higher (like Velvia does to the G/ and B/ channels), but it still has a very peculiar blue shift. So in effect, this lift is perhaps not satisactory to you?

Colour is always set for the destination device (e.g. web or printer) when working on scans (via Photoshop or some other software). Additionally, the rendering intent must be correct (absolute, relative or perceptual), as must the profile, all again to match the destination device and desired use. If not, the colours you see at your end will not be the same as the printer or screen (not all screens display colour correctly) — it's not rocket science to line everything up, but it does require a bit of training.

The negatives are not beyound redemption. Straighten, correct colour, crop and reassess.

Hi Poisson Du Jour,

Those scans were for web, but i think it should be fair to try to print this out and check a difference. Even with all monitor failures this film is obviously very velvic for such scenes i.e. it will be very hard to get resluts with soft evening natural colors like this (Ektacolor pro 160):

[Poisson Du Jour]

If your scans are solely for viewing on the web, you accept that colours may shift or be displayed differently from one monitor to the next. That's not your fault, of course.
Maybe you should align your scans for prints to give you an idea of the many factors that come into play that govern how well a scan displays vs prints. Film scanning is not an art, but it does require a lot of patience and observation. What I see here is a reasonably attractive representation of the image in terms of colour and tonal gradation, and a subtle renderingn of yellow and red. Is there anything else you were attempting to achieve by way of colour nuances?

[hpulley]

Hi hpulley,
thanks for a great example.
What kind of film have you been using?

Those examples are both Kodak Ektar (100) 120 in my Mamiya RB67 ProS.

I've also been shooting a lot of Fuji Reala and Kodak Portra 160NC lately when I shoot color, which seems to be the majority of what I shoot with spring flowers and short sleeves coming out (I've been shooting Ilford Delta 100 and 3200 and FP4+ mostly of late when doing B&W).

[T0MA]

I did some fast testing and here are my results. I measured sunny area at front and took three shots. I scanned to 16bit tiff and used negfix8 script for positive conversion. From those I would say that overexposing looks best and has least "blue" cast. All images are directly after negfix8 without other modifications.

Now I have to say that actually all looked really good after I just played bit with blue color curve and I got pretty much those color that I was expecting. Also those were really close to prints I got from local lab. So I would say that bit blue color cast will show easier in dark areas if just some basic color manipulation is done (like invert and autolevels). I suppose good scanner software with profiles for current film would handle this without problems also but currently I don't have those.

So I would conclude that overexposing bit would maybe make scanning easier. Also getting rid of that color cast was not too difficult at least in this case so even underexposing did not cause any more trouble.

[Photo Engineer]

This is how my software handles that example.

No doubt a wet print with suitable exposure changes would render them all virtually identical. I have done it with Portra VC 160. Examples are posted here on APUG.

PE