“Set Up”- Bruce Willis, Ryan Philippe and Curtis “50 Cent” Jackson.

[Zé Caralho]

It's not the most stunning looking film, but it's not as bad as some of Soderbergh films, such as The Informant! To be honest, I can see why some people might not like the look (or grade) of this film, but it's not that bad. As someone else stated, The Book of Eli was all over the place, regarding color and highlight, but it fitted the genre of that particular picture. With this one, its the kind of gritty, dirty kind of look they went for.

Granted that if someone else had shot it, someone like Cronenweth or Deakins, even going for the same image treatment, it would probably have looked a lot more classy, but hey, it is what it is now.

 

[Jake Bastian]

Your Nick is offensive, and to the portuguese community that read this forum it can only do harm. I suggest the mods look up Caralho.

I'm guessing Zé may be his first name but Caralho is more of a slang term - it's probably not overly offensive as long as you're not looking at the word "caralho" in the absence of the name "Zé". Solo it's similar to the F-bomb or talking about your member but with "Zé" which I believe is the english equivalent of the name "Joe" it more like an expression of surprise - saying DAMN Joe (like Martin Lawrence's "DAAAAAMN GINA!" for those looking for an old school throwback) I'm sure with the rules it'd be better to change names to something like '"Zé Caralho' surname".

The guy doesn't exactly come across as pleasant in this situation, but I wouldn't turn a little anti-fanboy trolling into an international incident when many of us don't really understand Portuguese slang. (Myself included)

 

[Paulo Kobayashi]

Hey a Brazilian here. Indeed it's hard to believe that he is using his real name, although it's not impossible.
It doesn't bothers me that much, but I like the idea of using real names because it makes people accountable for their words and opinions. I'm sure that if reduser allowed people to use nicknames it would be infested of childish trolls that use forums and social media just to satisfy their lack of motherly love.

@ Ze. About your personal taste for Alexa, it seems reasonable, I like the looks of it very much too. But your insistent remarks that the MX sensor clips on highlights seems lacking technical details and it makes me believe that it is a problem with your workflow or your knowledge in color correction and post using r3d in general.

Just to state your credibility why don't you tells us what you do? Are you a DoP? Are you a Colorist? What tools do you use? Nucoda? Da Vinci Resolve? Apple Color? Are you using an 8 bit or 10 bit monitor? Are you monitoring in YCbCr 422 or RGB 444?

What is your workflow with Alexa and the MX? Do you color correct the proxies directly in FCP? Because Arri Log-c ProRes workflow will not work the same way with r3d. In fact I can tell you that if you treat r3d the same way you treat an Alexa Log-C ProRes you will get exactly what you are saying.

Can you show us some examples of your work with both cameras and the highlight clipping on MX? I tried to find your name on IMDB and couldn't find anything.

R3D is a digital negative, a very powerful one. It makes a huge difference what you do with it.

http://colorgrader.net/index.php/di...avinci/265-aces-vs-yrgb-using-red-camera.html

Here is an example of a test I came across the other day on COLORGRADER NETWORK, where the same MX r3d footage in YRGB had much less details on the highlights than in ACES. And the reason was just because the YRGB, even being 10 bit, was not robust enough to "resolve" all the information contained on the 12 bit Raw of r3d. This is easy to imagine because the difference between a 10 bit rgb and 12 bit XYZ color space will be on the very end of the edges of the color gamut and where you need a lot of "space" to maintain a soft gradation, like the clouds in the example.

(BTW the conclusion of the reviewer on the article was incorrect. He believes that the MX sensor had more details that the EPIC, but in fact what he did was comparing 2 very bad lit Epic footages with 2 direct sunlight MX. It makes no sense to compare the details in the highlights of a bad CFL practical lamp to a sunlight FULL spectrum footage. Anyways aside from that it was a very nice experiment from his part.)

 

[Paulo Kobayashi]

One more thing. The ACES look desaturated because we can only see it in 8 bit from the JPG. Even in a 10 bit system it would still "seem" desaturated but in reality the saturation is still there. What it does is to "push" everything "down" relatively to the rest of the image so we can see the details on the edges (for example the clouds.) In other words, if you could see in a 12 bit ACES monitor (which is imposible btw, ACES is only an GIGANTIC theoretical color space) you would be able to see the color equally saturated.

ACES workflow is very useful to match VFX and / or footage shot on different cameras, but it's not a format for deliverables. One should "convert" the footage to the target color space, apply the 3D Lut for your monitor and the intended output and from there fine tune to the desirable final color.

 

[Zé Caralho]

Hey Paulo, thanks for your post. It's a very interesting test that shows some things I wasn't aware of. I'm not a Dp nor a colorist, I'm a director. But I do nonetheless love the etch side of things.

I will read this article.

And yes, it does make sense to grade on a wider color space and its good to know that the highlight information is actually there and not lost during actual shooting. But what happens when I see these kind of clipped highlights on a motion picture like 'Contagion' ? http://trailers.apple.com/trailers/wb/contagion/

I would guess that the colorist (he color corrected everything from Harry Potter to Black Swan, and has been timing Steven's films since Che), so I believe that he must have done his workflow right. This is why I concluded this was mainly a consequence from the MX highlight algorithm design.

