Ad
Ad
Ad
Pages: « 1 [2] 3 »   Bottom of Page
Print
Author Topic: Using 'Linear Response' curve rather than 'Film Standard' in C1 ?  (Read 18581 times)
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #20 on: December 05, 2013, 10:20:00 AM »
ReplyReply

one more thing...

C1 also provides an "Auto" correction to any curve you like. Perhaps some could see if it's the stepping off point they desire. One could use Auto and then modify it to suit their general tastes and save that as their stepping off point as well

I find Auto amps exposure a bit more than I like (esily corrected though). However it is my understanding that Auto isn't a constant, but rather varies with the input... but I'm not sure on that
Logged
rhadorn
Newbie
*
Offline Offline

Posts: 5


« Reply #21 on: December 05, 2013, 02:51:28 PM »
ReplyReply

Thank you for your suggestions.

In a blog entry at C1, I read that a gamma curve is also applied to obtain the so called linear rendition. So this one is not linear in the sense of the original distribution of tone levels from the sensor. Unfortunately, that blog entry does not say more about it. I suppose that the curve they call linear is calculated so as to cancel the long shoulder and the small foot of the standard curve ('S' shape), intended at imitating the good ol' slide. Stretching the S reduces contrast of the mid-tones and this is what impacts the colors themselves, hue as well as saturation.

As to the support the automatic adjustment may give, I made a simple test, applying the auto adjustment for the standard and the linear rendition to the same picture. Contrast and saturation are corrected to the same extent, only the exposure varies, actually pushing to the highlights, as you observed yourself. Sorry, the highlights slider slips also a bit more right with the standard curve. I purpously slected a picture with overexposed sunlit clouds.

Reto
Logged
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #22 on: December 05, 2013, 03:20:12 PM »
ReplyReply

Not sure where David is but I know he can help

IIRC, he said it's "linear" because it is the RAW sensor's output. All others are actually "curves" which is why they look as we are used to seeing them. They even went into discussing that some houses now limit to starting with "linear" since it is pure

In any case, the LULA course is great. There is so much to the software. The program is so rich and modifiable vs LR. Even Michael seemed impressed at some points. I'm sure he can answer our question as well...

Thank you for your suggestions.

In a blog entry at C1, I read that a gamma curve is also applied to obtain the so called linear rendition. So this one is not linear in the sense of the original distribution of tone levels from the sensor. Unfortunately, that blog entry does not say more about it. I suppose that the curve they call linear is calculated so as to cancel the long shoulder and the small foot of the standard curve ('S' shape), intended at imitating the good ol' slide. Stretching the S reduces contrast of the mid-tones and this is what impacts the colors themselves, hue as well as saturation.

As to the support the automatic adjustment may give, I made a simple test, applying the auto adjustment for the standard and the linear rendition to the same picture. Contrast and saturation are corrected to the same extent, only the exposure varies, actually pushing to the highlights, as you observed yourself. Sorry, the highlights slider slips also a bit more right with the standard curve. I purpously slected a picture with overexposed sunlit clouds.

Reto
Logged
narikin
Sr. Member
****
Offline Offline

Posts: 858


« Reply #23 on: December 08, 2013, 08:52:17 AM »
ReplyReply

Am currently taking the LULA C1 Course

In about lesson 10 or 11 or so, Michael and David explain the "linear" is the actual RAW file curve. The Others, "film standard" are curved renditions of the RAW

How's the LuLa course?  I know and have used C1Pro for something like 8 years, but there's always something new to learn.

For example today I recently learned that the tilt sensor information (the built in level you can use on the digital back) is recorded for every shot, and... you can tell C1 to access that and correct the alignment based on it, automatically.

Yes linear response is interesting and probably the real raw file, not an interpretation of it ('Film standard'), but like many, I am put off by the clear drop in saturation/color in the vital mid tones of every image, and all the hard work to restore that.  It is however a great option for sky/cloud/high tone images.
Logged
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #24 on: December 08, 2013, 10:18:42 AM »
ReplyReply

How's the LuLa course?  I know and have used C1Pro for something like 8 years, but there's always something new to learn.

