Hi Steve,
Just to put things in perspective. I am mostly a landscape photographer. It may be that a lot of landscape photographers adjust colours, but I don't have the slightest idea what colours look like. So, I want to have the colours of the camera as close as possible to reality.
I had some discussions with a guy in GB called Tim Parkin, he has a landscape photography site called OnLandscape. He is a photographer, writer,
painter and mathematician.
Tim found that the P45+ had about the worst colour of any camera. Tim works often with Joe Cornish and they had long discussions about fixing color coming of the P45+. I was suggesting that the issue could be handled by profiles, but Tim did not agree. I did dig into that a bit.
One of the aspects I stumbled on was that vegetation has high content of IR. The ColorChecker vegetation patch has no IR. So a weak IR filtering would have an effect real vegetation images but no effect at all on the vegetation patch on the ColorChecker.
Later, I have noticed that the P45+ has some tendency to browns on perfectly black cloths, also indicating weak IR-filtering.
When Joe Cornish started shooting with DALSA based backs the issues went away. I still have my P45+ back. At one time I was preparing an article for OnLandscape and made some test shoots. With some interesting results. I used a deep blue purple flower as a subject, as I knew that it was a kind of difficult colour and it also had green leaves.
The P45+ delivered this image:
I of course also shot it on my Sony Alpha 99, yielding this image:
One may be better or worse... But, those images were processed in Lightroom using colour profiles generated by Adobe DNG Profile Editor.
So, what did we get using Capture One, this was P45+:
And this is what Capture one made of the Alpha 99 image:
What did that flower look like. Here are the correct colours for the petals:
And for the leafs:
So, I guess that Capture One missed on that scene miserably.
Now, getting back to the IQ Thrichromatic. Phase One published some info about the Thrichromatic that made a lot of colour scientist raise their hair in astonishment. So much that Jim Kasson and Jack Hogan developed a set of programs for comparing CFA designs.
https://www.dpreview.com/forums/post/60253992https://www.dpreview.com/forums/post/60278621One of the early findings was that this response curve was optimal:
Well, that is the response curve of human vision. For some technical reasons, it is not very practical.
Now that Jack Hogan had some emulation software for analysing data he did some analysis based on two sample image from "DChew" and found that the IQ3100MP back is more colour accurate, but the Thrichromatic has a compromise matrix that is more ideal.
In DChew's images I have seen the brownish contamination of leafs I would attribute to IR in the IQ3100MP, while the IQ3 Thrichromatic had cleaner greens. That may indicate that IR cut off frequency is steeper on the IQ3 Trichromatic.
Usman Dawood published some comparisons between the IQ3100 and the Thrichromatic and showed that the IQ3100 had bokeh fringing that the Thrichromatic did not display. That is a strong indication of a different IR/UV filtering strategy.
If the Thrichromatic has more accurate colour, I would assume that "Digital Transitions's Cultural Heritage division" would switch over to it. That may of course be the case they do.
Doug indicated that "lime greens" are better handled by the Thrichromatic, and that is indeed visible also in "DChew's" test. I bought some limes and green peppers etc and made a set up with some similarity to Doug's and shot it with three cameras I have:
- P45+ - that was known to have problematic colour
- Sony Alpha 900- same generation as the P45+, but known to have excellent colour renditon
- Sony A7RII - my most used camera right now
All three handle lime greens perfectly well, providing perfect match with measured data from the limes themselves. So, yellow contamination was no issue under the test illumination that was studio flash.
So, to sum up. The improved IR and UV filtering are probably useful. The only real world data that we have indicates that the IQ3100 MP is the one that yields the more accurate colour.
Best regards
Erik
Yes, but what I am saying is that there are a ton of photographers who will never utilize objective - based color measurements on the way to final production. In fact, probably the majority of photographers evaluate color visually, and that's it. Certainly this leaves room for error, given the wide variety of devices and conditions all these photographers render and view in (not to mention their own visual acuity). And they'll never change.
But for someone who wants to compare an IQ3 100 to an IQ3 100 Trichromatic, I am saying it is possible to arrive at a true and accurate conclusion without objective measurements if the devices and conditions are optimal and consistent. That's the whole point of you showing the image samples, is it not? We produced our own image samples and from our own CI samples, I could see the difference and the difference was in favor of the IQ3 100 Trichromatic, as your article did, just with different data.
Steve Hendrix/CI