Because when I saw In Time, http://trailers.apple.com/trailers/fox/intime/ and Anonymous, http://trailers.apple.com/trailers/sony_pictures/anonymous/ I didn't see the same highlight behavior, be it direct sun or practicals.

I know that grading has a lot of say in these clipped electronic whites, and I didn't see this in either The Social Networks or The Girl With the Dragon Tattoo, which look absolutely perfect. But I guess that Jeffrey Cronenweth dealt with this issue by underexposing his main exposure, very much like pushing film, where you underexpose and bring back the underexposed stops in the development stage to preserve more texture and information in the shadows, but with digital, i's the opposite, you loose the highlights, not the shadows, so the same method was used, but this time to protect the highlights, like Harris Savides did so amazingly in Zodiac.

But I'm speculating here. But surely is confusing, because on one side, the test done by Colorgrader Network makes perfect sense, but on the other, there are a good handfull of films that have these clipped highlights, from major motion picture studios, that should not have any problems getting their workflow right.

I'm confused...

 

[Jeff Kilgroe]

Going to have to retract some of what I said above. I guess there is a very slight possibility that it's the guy's real name, but on further examination I'm going to say this is probably not the case. Previously he made a comment about being able to go by whatever name he chooses. Then seems to waffle and say it's his real name. But when I look around, there are too many "Zé Caralho!" t-shirts, bumper stickers, etc.. Gonna have to side with the large number of complaints about the guy's name. Sorry, "Zé", but if you want to go by that, you'll have to show us some bona-fide credentials.

 

[Paulo Kobayashi]

I don't know what happened on those films. What they did before with other footage does not guaranty that they know what to do with r3d. The workflow is always changing and new technology is being created to tackle the new chalenges. What I know for sure is that the MX sensor and the Epic/Scarlet are capable of capturing more than enough information to never clip highlights, and I'm not even talking about HDR. And another truth is that nobody ever admit making mistakes, even when is their first time working with a new technology.

In general, people that came from film centric workflows don't like to learn how to work with r3d and make a lot of mistakes. Arri had the knowhow and the desire to create a workflow that emulates what film feels like on a 2K digital scan in post, and they succeeded.

New comers are blank canvas, they learn the new techniques and embrace the workflow. If you want to compare something get works done by pro Epic teams that know how to treat their footage and pro Alexa teams that know what to do with it. I'm pretty sure both will look amazing.

I bought a Scarlet, but that does not mean I don't like Alexa. Both create great pictures, but Scarlet let me use it for Still works, I love the power of r3d, I can upgrade the hardware and firmware is free, I can get away with bigger crops, and IT COSTS MUCH LESS.

If you are going to BUY something it's a no brainer, but if you are going to rent, your budget is the limit. Go with Epic, film, Alexa, Scarlet, 5D, whatever you can afford and fits your desired look.

All the best. Sincerely.

 

[Peter Johnson]

I worked on this movie as the A-cam 2nd. Lots of stories to tell. I wouldn't give Steve Gainer too much grief. Our schedule was incredibly tight as we did the shoot in about 25 days. The footage was all exposed just fine. So you must be referring to the grade. Also we did lots of action shots with a Go-Pro and several 5D's. I mounted the Go-Pro on the bumper with crap loads of gaff tape. The stunt where the girl is thrown through the glass table chipped the front element of our 10mm on the B-Cam (always use optical flat when glass and explosive charges are a mere 2 feet from the lens. - helpful tip to remember).

Also, Anton Baur batteries won't charge if they are below 45 degree's Fahrenheit. We found that one out the hard way.

 

[Graeme Nattress]

I'm confused...

You can get ugly clipped highlights and nasties on ALL cameras.

"argument by trailer" is not a particularly good one.

"This is why I concluded this was mainly a consequence from the MX highlight algorithm design." - there's no issue there. That you guess this from trailers is not the best argument.

"which look absolutely perfect." - shows exactly that there's no issue with how MX sensor handles highlights. We give you many tools to help ensure correct exposure. Some people need them (like me), so use them, others don't, everyone has a different idea what an image should be like.

REDUser is a nice place to come and ask questions and learn - let's keep it that way!

Graeme

 

[Christian Munoz D]

I worked on this movie as the A-cam 2nd. Lots of stories to tell. I wouldn't give Steve Gainer too much grief. Our schedule was incredibly tight as we did the shoot in about 25 days. The footage was all exposed just fine. So you must be referring to the grade. Also we did lots of action shots with a Go-Pro and several 5D's. I mounted the Go-Pro on the bumper with crap loads of gaff tape. The stunt where the girl is thrown through the glass table chipped the front element of our 10mm on the B-Cam (always use optical flat when glass and explosive charges are a mere 2 feet from the lens. - helpful tip to remember).

Also, Anton Baur batteries won't charge if they are below 45 degree's Fahrenheit. We found that one out the hard way.

Thank you peter for the insights. So, I wasn't to far when I was guessing the movie was shot with DSLRs, etc. That maybe explain the motion artifacts on several scenes.

Can you give us more info about shooter speeds, glass and lighting used on the film? Will be great to use this film as "case of study" Thanks.

BTW: I added and edit to my first post to put in context the question.