For example today I recently learned that the tilt sensor information (the built in level you can use on the digital back) is recorded for every shot, and... you can tell C1 to access that and correct the alignment based on it, automatically.

Yes linear response is interesting and probably the real raw file, not an interpretation of it ('Film standard'), but like many, I am put off by the clear drop in saturation/color in the vital mid tones of every image, and all the hard work to restore that.  It is however a great option for sky/cloud/high tone images.

as above, the course seems almost mandatory for such a rich and open-ended bit of software.

as far as a "drop in saturation and color" one may also consider the "curves" as a cheat of sorts and the Linear as an advantage since the curves no doubt are arbitrary in some senses and the RAW allows you to decide. Sort of like cooking from scratch vs. boxed pre-recipes.
Logged
rhadorn
Newbie
*
Offline Offline

Posts: 5


« Reply #25 on: December 09, 2013, 02:34:35 AM »
ReplyReply


I am glad the discussion continues, perhaps we will come to good ideas about the color compensation necessary when using the so called linear curve. At this point I would like to make my point again: the linear option in C1 does NOT deliver the raw data from the sensor.

As a first demonstration, we can read again a very good article by Bruce Fraser: http://www.adobe.com/digitalimag/pdfs/linear_gamma.pdf.
- You will see that the image corresponding to the raw data is ways darker than the image delivered by the linear curve in C1.
- If you apply to a 'linearized' image in C1 the gamma curve shown on p. 2 in the cited paper, you get a much too bright image.

Second, we can also dive into C1's blog. I recommend the reading of following article: http://blog.phaseone.com/take-full-control-of-the-tone-mapping-by-using-a-linear-film-curve/. Let me pick a few citations:

"By default the curve “Film Standard” is used. This curve is designed to give a similar tone mapping as when using transparency film. The curve has a slight S shape in order to achieve deep blacks while still maintaining details in the highlight, giving a gentle transition into overexposed areas." This shows that the Film Standard curve is not just arbitrary but... reproduces the limitations of transparency film as well as its brillance. The author adds: "The default ICC Color Profile is designed to work well together with the default curve “Film Standard”". This is the reason why I would greet the definition of a second ICC profile, which would enhance the results obtained with the linear curve.

In the fourth illustration, we see the curve which must be applied to the linear rendering to obtain the standard rendering. It is an S-shaped curve, very distinct from the gamma curve applied to true sensor data in Fraser's article.

Later on, in the discussion, Niels answers a question: "Yes, you are absolutely right. Gamma correction has been applied. "

So, the sequence seems to be the following:

RAW    gamma   LINEAR    S-shape    FILM STANDARD

In C1, linear does not refer to the linearity of the sensor data but to perceptualy equally spaced tone levels of the gamma corrected data, which nowadays cover 12-14 Ev's.

I interpret the Film Standard curve as the expression of the transition between analogue and digital photography, a time in which a good picture is still, for many people, a picture which could have been made with analogue means. In my eyes, the linear curve is as legitimate as a standard as the Film Standard in C1 and should be paired with a specific ICC profile.


Reto
Logged
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #26 on: December 09, 2013, 09:14:32 AM »
ReplyReply

I'm certain that you know far more about image processing than I do. And can see what you're inferring

Here is the script from LULA's C1 course at about 3 min into Chapter 10. Basic Editing

Michael; (reference to "Linear") - "This is actually what the sensor is producing"

David; "yes"

Bruce; "Yes"

Michael; "a sensor produces a linear image...."

If I give away any more they may send a hit man

they do not address ICC profiling

cheers

h
Logged
narikin
Sr. Member
****
Offline Offline

Posts: 858


« Reply #27 on: December 09, 2013, 10:57:15 AM »
ReplyReply

I am glad the discussion continues, perhaps we will come to good ideas about the color compensation necessary when using the so called linear curve.
I interpret the Film Standard curve as the expression of the transition between analogue and digital photography, a time in which a good picture is still, for many people, a picture which could have been made with analogue means. In my eyes, the linear curve is as legitimate as a standard as the Film Standard in C1 and should be paired with a specific ICC profile.

Reto

Agree.

It seems there should be a profile for linear response for the major IQ series backs, not leaving us using one made by another route.  Have you suggested it to Phase One, or 'The Image Professor' ?  Though maybe someone very smart can explain why it might be trivial or pointless?


Logged
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1320


« Reply #28 on: December 09, 2013, 12:20:54 PM »
ReplyReply

Have you suggested it to Phase One, or 'The Image Professor' ?  Though maybe someone very smart can explain why it might be trivial or pointless?

there are tehnical (software,no hardware) engineers from P1 present here (not very active though), for example this individual  = http://www.luminous-landscape.com/forum/index.php?action=profile;u=62645   or this individual = http://www.luminous-landscape.com/forum/index.php?action=profile;u=55954
« Last Edit: December 09, 2013, 12:25:07 PM by Vladimirovich » Logged
tho_mas
Sr. Member
****
Offline Offline

Posts: 1696


« Reply #29 on: December 09, 2013, 12:59:27 PM »
ReplyReply

in C1 you can't view an image without a tone curve nor withaout an input profile ("camera profile") applied.
All camera profiles in C1 are Gamma 1.8.
The "linear" curve produces an image with equidistant tonal values when viewed with a Gamma 1.8 profile assigned.
The term "linear curve" has nothing to do with the linearity of the sensor (or raw file).
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7424


WWW
« Reply #30 on: December 09, 2013, 01:33:44 PM »
ReplyReply

Hi,

That video lacks Jeff Schewe…

Best regards
Erik

there are tehnical (software,no hardware) engineers from P1 present here (not very active though), for example this individual  = http://www.luminous-landscape.com/forum/index.php?action=profile;u=62645   or this individual = http://www.luminous-landscape.com/forum/index.php?action=profile;u=55954
Logged

rhadorn
Newbie
*
Offline Offline

Posts: 5


« Reply #31 on: December 10, 2013, 12:32:10 PM »
ReplyReply


@ Narikin

Yes, I posted the question into the comments of the blog entry handling the linear curve on December 4th. I did not get any answer for now. I suppose Niels is on some trip in a warm country.

@ tho_mas

Interesting to read that C1 applies a gamma of 1.8 to the sensor data. Can you tell what is your source?

My system is calibrated for a gamma of 2.2, which is actually very common. This raises new questions:
- would the colors from the linear curve look more vivid on my system if the gamma applied by C1 was 2.2?
- is the pretty steep S-curve for the Standard rendering also a way to fill the gap between 1.8 and 2.2?
These questions may be naive, I am grateful for any expert advice.

Reto
Logged
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #32 on: December 10, 2013, 01:48:42 PM »
ReplyReply

in C1 you can't view an image without a tone curve nor withaout an input profile ("camera profile") applied. This is not correct if you mean that I can't just import without designating some camera. Otherwise not sure what you mean
All camera profiles in C1 are Gamma 1.8. source?
The "linear" curve produces an image with equidistant tonal values when viewed with a Gamma 1.8 profile assigned  Again, not what the tutorial says, can you provide a source for this statement?
The term "linear curve" has nothing to do with the linearity of the sensor (or raw file). So you say...

[/quote

I don't know much... but I do know what I am told, and what I can manifestly demonstrate. It gets silly sometimes when folks seem to just want to say "you're wrong". Either you have a source for what you say or you do not. I may not be correct, it's happened before far too often, but at least I have a reasonable source that must also be wrong too. Could you provide yours?
« Last Edit: December 10, 2013, 01:51:12 PM by allegretto » Logged
tho_mas
Sr. Member
****
Offline Offline

Posts: 1696


« Reply #33 on: December 10, 2013, 07:26:06 PM »
ReplyReply

@ tho_mas

Interesting to read that C1 applies a gamma of 1.8 to the sensor data. Can you tell what is your source?
you can inspect the input profiles with dedicated software. As a practical test you can also create a new blank image in Photoshop and assign any Capture One camera profile (for instance the "no color correction" or any other input profile of your choice) to it. Now create a black to white gradient and afterwards assign for instance ProPhotoRGB (which is Gamma 1.8 ) to the file. You'll see that the distribution of tonal values does not change after assigning ProPhoto... which shows that the original file is also Gamma 1.8.

I assume herewith I also address "allegretto's" questions...
« Last Edit: December 10, 2013, 07:28:29 PM by tho_mas » Logged
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #34 on: December 10, 2013, 08:16:20 PM »
ReplyReply

so, one should infer that the source is you?

If so, maybe you're correct and Michael and David are not. Not choosing sides, simply asking for a source. Which you have provided, correct?

Thank you. That was easy.
Logged
SecondFocus
Sr. Member
****
Offline Offline

Posts: 467


WWW
« Reply #35 on: December 10, 2013, 10:06:06 PM »
ReplyReply

Interesting discussion, much of it way technically past me. However I was just having a difficult time with one series from a shoot with someone wearing a red shirt. In this series the red reproduced in C1, Adobe and Canon DPP was not very close to the actual color by a long way. These were shot on Canon. I changed to Linear in C1 and was able to very closely match the exact color with a few slight adjustments. (White balance was not the issue). I don't have any idea as to the how and why behind it, but it is a new tool in my toolbox!

Maybe Capture One could do a new webinar on more unusual features like this. As well as addressing such problems we face.
Logged

Ian L. Sitren
SecondFocus
allegretto
Sr. Member
****
Offline Offline

Posts: 447


« Reply #36 on: December 10, 2013, 10:47:27 PM »
ReplyReply

Interesting discussion, much of it way technically past me. However I was just having a difficult time with one series from a shoot with someone wearing a red shirt. In this series the red reproduced in C1, Adobe and Canon DPP was not very close to the actual color by a long way. These were shot on Canon. I changed to Linear in C1 and was able to very closely match the exact color with a few slight adjustments. (White balance was not the issue). I don't have any idea as to the how and why behind it, but it is a new tool in my toolbox!

Maybe Capture One could do a new webinar on more unusual features like this. As well as addressing such problems we face.

There is to some extent! A webinar that tells you perhaps not everything, but gives you a comprehensive overview and makes suggestions about how features might be used (as you have noticed). It's called the LULA course. It isn't free, $60. but if you're going to try to really use the program it's a bargain.

Now as noted above, maybe they are not correct about the fine technical stuff (sooner or later David or Michael will probably help us here) but they sure have great ideas on rescue. Raber goes into a lot of this kind of stuff and how you can fine tune specific tonal metrics.

Good stuff...
Logged
tho_mas
Sr. Member
****
Offline Offline

Posts: 1696


« Reply #37 on: December 11, 2013, 04:47:11 AM »
ReplyReply

so, one should infer that the source is you?
the source are the actual input profiles available in Capture One (all of them).
Logged
Doug Peterson
Sr. Member
****
Offline Offline

Posts: 2803


WWW
« Reply #38 on: December 11, 2013, 05:04:15 PM »
ReplyReply

Maybe Capture One could do a new webinar on more unusual features like this. As well as addressing such problems we face.

We address a variety of such niche tools and workflows in our Capture One Masters Program.
« Last Edit: December 12, 2013, 09:01:13 AM by Doug Peterson » Logged

DOUG PETERSON (dep@digitaltransitions.com), Digital Transitions
Dealer for Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Profoto
Office: 877.367.8537
Cell: 740.707.2183
Phase One IQ250 FAQ
rhadorn
Newbie
*
Offline Offline

Posts: 5


« Reply #39 on: December 12, 2013, 07:53:54 AM »
ReplyReply

We address a variety of such niche tools and workflows in our Capture One Masters Program.


Hi Doug,

Nice to read somebody having a deep knowledge of Capture One here. And thank you for the hint, that we may find an answer to our questions in your tutorials.

Perhaps you would be more successfull in selling them to the folks here if you gave us a demonstration of your capacity to answer the question asked about the icc profile, which would serve le linear curve as well as the default icc curve serves the standard curve. I asked that same question in Niel's blog and am still waiting for an answer, so I start considering that there is perhaps no answer at all on the side of Phase One.

By the way, the e-mail address on the page you refer to is false, it probable should be dep@digitaltransitions.com and not dep@digitransitions.com.

...

Welcome


Reto
Logged
Pages: « 1 [2] 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad