Improving the Camera: The Image

[Hrvoje Simic]

1. This is a subjective 3dLUT as it is not a transform to a known target like REC709. As such it's usefulness is limited by how well your subjective corrections play out on different shots and by how destructive the LUT is.
2. Having looked at the images posted here and on your site and having read through the thread, it becomes clear that what needs to be explained is that there is in fact no way to Improve the image outside of the SDK. Change perhaps, but your LUT will only manipulate what is already there and is destructive, meaning there is a penalty for every change. As such it is not a "Primer"but a Artisitc LUT similar to a film emulation or any other LUT that offers a subjectively more pleasing image.

Dear Aharon,

By your rhetoric it is evident that you didn't quite get what was written here and what is this all about so you also compensate with projections of concepts you are familiar with, for some reason insisting that this is actually what you assume it is, despite the fact that the very source this actually originates from says completely the opposite.

That is okay though, it takes a while and it gets better with time. If there is a will and capacity to understand.
Until that happens, it would be productive and not trolling if you saved your guesswork and what you think or *prefer to think* this is for yourself.

Btw, SDK tells you nothing about camera + signal mapping-specific numerous signal deviations.
It serves a different purpose.

It would be great if things were that simple.

 

[Hrvoje Simic]

Anyone wants to beta test Alexa Primers ?

 

[Aharon Rothschild]

Hvroje I understand that these LUTs are simliar to these:
https://www.hurlbutvisuals.com/blog/product/red-dragon-luts-neutral-and-skin-tone-bundle/?tm=ic

and may be useful to the extent that your subjective color decisons play out on different shots. I understand that all LUTs are destructive and there is no way to add anything to a debayered signal that isn't already there.

Please edit your first post to reflect the following:

"Dragon response is optimized for monitoring and post.
Significant improvements across the signal range."

TO:

The signal is neither optimized or impoved. The LUT is a post debayer transform for a subjectively pleasing image.

Thanks

 

[Mike Nagel]

Btw, SDK tells you nothing about camera + signal mapping-specific numerous signal deviations.

of course it does not "tell you signal deviations". the SDK cannot know what a deviation is, b/c for that it would have to have a reference to compare to in order to assess a deviation.... plus a known target... plus other compensation values...

a SDK (Software Developer Kit), is a set of developer tools. One needs to be a developer. And understand basic principles and concepts.

;-)

 

[Mike Nagel]

And how do You compensate for hardware errors? With software? Just by remaping c.v. around?
Aren't You inducing banding by doing so?
My point is if it's acceptable with monitors why wouldn't be acceptable with cameras?

ultimately, the calibration sw relies on the probe, ergo hardware. there is no way around it. I use only high-end equipment, but in order to know how accurate it is - for the spectros (!), for the colorimeter u can run tests on ur own - u need to evaluate against an even more accurate reference. which is what the yearly manufacturer maintenance is for (to assess drift and re-calibrate the probe).

Re charts: I'm not disagreeing. I shoot charts on every single shot. My point was that the analogy of shooting charts on set with display calibration is off, because on set you will deal with non-uniform illumination of the reference, ergo ur chart + plus possible mixed lighting. So compensation/optimization can possibly be very tricky or impossible.

In display calibration, the probe takes measurement samples from a single, small spot and returns a set of direct XYZ values. Very different.

 

[Hrvoje Simic]

of course it does not "tell you signal deviations". the SDK cannot know what a deviation is, b/c for that it would have to have a reference to compare to in order to assess a deviation.... plus a known target... plus other compensation values...

Excellent Mike, we're slowly getting there.
I have faith in you.

I do that other thing.
And I already wrote that.

Hvroje I understand that these LUTs are simliar to these:

You didn't understand anything and just keep repeating the same as if it is going to change reality and my posts were written in Swahili.

The tool wrapper is misleading you.
The analytical property of your mind relies on known references, grabs on to the one thing you know, which is the wrapper and attempts to rationalize it by putting it into the category you are familiar with.

Due to the nature of the project, much of which is pioneering, I gave three factors in the start of the thread necessary for successful assimilation. You are failing in all three so far which creates an issue in this discussion. Some didn't.

 

[Sabovic Adis]

ultimately, the calibration sw relies on the probe, ergo hardware. there is no way around it. I use only high-end equipment, but in order to know how accurate it is - for the spectros (!), for the colorimeter u can run tests on ur own - u need to evaluate against an even more accurate reference. which is what the yearly manufacturer maintenance is for (to assess drift and re-calibrate the probe).

Re charts: I'm not disagreeing. I shoot charts on every single shot. My point was that the analogy of shooting charts on set with display calibration is off, because on set you will deal with non-uniform illumination of the reference, ergo ur chart + plus possible mixed lighting. So compensation/optimization can possibly be very tricky or impossible.

In display calibration, the probe takes measurement samples from a single, small spot and returns a set of direct XYZ values. Very different.

And my point is that You don't even have to compensate for uneven illumination once You've calibrated (or in this case primed) the whole thing. Uneven illumination and mixed lighting is part of the look created by DOP. It's "ambiance". It's "atmosphere"." Play of shadows"...
That's DOP's "doing". :rofl:
Just imagine a shot under red light (cuz the damn script says so) - not gonna whitebalance that, aren't Ya? Or "dimly lit", or "foggy". How do You corect for that - You don't. :biggrin:

But "artistically" You do whatever feels right to do, right? :thumbsup:

 

[Aharon Rothschild]

Excellent Mike, we're slowly getting there.
I have faith in you.

I do that other thing.
And I already wrote that.




You didn't understand anything and just keep repeating the same as if it is going to change reality and my posts were written in Swahili.

The tool wrapper is misleading you.
The analytical property of your mind relies on known references, grabs on to the one thing you know, which is the wrapper and attempts to rationalize it by putting it into the category you are familiar with.

Due to the nature of the project, much of which is pioneering, I gave three factors in the start of the thread necessary for successful assimilation. You are failing in all three so far which creates an issue in this discussion. Some didn't.

Oy vey.
Using color function association to affect the molecular assimilation of light waves and through a long and labourous process of study, Hvroje has arrived at his version of visual alchemy which attempts the impossible: transmutation of one property to another, like gold to lead or RWG or RedLog to some optimised version of REC709 containing better more optimised color than before without paying any of the penalty that destructive color manipulation entails. To understand this mystery we must only grasp the three principles of understanding.

 

[Mike Nagel]

Excellent Mike, we're slowly getting there.
I have faith in you.

;-)))))))))))))))))))

I admit, obviously by the continued participation in this thread, you do amuse me. Quite a bit. You are - inadvertently - wildly entertaining to me.

You've been slapped on the finger multiple times. You post "alternative facts". Then, when (VERY FRIENDLY) pointing it out, you argue like a 2 year old child, getting ultra defensive and at the same time insult everybody else here. Yes, the irony. Then you flip/flop. Then you contradict yourself. It is clear that you don't understand certain principles and concepts. I personally do not care (about ur irrational reasoning) because I have no skin in this game and like I said this solution can be tried, like many other ones that already exist. Or, you just learn how to grade. ;-)

But it is a mystery why you continue to try playing baby mind games with these twisted responses... ? ;-)))))))))) Are u incapable or learning from ur mistakes ?

Clearly, you have no clue about the SDK. Or probably any SDK, for that matter, given ur naive statements. And I'm not saying that the RED SDK holds the holy grail for an Arri debayer.

95% of any of ur posts is hot air. It means absolutely nothing. Smoke and mirrors garbage packed in meaningless metaphors and analogies. Diversion and distraction alongside your inner desire to try to portray yourself as intelligent and put urself on a high horse, or rather: a large donkey that u ride towards the windmills that ur delusional mind creates. On a tech forum. Selling a tech solution. Not bad ;-)))))))

But buddy, after u've been adjusted a few times, I have to ask: who do u think u can fool with this here on this forum ?

It would appear that.... THE DOMAIN ABOVE THOUGHT.... did a serious number on u, old chap.

Re the girl shot. I already saw the results. But ur response showed how twisted u are. Real footage can and will be 1000x worse, and u're already crying like a little girl about that shot.

Nobody here wants u any bad, this was supposed to be an open conversation (workflow, approach, usage, caveats, etc) with a developer - that unfortuantely turned out to be delusional and incapable of answering normal questions (ergo: product support). U seem to be very new to any of this.

I wish u the best of luck. Only that can help u.

;-)

 

[Mike Nagel]

And my point is that You don't even have to compensate for uneven illumination once You've calibrated (or in this case primed) the whole thing. Uneven illumination and mixed lighting is part of the look created by DOP. It's "ambiance". It's "atmosphere"." Play of shadows"...
That's DOP's "doing". :rofl:
Just imagine a shot under red light (cuz the damn script says so) - not gonna whitebalance that, aren't Ya? Or "dimly lit", or "foggy". How do You corect for that - You don't. :biggrin:

But "artistically" You do whatever feels right to do, right? :thumbsup:

You can correct it, and I do. So that u actually get red light and not RED's light, which is yellow/green contaminated ;-))))))

My response was towards being aware when using charts in mixed + non-uniform lighting. If u would know how the correction algorithms work, u would understand what I'm saying. Shooting charts is very important, u just have to understand how to evaluate and approach each shot and the chart within the shot when using it for color and 1D adjustments. That is my point. Display calibration does not work like that. It is always the same after the initial screen pre-cal setup - all patches are treated equally by the probe.

 

[Sabovic Adis]

You can correct it, and I do. So that u actually get red light and not RED's light, which is yellow/green contaminated ;-))))))

Isn't that the whole point of "priming"? :biggrin:

My response was towards being aware when using charts in mixed + non-uniform lighting. If u would know how the correction algorithms work, u would understand what I'm saying. Shooting charts is very important, u just have to understand how to evaluate and approach each shot and the chart within the shot when using it for color and 1D adjustments. That is my point.

Why would You even *touch gamma or any of them color wheels if You had a perfectly working camera?

* Artistic coloring excluded, of course. :biggrin:

Display calibration does not work like that. It is always the same after the initial screen pre-cal setup - all patches are treated equally by the probe.

What if somebody lights a cigarette during color grading - does that count as "mixing the lights" & "unevening the illumination"? :rofl:

:thumbsup:

 

[Mike Nagel]

Why would You even *touch gamma or any of them color wheels if You had a perfectly working camera?

Ah, drugs !

What if somebody lights a cigarette during color grading - does that count as "mixing the lights" & "unevening the illumination"?

Absolutely not, not in ur house. Smoking and flashlights are mandatory during the grade. Turn on the disco lights and just turn the screen off, what do u need a screen for ? Use it as an ashtray ! Adjust colors by "feeling the controls". It can only improve ur current results. :biggrin:

 

[Brett Harrison]

Post it up on Lift Gamma Gain.

The hues make me feel a bit queasy at times, but I do like the contrast curve for some of the example images.

I think there's too much saturation present for something that is considered a calibration pass on the image. A good example of a calibration of the image is Light Iron's dxl.panavision.com - the change is subtle but profound (scroll down a bit).

 

[Eric Z]

Hrvoje,

In addition to my previous thoughts, looking at the latest photos posted here, and the before/after on omeneo.com, it now seems to me that part of what you are trying to achieve with Primers is somewhat like the effect of a Polarizer filter - without actually using one.
So besides fixing the nasty yellow-greenish tint of skin tones, among other things, the Primers also seem to try and preserve colors in highlights, like a CPL would do...

 

[Hrvoje Simic]

;-)))))))))))))))))))

I admit, obviously by the continued participation in this thread, you do amuse me. Quite a bit. You are - inadvertently - wildly entertaining to me.

You've been slapped on the finger multiple times. You post "alternative facts". Then, when (VERY FRIENDLY) pointing it out, you argue like a 2 year old child, getting ultra defensive and at the same time insult everybody else here. Yes, the irony. Then you flip/flop. Then you contradict yourself. It is clear that you don't understand certain principles and concepts. I personally do not care (about ur irrational reasoning) because I have no skin in this game and like I said this solution can be tried, like many other ones that already exist. Or, you just learn how to grade. ;-)

But it is a mystery why you continue to try playing baby mind games with these twisted responses... ? ;-)))))))))) Are u incapable or learning from ur mistakes ?

Clearly, you have no clue about the SDK. Or probably any SDK, for that matter, given ur naive statements. And I'm not saying that the RED SDK holds the holy grail for an Arri debayer.

95% of any of ur posts is hot air. It means absolutely nothing. Smoke and mirrors garbage packed in meaningless metaphors and analogies. Diversion and distraction alongside your inner desire to try to portray yourself as intelligent and put urself on a high horse, or rather: a large donkey that u ride towards the windmills that ur delusional mind creates. On a tech forum. Selling a tech solution. Not bad ;-)))))))

But buddy, after u've been adjusted a few times, I have to ask: who do u think u can fool with this here on this forum ?

It would appear that.... THE DOMAIN ABOVE THOUGHT.... did a serious number on u, old chap.

Re the girl shot. I already saw the results. But ur response showed how twisted u are. Real footage can and will be 1000x worse, and u're already crying like a little girl about that shot.

Nobody here wants u any bad, this was supposed to be an open conversation (workflow, approach, usage, caveats, etc) with a developer - that unfortuantely turned out to be delusional and incapable of answering normal questions (ergo: product support). U seem to be very new to any of this.

I wish u the best of luck. Only that can help u.

;-)

Can't hear you, I have bananas in my ears.

 

[Hrvoje Simic]

Hrvoje,

In addition to my previous thoughts, looking at the latest photos posted here, and the before/after on omeneo.com, it now seems to me that part of what you are trying to achieve with Primers is somewhat like the effect of a Polarizer filter - without actually using one.
So besides fixing the nasty yellow-greenish tint of skin tones, among other things, the Primers also seem to try and preserve colors in highlights, like a CPL would do...

All the effects you see are outcomes when the image is re-mapped from base to a viewable space in a way which:

a) corrects the fixed part of measurable camera + signal mapping originated signal deviations
b) negates certain effects of digital image quantization and processing

...to a possible extent which doesn't cause visible deteriorations of the signal and allows further post work.

That nasty yellow tint on skin you see is one of many signal deviations which happen when digital signal reaches extremes and are very hard to avoid or correct. Primers among other things correct those and "iron" the signal so it keeps color consistency across its dynamic range.

Signal can be manipulated and shaped at base level, transformation level and in grading.
Each level allows specific approach, advantages, manipulation range and limitations.

What you see with skies is the effect of digital's boosting of image parameters due to properties of mathematical functions and properties of color space they manifest in.

Those skies weren't originally as on the first image > so the process would somehow remove that from the image and gathered light, they were as on the second but you didn't see it the first because of what digital typically does.
That can be modified from base gamma and gamut image to get back to the natural image, to extent dependent on properties of the signal. With some cameras less, with some more. With Red - more. The tools accomplishing this have to use the body of a 3DLUT with 64^3 precision, to have precise independent access to signal particulars and remap the signal properly with numerous micro-corrections in 3 axis.

You can grade all day, but if you have a blob of yellow you''ll push it around into a blob of something else. Or do a Sisiphus job with mask, qualify or even paint, attempting to correct this, while inheriting more signal distorsions by the properties of modification tools and color space they operate in, and what ever you do will never be as if the tonality and separations were correct from the start.

This is what Primers are for.

 

[Hrvoje Simic]

Post it up on Lift Gamma Gain.

The hues make me feel a bit queasy at times, but I do like the contrast curve for some of the example images.

I think there's too much saturation present for something that is considered a calibration pass on the image. A good example of a calibration of the image is Light Iron's dxl.panavision.com - the change is subtle but profound (scroll down a bit).

You may interpret strange looking saturation with Primer which doesn't have toe and knee fully defined. That is normal because contrast is still unfinished while saturation mapping is done and it all sinks in together as soon as you fix your own toe and knee or just dial contrast and bias per your taste into the base Primer.

These are not final "looks" but per-camera designed image corrective tools you use for making your own look with improved final results. Four contrast distributions are there to offer four starting points and image manipulation approaches, each pre-defining the outcome with a healthy signal. I use these in our post.

These don't do creative decisions for you but fix the things which can be fixed so you get more neutral starting point and can focus on other things. Things which can be fixed are color accuracy, separation and consistency. Particular contrast distribution, light compensation, creative toning and accentuation etc are left to the Creative.

 

[Aharon Rothschild]

Reading though Hvroje's response it seems clear that he does not understand that this is a Camera Patch or Input LUT which has certain uses but does not improve the camera image. Graeme has been working to improve the camera image and there have been many iterations of REDcolor over the years to improve the camera at the debayer stage.
Input Luts have existed as a quick way to neutralize a camera's "issues" and the usefullness of any particular LUT is limited by the fact that thay are destructive. They cannot improve or change or optimimise or magically enhance the camera. If they could they would be part of the existing color science. The reason they are not is that there is a price to pay for correction, pushing for better skin costs somewhere else in the image and fixing one problem creates another. If Hvroje can explain how he has created a non-destructive LUT he should, as that LUT is worth far more than $55. If he cannot he needs to acknowledge that this LUT will not enhance all images, it will in fact harm some images, that this LUT works to change exising data, and that is serves the same purpose as any correction would except that it has been saved as LUT and is being sold on a nice website.

EDIT: again, this LUT may be usefull, It's being described in a way that will lead many to find it disapointing when it does not deliver on the promise of one stop image enhancement. I would hope that we can agree on what a LUT is and is not before engaging in further discussion.

 

[Hrvoje Simic]

Reading though Hvroje's response it seems clear that he does not understand that this is a Camera Patch or Input LUT which has certain uses but does not improve the camera image.

It only seems that way because it is more difficult to accept something else.

It is already explained what these do, not by using wishful thinking and assumptions but research and analysis and I have zero intentions of running in circles with you or Mike.

For convincing you came at the wrong address and according to seen rhetoric and reasoning those attempts would not bring success. You already painted your picture.

They cannot improve or change or optimimise or magically enhance the camera.

LUTs are data sets and cannot do anything.
This is done by a live Being with a brain and applied with a live Being with a brain using what is done and embodied into a form which is causing issues in your reasoning process.

The term "enhancing" came from your imagination, as well as a bunch of other strawmen you continue to battle.

If they could they would be part of the existing color science.

You are using either/or reasoning and flunking at synthesis.

This is made possible by base color science.
I already did this last year with RLF and DC2, Log3G10 & RWG brought more image potential, while making some things in the process more difficult and some easier.

 

[Aharon Rothschild]

As such please change your description at the beginning of this thread and remove the link to the product. Either it does what you claim it can or it's false advertising.