Trichromatic vs 3100

[dchew]

I have the Trichromatic for a demo over the weekend, returning 11/28. I will be doing some comparisons with my 3100. If anyone has any particular tests they want me to run, let me know here. It is a kinda gray weekend, but you never no. Still some shades of brown around...

What I plan to do:

1. Shift to 18mm and check color cast.
2. I will focus on three lenses: sk150, 90hr and 60x. I also have the 35xl and 40hr, so will do a bit with those.
3. The Erik Kaffehr fruit test. I can scrounge up at least a lemon, lime, pepper, orange and apple.

I do not have critical color testing equipment, just a classic color checker and passport.

Dave

 

[Vivek]

Dave, Could you check id it has the blue-green UV/IR cut cover glass on it. The manufacturers claim thta the bayer dyes in them respond only to to the visible region. So, there shoudl be no need for the conventional UV/IR cut filter.

 

[dchew]

Dave, Could you check id it has the blue-green UV/IR cut cover glass on it. The manufacturers claim thta the bayer dyes in them respond only to to the visible region. So, there shoudl be no need for the conventional UV/IR cut filter.

Do you mean check visually? The sensors look the same and have the same cast when I view them side by side. But I have no idea how I would verify your question.

Dave

 

[Vivek]

Thanks. Yes, visually.

The claims are a bit too tall, I guess.

 

[jng]

Dave,

Ruh roh - sounds like this could wind up being an expensive weekend for you!

I'd be interested in seeing how the 40HR performs wrt lens casts with and without shifts. In fact, it would be useful to see the uncorrected LCC files themselves. Hopefully you'll have time to shoot some more interesting subjects as well.

Have fun!

John

 

[dchew]

Dave,

Ruh roh - sounds like this could wind up being an expensive weekend for you!

I'd be interested in seeing how the 40HR performs wrt lens casts with and without shifts. In fact, it would be useful to see the uncorrected LCC files themselves. Hopefully you'll have time to shoot some more interesting subjects as well.

Have fun!

John

Nah, I doubt it. But I will have fun. I will take the wides with me. I am shooting with John Lytton on Sunday. He's nearby for Thanksgiving so that will be even more fun!

Dave

 

[jng]

Nah, I doubt it. But I will have fun. I will take the wides with me. I am shooting with John Lytton on Sunday. He's nearby for Thanksgiving so that will be even more fun!

Dave

Sounds like fun! Give John L. my best -- from a fellow Pig!

John

 

[DougDolde]

So what do they want as a trade in difference when you swap your IQ3100 for the new model? Next it wlll be the 150 megapixel model and likely soon

 

[dchew]

I don’t know, Doug. I am afraid to ask.

 

[vjbelle]

Since you have the 35XL it would be interesting to see if CC is any better with the Tricho at 8mm.....

Victor

 

[dougpeterson]

Dave, Could you check id it has the blue-green UV/IR cut cover glass on it. The manufacturers claim thta the bayer dyes in them respond only to to the visible region. So, there shoudl be no need for the conventional UV/IR cut filter.

I can see where you got confused.

But it’s the response of G&B to IR and R and the response of R&G to B and UV that have been attenuated as part of the improved color filter design.

An IR filter is still used over the overall sensor package. It was never claimed otherwise.

 

[Vivek]

I was not confused. Rather, I read your report on this back accurately. A clarification there might be useful! Thanks!

I can see where you got confused.

But it’s the response of G&B to IR and R and the response of R&G to B and UV that have been attenuated as part of the improved color filter design.

An IR filter is still used over the overall sensor package. It was never claimed otherwise.

 

[0beone]

Dave,

Ruh roh - sounds like this could wind up being an expensive weekend for you!

I'd be interested in seeing how the 40HR performs wrt lens casts with and without shifts. In fact, it would be useful to see the uncorrected LCC files themselves. Hopefully you'll have time to shoot some more interesting subjects as well.

Have fun!

John

I have found that with both the 40 and 70HR's you cannot move more than a few mm without CA creeping in and the need for an LCC shot every time.
I think that with the impending 150MP backs at some stage, we will see Rodenstock create a new set of lenses with a greater image circle and/or a redesigned lens geometry to allow greater movement on tech cameras.

Cheers
Frank

 

[dchew]

Simple test in the shade. Alpa STC, 90hrsw. My interest is in comparing the two backs based on how I would actually use them in the field, which is not the same. For example, I would use the TC @ 35 ISO, so that's where I shot it, and the 3100 @ 50. I tried as best I could to balance the exposure for both. Brought them into C1 and white balanced on the second to left gray square, then fine tuned exposure by reducing the 3100 image by -0.15 stops so each gray square read 211. In all cases, the TC images use the canned "Landscape" profile and "film Standard" curve. In this first image, the 3100 uses my custom Lumariver profile and Film Standard Curve.

Here are the two full images (TC left):

First screenshot: TC on the left, 3100 on the right w/ P1 Outdoor Daylight profile:

TC on the left, 3100 with Leaf Landscape profile:

TC on the left, 3100 with my Lumariver profile:

I haven't had time to assemble any thoughts yet, except to say profiles really matter when discussing the differences between these two backs. More to come.

Dave

 

[dchew]

Here are some color tags. Top TC Outdoor DL profile, bottom 3100 Lumariver profile:

 

[dchew]

More ramblings...

White balance:
These backs have different white points (not sure if that is the correct use of the term, but oh well). For the image above, CC20 (second gray from the left):
TC: White balanced to 6574/3.9 (OOC= 4845/-3)
3100: White balanced to 6803/17.7 (OOC=5049/-0.1)

I struggle with white balance in Capture One. It is a more iterative process than LR. To me, LR really is blue/yellow in temp and magenta/green in tint. In C1, moving temp affects the magenta/green cast, so there is more back and forth.

Point is, I liked the TC's OOC “white point” better than my 3100. A least with gray overcast skies. Getting to a “pleasing” white balance was faster on the TC. Obviously, for the images above it was just a click. But I prefer things not that warm because shadows need a little blue in them. For example, I would lower the temp in the images above by about 800. In the TC image below I posted in the other thread, it took just took a few clicks. I still do not have the 3100 version where I want it.

As for "accurate" color, I am very afraid to go down that rat hole. I do not have the skills or the equipment. The point of me going through this exercise has nothing to do with accurate color. It is purely to see if I "like" this back's color significantly better than the 3100 (well, that and bland curiosity). Viewed alone, any one of these renderings would look fine. I will say this: of the four options (TC and the 3100 with three different profiles), I like the TC best. There is a lot I do not understand, however:

• The orange looks far better in the TC image than any others. All the 3100 versions appear too saturated; the orange gets "crunchy." P1 Outdoor DL profile is second-best for the orange. This is curious because the difference between yellow and orange-yellow (colorchecker squares CC16 vs CC12) is too muted on the TC. They almost blend together. In real life CC16 and CC12 look much more like what is displayed with the Lumariver profile.
• On the other hand, I still think reds are too punchy in all but the Lumariver profile (which frankly is too dull). The apple looks "right" in the Lumariver profile.
• Others in previous threads both here and on lula focused on the greens. Sure enough, the limes and foilage go yellow on the 3100. From yellowish green to blueish green: 3100-Leaf, 3100-P1OutdoorDL, 3100-Lumariver, TC. This is all with a neutral CC20, 201/201/201.

I lowered the exposure to match the published luminance value of the colorchecker square CC20 (=200)
Found here:
X-Rite: +Colorimetric values for ColorChecker Family of Targets
Of course, none of the readouts for the color squares are anywhere close to what they are supposed to be!
:loco:

Here are some revised color readouts for the orange. TC top, 3100-Lumariver profile bottom:

Tomorrow I will try to do some shifting with the 35 and 40 for you all.

Ciao,
Dave

 

[dougpeterson]

I was not confused. Rather, I read your report on this back accurately. A clarification there might be useful! Thanks!

Please help me understand where I wrote that there wasn’t an IR filter in front of the sensor so I can add a clarification or edit.

 

[vjbelle]

Dave..... first of all thanks much for your taking the time to test these two backs for color accuracy. I have yet to own a Phase 1 back that does not favor a yellowish hue with the IQ180 being the worst and the 3100 being the best. I always corrected by just adjusting white balance a little towards a cooler hue. This new back seems to be another evolution in the never ending quest for better color balance.

Victor

 

[Paul2660]

Hi Dave

Thanks the testing.

Have you compared shadow push at base ISO 35 on the TC and 50 on the 3100? And higher ISOs both in good lightning and pushed?

I know that 150mp will not be on my list as I don't need it or want to work with the larger files and diffraction issues. But if the TC still interests me. I have only one other user report on the ISO ranges at 200 and beyond on the TC and it was not what I had hoped for but logically makes sense as the base ISO of the TC is as much as 2 stops below the 3100 if you assume the true base of the 3100 is 100 not 50.

Thanks
Paul Caldwell

 

[dchew]

Hi Dave

Thanks the testing.

Have you compared shadow push at base ISO 35 on the TC and 50 on the 3100? And higher ISOs both in good lightning and pushed?

Thanks
Paul Caldwell

Paul,
I haven't checked that. Tomorrow I will try to find some suitable subjects when I am out with John. I will post some raw files, but it might take a while to do that on my DSL.

Dave