Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Improving the Camera: The Image

Luigi rocked with material and last sample shots are also his.


luigivaltulini said:
BOOOOOMM ;)

MAgic Hrvoje.

... It was time I was expecting this ....


It would be a nice 17x17, for the "boys" of Scarlet-w eheh ;)
Click to expand...


That was the original plan, a set of 12, with the free 17^3 cube for testing, but the level of precision required to properly fix the signal is not met by 17^3 and following that desire, regardless of the best intentions would bounce back in my face. 33^3 is already borderline.

Hats off for a leap of faith, good will and all the imagery.
Your set is heading to you soon.
 
So to use these in camera, on weapon magnesium, I'd need to not use redlogfilm correct, is redwidegammut available for magnesium weapon?

I would also have to convert to 17x cube using lattice
 
Would you say the 33^3 Log3G10 RWG is good enough for burning in to prores or should you 100% only use the 64^3 version in post even for those projets that only requiere prores.
 
Joel Arvidsson said:
Would you say the 33^3 Log3G10 RWG is good enough for burning in to prores or should you 100% only use the 64^3 version in post even for those projets that only requiere prores.
Click to expand...

I have to do more comparative analysis to see what is affected and to which extent.

Kenny Mosher said:
Will Helium primers be up soon?
Click to expand...

Can't tell, the whole process consists of gazzillion variables and I'm doing this between two jobs, with a new feature on my desk.

I've done a few Helium based cameras profiling and analysis. At this point various r3d grabs speed up the testing process so who wants to help with that can reach out at exert at omeneo dot com.
 
Dragon.

Tommaso Alvisi


syxdmb.jpg



25zs278.jpg
 
Take vacation and gets your hands dirty on the Helium version!
 
:)



Dragon.


Jacobo Martinez




t7iq6h.jpg


i1cil3.jpg








2zejx5h.jpg


feqhpu.jpg
 
Dragon.


Luigi Valtulini


11smus3.jpg




2z51yx1.jpg
 
Interesting work.

Some of the skin samples don't work for me for the reason why the outdoor samples work. The tone mapping of the highlights into range looks great on skies but the highlights on skin are reduced in a way that looks a bit un-natural IMHO. The skin colour seems off at times too, but in other images looks good. Skin can be so difficult that one LUT will never do.

It's a shame this is all LUT based. Ideally you want a nice plug-in that affects the image using local contrast and adjustments, the LUT approach, whilst simple can be a bit heavy handed.

Nevertheless it's a great looking image for the most part...

cheers
Paul
 
Hey Paul, thanks for the input.

Skin is corrected in all examples, these fix the offsets and the rest is light, WB, environment cast etc. Subjective impression varies depending on many factors. You get to choose more specific contrast grading into the Primer. There are 4 signal distributions for this very purpose allowing different approaches in post. Hues, separation and DR are mapped more accurately and you get to choose the rest on a stable signal. That is the whole point.

There is no single contrast distribution which works for all shots. That is what color grading is for. This doesn't replace a colorist but give him a more accurate basis to start from.
There is a reason these are called Primers and not "looks".

One should perceive this as a new type of tool, and not see it through previous impressions of LUTs. This is why I avoided mentioning LUTs form the start of the thread and gave three important points in a first post.
LUT is only a form, body, which can carry different signal modifications. This very specific signal treatment requires this form.
 
I think I understand it now. Perhaps it's a language barrier thing, but let me try and put it plainly for those that are confused.

Hrvoje is not making a subjective "look". He's not using a "creative licence" to make the image look nicer to his personal taste. He's taking an objective approach to put the colors captured by RED cameras in a more comfortable place before grading. This is why he's calling it "optimization" and why it's called "Primers." In real life, you use a primer undercoat before you do any serious painting. Same idea here.

Theoretically, this will lead to less manipulating needed in the color chain for a basic look. People always compliment Alexa footage for being able to grade easily with just some contrast and saturation.

All of this said, I think I speak for everyone here in wanting to know what your methodology, is, Hrvoje.

If it's a totally objective approach, what is it you are doing exactly to look at the information recorded and know exactly where each color needs to align? For example, how do you know for sure that butterfly needed to be more orange than yellow? How do you know that boxer wasn't actually bathed in yellow/green light? If I have yellow lights in my shot purposefully, are they going to skew orange against my will?

You don't need to give away your secrets, just convince us why your approach is correct.
 
Do you plan to do primes for epic mx aswell once you finished the Helium version? I will keep using them side by side.
 
Excellent post.

Aaron Lochert said:
Hrvoje is not making a subjective "look". He's not using a "creative licence" to make the image look nicer to his personal taste. He's taking an objective approach to put the colors captured by RED cameras in a more comfortable place before grading. This is why he's calling it "optimization" and why it's called "Primers." In real life, you use a primer undercoat before you do any serious painting. Same idea here.
Click to expand...

Yup.

And now people can do that with other cameras as well.

Aaron Lochert said:
Theoretically, this will lead to less manipulating needed in the color chain for a basic look. People always compliment Alexa footage for being able to grade easily with just some contrast and saturation.
Click to expand...

Exactly.

It is already in practice for a year or so. :)

Aaron Lochert said:
All of this said, I think I speak for everyone here in wanting to know what your methodology, is, Hrvoje.

If it's a totally objective approach, what is it you are doing exactly to look at the information recorded and know exactly where each color needs to align? For example, how do you know for sure that butterfly needed to be more orange than yellow? How do you know that boxer wasn't actually bathed in yellow/green light? If I have yellow lights in my shot purposefully, are they going to skew orange against my will?

You don't need to give away your secrets, just convince us why your approach is correct.
Click to expand...


Firstly, convincing people is a waste of energy. That's what politicians, media folks and sellers attempt to do. :)
I can share my findings and results.

I developed a methodology to see what is hitting the sensor when mapped in particular gamma and gamut. Then I developed a methodology to re-configure that through a highly particular transformation to a usable gamma and color space and wrap it in a body of a 3DLUT.
So the image gets improved in accuracy, separation and consistency without falling apart.

Greater part is objective, smaller a mix. That part which is less objective must be that way since there isn't a single goal which this has to meet but a range of those so some decisions have to be made. Camera and post experience used with multiple points of reference and factors of practical purpose are used to ensure those decisions meet overall usability and neutrality and not a personal whim.

Objective part covers hue accuracy, hue separation, DR mapping, signal "health" etc.
The part which is more subjective is additionally designed within the context of shooting and post so there are 4 signal luminance distributions useful for different scenes and approaches in shooting and post, while being friendly for the particular camera's signal.

This way the shooter knows what he's getting and colorist won't have an adventure when having to achieve something the shooter thought he'd get. Or meet what someone assumed would be possible. This way, the shooter and the colorist get more accurate and stable basis to rely on. No need to fix many things, signal deviations which the digital domain brought and we can get back to exposure, WB, scene adaptation and creative toning.

For example, in digital, missing WB can get messy quickly because of particular color offsets which further deviate and make the image intuitively unappealing and exponentially more difficult to correct. When this treatment is applied WB gets more creative since overall colorimetry is stable and there is only a pleasant toning impression.
One of the benefits of this solution is more creative range with color temperature.

Signal modifications and optimization are not done based on eyeballing and guessing the imagery, this type of real scenario imagery is used to apply what is made through usage of multiple custom-designed reference material. When the results are applied on the imagery natural colors pop out and everything falls in its place. As with everything, there is of course the teething phase so some things may need re-tweaking and there is some kicking and screaming along the way. That is normal. I've done this before.

The answer to your question "How do you know what the color should be" (if focusing only on rational part) is - because through detailed camera analysis I know what each camera does with the signal and I've learned how to improve it on micro levels before it hits post. So there is no need any more to waste time, focus and energy on repetitive "fixing" of digitally acquired image.

So when that is done with Dragon, there is no more this (and related image properties):


28rfuc0.jpg



But this:


14b2uth.jpg



Image by Lauri Hakala.






Btw, Alexa also has its quirks and trainwreck scenarios and yes its results can also be improved before the material hits the grading suite. And no, none of this replaces a creative Human, shaping the material to context. It improves his workflow, bringing the feeling...which is best left to be felt directly. :)
 
Joel Arvidsson said:
Do you plan to do primes for epic mx aswell once you finished the Helium version? I will keep using them side by side.
Click to expand...

MX is on the way. I have other cameras to take care of first.
 
Thanks. Looking forward for the MX & Helium versions.

Could you post one image but with all versions of the prime on? So its easier to just se how they affect the image. You can alway mark the one that you think it appropriate to use on the given image.
 
You must log in or register to reply here.
Back
Top