Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Merrils and A7r color differences  (Read 1466 times)
mdijb
Sr. Member
****
Offline Offline

Posts: 392


WWW
« on: April 08, 2014, 06:55:16 PM »
ReplyReply

I recently did some shooting with both my Merills and my new A7r and noticed some significant color differences in the images from each camera

Attached are the real scene from each camera and Photos of a color checker I took today. All were taken with AWB and I adjusted only the brightness of the color cecker shots to bring them close together for comparison.

Nothing comes close to resolution and detail form the Merrils,  Not even my A7R.  With processing and sharpening, the A7r images however are very good and sharp.

However the color differences are something I never  noticed before and are concerning.  It takes some severe adjustments in LR to get the Merril images to get sort of close to the results from the A7r, and frankly most times I prefer the colors form the A7r because they were real close to the real scene and seemed much more pleasing.

I would appreciate your thoughts and comments

MDIJB
« Last Edit: April 08, 2014, 07:03:33 PM by mdijb » Logged

mdiimaging.com
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1320


« Reply #1 on: April 08, 2014, 10:35:03 PM »
ReplyReply

I would appreciate your thoughts and comments

and why do you expect the colors to be the same ?

Logged
mdijb
Sr. Member
****
Offline Offline

Posts: 392


WWW
« Reply #2 on: April 09, 2014, 12:04:44 AM »
ReplyReply

I do not expect them to be the same.  My concern is that the colors in the Merill images are not even close to the real scene, while the A7r images are much more like the actual colors I observed in reality. 

Is there a setting in the Merrils I am missing?

Even though the merrils produce outrageous sharpness, the colors are not what I expected and are not close to real thing.

However, I have shot a series of flowers in market with available light and the colors of these were very good and much more accurate that I am seeing now with the landscapes.

MDIJB
Logged

mdiimaging.com
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #3 on: April 09, 2014, 12:25:46 AM »
ReplyReply

I do not expect them to be the same.  My concern is that the colors in the Merill images are not even close to the real scene, while the A7r images are much more like the actual colors I observed in reality.  

Is there a setting in the Merrils I am missing?

Even though the merrils produce outrageous sharpness, the colors are not what I expected and are not close to real thing.

However, I have shot a series of flowers in market with available light and the colors of these were very good and much more accurate that I am seeing now with the landscapes.

MDIJB
Without trying to start a flame-war here... If you are after a "colorimetrically accurate" camera (i.e. one that mimics our understanding of human color vision), then I think that Foveon is a strange choice, while Sony system cameras (in general) is a good one? It seems that Foveon color images can only be made sort of believable after some pretty hefty nonlinear processing, while other cameras raw files need only a modest 3x3 linear correction.

Since you have the xrite passport: did you try to make a profile of both shots, using that profile to correct the image under the same illumination?

-h
Logged
SZRitter
Full Member
***
Offline Offline

Posts: 156


« Reply #4 on: April 09, 2014, 09:21:52 AM »
ReplyReply

Honest question, can you even create a profile for a Merril or process it in Lightroom/ACR?

That aside, my understanding of sensors is that they will all vary in response, much like different film types. So it is your job as the photographer to bring it back into an acceptable standard, if that is your priority. And since the Foveon handles light very different than a standard Bayer array, I am not surprised one bit that it's color response is so different.
Logged
Herbc
Full Member
***
Offline Offline

Posts: 110


« Reply #5 on: April 09, 2014, 09:28:29 AM »
ReplyReply

Ok, since I don't pay a lot of attention to color, being in the 5% that are red/green colorblind (only for nuanced colors, I am just fine for traffic lights),  is it the case that the Sigma has its own "look", and the Sony doesn't??

Seems like changing cameras would always show some differences, although I have seen raves about the Sigma colors,  I notice colors are different with different lenses, at least they seem that way to me.
Logged
SZRitter
Full Member
***
Offline Offline

Posts: 156


« Reply #6 on: April 09, 2014, 10:29:23 AM »
ReplyReply

Ok, since I don't pay a lot of attention to color, being in the 5% that are red/green colorblind (only for nuanced colors, I am just fine for traffic lights),  is it the case that the Sigma has its own "look", and the Sony doesn't??

