More Color Tests On Build 20

[Mark David]

Art has been doing a lot of tests with Build 20 lately and I found this one quite interesting. Take it for what it's worth, I haven't had time to try any of these tests myself but I feel it's important to know the strengths of any system you rely on.

http://provideocoalition.com/index.php/aadams/story/what_you_need_to_know_about_green_and_blue_and_red/P0/

[danbrazda]

Art has been doing a lot of tests with Build 20 lately and I found this one quite interesting. Take it for what it's worth, I haven't had time to try any of these tests myself but I feel it's important to know the strengths of any system you rely on.

http://provideocoalition.com/index.php/aadams/story/what_you_need_to_know_about_green_and_blue_and_red/P0/

Speaking from Switzerland....

What is your agenda? That was a REALLY long way to go to say Red ain't what it's cracked up to be. My theory:

-You are committed to a much higher cost, marginally better acquisition format and have no way out
-you work for Arri or Sony
-you have too much time on your hands

I, on the other hand, have just cued up Baraka on Bluray. Seeya......

[CJ Roy]

Why is it always a conspiracy? This is pretty important information for DP's.

[Gavin Greenwalt]

Speaking from Switzerland....

What is your agenda? That was a REALLY long way to go to say Red ain't what it's cracked up to be. My theory:

-You are committed to a much higher cost, marginally better acquisition format and have no way out
-you work for Arri or Sony
-you have too much time on your hands

I, on the other hand, have just cued up Baraka on Bluray. Seeya......

What's your agenda? That was a REALLY long way to go to say you don't care about the results of article. My theory:
- You can't afford a better camera and will do anything to justify your purchase as just as good as what you can't afford?

[Mark David]

I have no AGENDA other then informing others on this board about what I thought was a great article? Lets steer the conversation back to the results of the article if at all possible...

[Jannard]

I think it is great that Art is trying to understand what we are doing with our color. But I'm not sure, other than interesting science stuff, exactly what the point is. The RED ONE makes stunning images (Headline)? That, like all cameras, it is not perfect? That it is the best $17,500 camera?

While attempting to explain what Graeme is doing is certainly entertaining (but only partially correct), I keep looking at the article for something that helps one shoot better images. I can't find it. Maybe you can help me translate the science into something useful?

Jim

[Andrew Thomas]

I'm kind of in the same place as you, Jim. I do think it is important to know the strengths and weaknesses of your gear, but at the end of the day I don't know what Art's conclusion really means for people using the R1.

[jbeale]

The provideocoalition article is assuming something about the RED spectral response. I'm puzzled by that, why speculate about something you can measure directly? I did, and I show my results here (Test #6): http://bealecorner.org/red/test-080108/page2.html

I won't claim my results are definitive. I didn't use any lab instruments, and this was an old firmware version anyway. But anyone can do the same thing with a $5 diffraction grating and a shoebox, so if spectral response is important to anyone I'd think they would just measure it directly. I'd be curious to see what other people find.

By the way, as I recall Graeme or another red team member actually did publish a RGB sensor response curve here, some time ago. (I'm also surprised to find anyone else interested in this sort of thing- I did similar measurements on some other cameras some time ago- never had much interest from anyone in them.)

http://bealecorner.org/red/test-080108/pics/Red-color-parts.png
I suppose I should also say something like "the RGB that you can see is not the true RGB" because you don't have direct access to the sensor, just whatever the R3D decoder gives you, and the color matrix could be doing just about anything to give you the nominal "rgb" values in the output file.

[Blair S. Paulsen]

I have always assumed that part of the master plan for getting the best picture from a Bayer pattern sensor included techniques that could be termed channel mixing. Moreover, wouldn't a sophisticated algorithm for de-bayering carefully balance both color and luma data points to make the best extrapolations possible?

To me, the RedOne camera system has inherent characteristics that skew fidelity from nominal just like every other imaging device - including human vision. Understanding how that impacts everything from art direction to lighting choices is part of the craft.

With most electronic cameras there is a ton of on-board DSPs that compensate for a number of issues in the sensor/optical block long before any image comes out. The RedOne is a different beast and when you judge it using metrics designed for different topologies its hardly a surprise that some of what you discover does not fit your assumptions.

In the immortal words of Sean Fairburn - test, test, test.

Cheers -#19

[Mark David]

I have been following Art's articles on provideo for quite some time and had published one of his articles earlier, but forgot to follow up with the rest.

Personally this information hasn't made me want to shoot more or less with RED, but I have been a fan of Graeme since my days at Carleton University (Ottawa) when I was messing around with the nattress plugins...so I guess a "geek" attraction comes into play when I hear someone digging deep into his work.

As far as real world information goes - jbeale makes a good point, but I think Art was very curious to see if fixing the "blue" under tungsten would affect how we use the camera.

[Lauri Kettunen]

About the articles in Provideocoalition, is the argument in this

http://provideocoalition.com/index.php/aadams/story/the_vertical_bar_graph_raw_exposure_meter/

about the meter correct? When the pole hits the top and red color, does it mean a channel is already clipping? At some point Jim said the zebras in raw view are the reliable way to know of about true clipping.

[Antony M]

And why is there nothing in the Red Ops Guide (even in the latest one for Build 21), which explains (or even mentions) the "rainbow meter" to the left of the RGB traffic light display???

[Gunleik Groven]

(...)

While attempting to explain what Graeme is doing is certainly entertaining (but only partially correct), I keep looking at the article for something that helps one shoot better images. I can't find it. Maybe you can help me translate the science into something useful?

Jim

What is interesting - if true - and will need to be tested for "best practice" exposure & RAW development - is his suggestion that red RAW past B20 isn't as "raw" off the cam as it used to be.

That if you record with a WB of 5600 and a wb of 3200, the RAW file is no longer equal - like it used to be - given that exposure/subject/lighting is equal.

That would have some rather significant implications. And needs to be tested.

Other than that, generally understanding how the sensor and format behaves, has been a good thing for the operators and post people up untill now. I guess it still is. So in that respect, I think this is an interesting article....

[Lauri Kettunen]

That if you record with a WB of 5600 and a wb of 3200, the RAW file is no longer equal

IF this is true, then of course, this should be an option that can be switched off. Otherwise it's like Windows software which has all sorts of automatic features makinf the system to guess what the user really wants to do. Would like to have a button that enabled me to switched all such automatic features off at once.

Nevertheless, if the WB setting affected the raw file, then the foundations of the whole workflow are somewhat ruined. Such a choice should be left to the user.

[Roberto Lequeux]

From the last page: "The good news is that RED’s white balance really is meta data. Nothing is “baked in” to the look, and that’s the RED’s greatest strength."

The title sets a negative tone, followed by two pages of very repetitious digging. All the way at the bottom of the last page the full picture is finally addressed. Not the best structure. A little weird when build 20 is nothing but better. :confused1: I am really confused now. Is there something to watch out for other than completely messing up WB?

[Gunleik Groven]

OK.

My bad

Channel-switching as such is nothin' new as one of many tools for the right occation.

I agree.
B 20 looks only good.

I am one (of the very few, I guess) that really did not like what Build 16/17 did with the image (many reasons). Now it's more like getting my old cam back, only like muuuuuch better.

No complaints.

But I want my RAW RAW! -;)

G