Improving the Camera: The Image

[Hrvoje Simic]

Fascinated by this discussion. Any love for FS700 S-Log 2 owners, Hrvoje?

Gladly but depends if I can get it for proper analysis, I'll check with colleagues.

I like the results of your color pallet Hrvoje, it's a pleasing look.

Thanks, John.
Nature takes the credit for the palette, Red team for capture potential through sensor and base colour science, Erich for capturing it, I deal in signal optimization before the eye of the beholder.

If you haven't already, I would recommend stress-testing the LUTs on some saturated ramps & gradients, such as this free image we've created:
http://truecolor.us/downloads/lut-stress-test-image/

Cool stuff, thanks for your work.
I cannot use it for my work but hats off nevertheless.

Another potential of this lies in using it as practice in grading software, to understand the principles of a digital image and color grading tools.

 

[Hrvoje Simic]

Hrvoje, can u post the LUT and/or solution here or is this still work in progress ? what exactly is the implementation ? is it a LUT ? is this paid or free... ?

in short: WHAT EXACTLY IS THE DEAL IN PLAIN ENGLISH ?!? ;-)))

Info in appropriate time.
Guesswork and relying on assumptions based on abstraction of the known is likely to mislead.
In this case it is better to start from the domain above thought and the images themselves.

 

[Joel Arvidsson]

Hrvoje, is this a 3D LUT? When do you think its out of beta and when can we se higher rez "before&after"?

 

[Hrvoje Simic]

Out of beta for Dragon depends on beta tester feedback and before/after samples on willingness of shooters to share publicly usable r3d grabs. Those are welcome on exert_at_omeneo_dot_com. Material I use is either from own internal analysis or production jobs and not for public display. Helium work is still in progress, I've been doing more camera testing and sensor analysis during last few weeks.

 

[Bob Gundu]

If you need more Beta testers, let me know. I'm still on MG Dragon Weapon.

 

[Cameron K]

Gladly but depends if I can get it for proper analysis, I'll check with colleagues.

I hope so, Hrvoje. Your prices are good and there's still an awful lot of FS700 Odyssey 7Q+ owners out here (despite your very legitimate criticism of its ergonomics). You might save us some grief and extend the life of this tool. Count me in.

 

[Mike Nagel]

Guesswork and relying on assumptions based on abstraction of the known is likely to mislead.
In this case it is better to start from the domain above thought and the images themselves.

Hrvoje, you have been informed that I will steal these 2 lines from you for a new script. It's for a character with a massive full beard that lives in a barrel behind the liquor store.

he keeps screaming: "BEWARE... THE DOMAIN ABOVE THOUGHT !"

;-)))) hahaaa


Now, feed the fishes some breadcrumbs: is this going through the API or LUT based ?

 

[Hrvoje Simic]

Hrvoje, you have been informed that I will steal these 2 lines from you for a new script. It's for a character with a massive full beard that lives in a barrel behind the liquor store.

he keeps screaming: "BEWARE... THE DOMAIN ABOVE THOUGHT !"

;-)))) hahaaa


Now, feed the fishes some breadcrumbs: is this going through the API or LUT based ?

Not sure if ridicule or benign joke but funny.

Behind-the-liquor-store-bearded-barrel-tennant removal attempt:

When confronted with the new matter to assimilate the analytical property of a rational part of mind automatically searches and grabs on to known data, abstracts, compares and categorizes, forming the impression of the new which can be rationally assimilated through relations of known references.
"ball = not square, orange = like ball".

With this automatical reasoning process in cases where there are key distinctions between the unknown and assumed, the result of that reasoning is likely to be innacurate and the great part of impression a projection. Difficult to remove because the person already "thinks" it knows. Which is why children always learn faster intuitively knowing they don't know and adults tend to stop learning at some point by "thinking" they do.

Image speaks to "supra-rational" mind.



Btw if I recall correctly I sent you a message a while back but the pidgeon got M.I.A.

 

[Mike Nagel]

Not sure if ridicule or benign joke but funny.

Behind-the-liquor-store-bearded-barrel-tennant removal attempt:

When confronted with the new experience the analytical property of a rational part of mind automatically searches known data, abstracts, compares, evaluates forming the impression which can be rationally assimilated.
"ball = not square, orange = like ball".

With this automatical reasoning process in cases where there are key distinctions between the unknown and assumed, the result of that reasoning is going to be innacurate and the great paet of impression a projection. Difficult to remove because the person already "thinks" it knows. Which is why children always learn faster knowing they don't know and adults tend to stop learning at some point by "thinking" they do.

Image speaks to "supra-rational" mind.

Btw if I recall correctly I sent you a message a while back but the pidgeon got M.I.A.

all I can tell you is this: know (--> understand) your audience !

and you are shooting way over the head here... since you've already invested the time, and I can clearly see from your repeated frustrated messages that now FINALLY Hrvoje u appreciate the hustle & grind ;-))) hehehe, it would be a waste, and very inefficient to lose your target out of sight at the finishing stretch

so, while the philosophical dabbling and the meaningless obfuscation cracks a warm smile on my face, there is a time and place to establish abstract concepts and ideas, and then to be very clear and precise on actual in-case application.

In addition, while the abstracts and approaches are all nice and peachy, ultimately only one thing counts: usable end results. Preferably in a wide range of scenarios.

So I'll patiently wait for more details and A/B tests, but more importantly details on the excact implementation... lots of dudes who know what they're doing grade wise, already implement many manipulations into the signal flow... as we all know: more than we want to with RED - out of forced necessity.

We will be very careful to implement further signal processing unless we know what's being done... and at what stage. The originally fat RED layer gets thinner and thinner every step...

Looking forward ! Release the beast !!!

P.S.: I think my reduser inbox was full.

P.P.S. I will now vanish into......... THE DOMAIN ABOVE THOUGHT !

 

[Hrvoje Simic]

and you are shooting way over the head here... since you've already invested the time, and I can clearly see from your repeated frustrated messages that now FINALLY Hrvoje u appreciate the hustle & grind ;-))) hehehe, it would be a waste, and very inefficient to lose your target out of sight at the finishing stretch

There is no frustration Mike. Erase those windmills please so we can have some fun.

so, while the philosophical dabbling and the meaningless obfuscation cracks a warm smile on my face, there is a time and place to establish abstract concepts and ideas, and then to be very clear and precise on actual in-case application.

Warm smile here as well. Patience is a virtue.
All in its time...

In addition, while the abstracts and approaches are all nice and peachy, ultimately only one thing counts: usable end results. Preferably in a wide range of scenarios.

Good news. That is the whole point and the reason this work takes months.
There was an offer for beta testing on "workflow" board last month and those who applied got to check directly.

So again - I cannot publicly post my working material, nor I'm going to chase folks around the forum nor interested in introducing mr. Sleeve Pulling. More before/after in this thread is no problem - so anyone interested for results on their own shots in this pre-release stage can send .r3d grabs with olpf and sensor info on exert at omeneo dot com.

So I'll patiently wait for more details and A/B tests, but more importantly details on the excact implementation... lots of dudes who know what they're doing grade wise, already implement many manipulations into the signal flow... as we all know: more than we want to with RED - out of forced necessity.

Now wouldn't it be cool if those undesired manipulations could be avoided...so there was more time and focus on desired ones...

We will be very careful to implement further signal processing unless we know what's being done... and at what stage.

No problem. All in its time.

 

[Mike Nagel]

re windmills: just teasing, because I share the pain brother, believe me... been through that many times in many fields, it is at times a mindfuck... especially when there is not a lot of credible source information available. But, what doesn't kill you make you stronger. Or: smarter ! ;-))))

I have to say, at a quick glance, on the pic with the Target building above, the treatment does seem to reduce the very artificial appearance of colors a bit, which has been one of the major caveats with Helium in very low light shots... good stuff !

 

[Hrvoje Simic]

Helium.

Erich Roland, DC Camera

Same solution on all.
Work in progress.



...

 

[Dan Kanes]

Mr. Simic,

I'm in.

Currently using Helium.

Wondering if I'll be able to use this in camera somehow while shooting - or if I can use a LUT box.

Either way, I'm in.

 

[Hrvoje Simic]

...

 

[Jake Wilganowski]

These pics are good. Stop teasing us.

 

[Mike Nagel]

color seem to be less artificial, but the 1D also gets lifted...

 

[Hrvoje Simic]

These pics are good. Stop teasing us.

No teasing, Jake.

color seem to be less artificial, but the 1D also gets lifted...

Good call again. There is a whole set of reasons for colours looking less artificial.
The impression of "lifted" also depends on particular image exposure and reference.

 

[Mike Nagel]

would be great to see the 1D to be matching or very similar for a more apples to apples comparison... perception plays funny games when u change too many parameters at the same time...

what was that last shot ? Helium ? And how was the reference developed ?

 

[Hrvoje Simic]

These all are Erich's Helium shots.

Don't take any of these as a reference for analysis, this is just to show some of the improvements in signal mapping so listen to the initial inner "feel" with these. Matching the latter to prior in what one might first associate with gamma is neither the priority nor useful for what I'm doing and would just confuse you. Both are results of a different base and routes.

Hey Dan, I'll catch you once I sort some more things with the Helium base signal. I see more stuff to improve.

Let's make it more fun.
Anyone with skin grabs ?

 

[Hrvoje Simic]

...