Seems like changing cameras would always show some differences, although I have seen raves about the Sigma colors,  I notice colors are different with different lenses, at least they seem that way to me.

*Waves* Hello from another of the 5%. Mine seems to be getting worse every year.

Due to how the Foveon sensor works, it responds to light completely different. The basic is that a bayer array has filters that allow each pixel to get red, green, or blue (I forget if white is also in the array) then uses it's neighbors to estimate what a pixel should be in full RGB response. A Foveon, by my understanding uses three layers of pixels stacked on top of each other, one blue, one red, and one green. This means each "pixel" is the composite of the stack, leading to hypothetically more accurate response to the collected light.

In a lab, this may be very true, but as a partially color blind person, you can agree that not all eyes respond the same, let alone how our brains view a scene.

As to the OP, have you tried to isolate what kind of conditions give you these "off" colors? Is it a certain color? Certain type of light?
Logged
uvl
Newbie
*
Offline Offline

Posts: 33



WWW
« Reply #7 on: April 09, 2014, 12:11:06 PM »
ReplyReply

To get somehow realistic colors with the Merrill you need to do a custom white balance before you shoot. Developing with Neutral color mode will give you well balanced result without exaggerating any color. Forget AWB. It's a guess. And you need daylight or flashlight.
Logged
Telecaster
Sr. Member
****
Offline Offline

Posts: 866



« Reply #8 on: April 09, 2014, 02:52:58 PM »
ReplyReply

...but as a partially color blind person, you can agree that not all eyes respond the same, let alone how our brains view a scene.

Yep. Somewhere around 13% of women are sensitive to four peak wavelengths rather than the standard three. Maybe the rest of us are all, in comparison, color blind.   Smiley

-Dave-
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #9 on: April 10, 2014, 02:56:14 AM »
ReplyReply

Honest question, can you even create a profile for a Merril or process it in Lightroom/ACR?
I don't know

Quote
That aside, my understanding of sensors is that they will all vary in response, much like different film types. So it is your job as the photographer to bring it back into an acceptable standard, if that is your priority. And since the Foveon handles light very different than a standard Bayer array, I am not surprised one bit that it's color response is so different.
My understanding is that Foveon is inherently less like "acceptable color" than other technologies with regards to color accuracy. Thus, more processing is needed in order to make it look acceptable.

-h
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #10 on: April 10, 2014, 02:59:13 AM »
ReplyReply

Due to how the Foveon sensor works, it responds to light completely different. The basic is that a bayer array has filters that allow each pixel to get red, green, or blue (I forget if white is also in the array) then uses it's neighbors to estimate what a pixel should be in full RGB response. A Foveon, by my understanding uses three layers of pixels stacked on top of each other, one blue, one red, and one green. This means each "pixel" is the composite of the stack, leading to hypothetically more accurate response to the collected light.
Bayer does not include white.

The spatial sampling has nothing to do with color tone accuracy. The problem with Foveon is that those stacked pixels are not "red", "green" and "blue", but some other desaturated colors. In order to make the three channels into something resembling accurate colors, you need to do some fancy processing that resemble the operations needed to make a blurry lense appear sharp (sharpening).

-h
Logged
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1320


« Reply #11 on: April 10, 2014, 12:31:02 PM »
ReplyReply

Bayer does not include white.

some implementations actually does (I think in some cell phones for example)
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #12 on: April 10, 2014, 12:36:19 PM »
ReplyReply

some implementations actually does (I think in some cell phones for example)
The usual layout for a Bayer sensor is to have 50% "green" sensels and 25% each of "red" and "blue". While there are many other possible color filter arrangements, I am not sure if it makes sense to call them "Bayer"?

-h
Logged
mdijb
Sr. Member
****
Offline Offline

Posts: 392


WWW
« Reply #13 on: April 11, 2014, 08:12:53 AM »
ReplyReply

To get somehow realistic colors with the Merrill you need to do a custom white balance before you shoot. Developing with Neutral color mode will give you well balanced result without exaggerating any color. Forget AWB. It's a guess. And you need daylight or flashlight.


Thanks for this idea--I will give it a try

Also thanks for the suggestion about a creating a camera profile--I will try this as well

MDIJB
Logged

mdiimaging.com
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad