puzzling ...

[Damian Grady]

We shot some VERY rough green screen footage today and after previewing using the Medium Proxy in Quicktime thought okay, that'll do for these purposes.

However we also pulled the R3Ds into Redcine X 158 and the images could not look more different. We were using the camera's metadata in Redcine X so I we can't understad why there is this huge difference. between what we see in quicktime and what we see in Redcine X.

I've supplied the images for you to look at if that helps. The washed out one is from Redcine.

PS ignore the crap focus.

 

[Mike Prevette]

The QT plugin that allows viewing of the proxies has not been updated in some time. Redcine-x is full of new "magic" that has not yet been implemented in QT, Adobe, or Apple RED plugins.

 

[Damian Grady]

The QT plugin that allows viewing of the proxies has not been updated in some time. Redcine-x is full of new "magic" that has not yet been implemented in QT, Adobe, or Apple RED plugins.

Cheers for the reply.

We tried it in Redcine X 104 too and it looked the same. The proxies in QT actually look a hell of a lot closer to what we shot than what we see in Redcine X. Even if we go back to RAW in Redcine the green in the screen is miles out from what we shot or saw in the LCD.

 

[Martin Weiss]

What build was the camera on when you shot the footage?

In which colorspace did you view the footage on the monitor when shooting, and in which colorspace and gammaspace did you process the footage?

 

[Damian Grady]

What build was the camera on when you shot the footage?

In which colorspace did you view the footage on the monitor when shooting, and in which colorspace and gammaspace did you process the footage?

The build is 21, and the colorspace and Gammaspace were both Rec709

 

[Damian Grady]

Anyone else able to get their head around this?

 

[Nook Kim]

I encounter the same situation everyday. Clients come in for the color session, and many times, the first thing they say is "that looks so much different from what we shot and saw on set!" Then, some of them give me funny look about it as if I processed the footage mistakenly somehow..

I learned very quick from this, and I now make it look like the QT proxy (or Offline) color, especially when I work with new clients.

It's most difficult though, when clients just want transcoded image sequences or tapes. I actually have to grade it to look like what they saw on set.

I guess we will see some updates from Apple on this soon. rogar-Smoke(DBG):

 

[David Battistella]

You can start by boosting the contrast in rc-x to .70 right off the bat. That will correct the gamma. You need to also perhaps add saturation. The image will look awful in rcx but will look just about like the qt.

Redcolor and Redgamma are not read corre ctky by qt when it is handed over for rendering to qt.

This is a problem with how qt guesses at the gamma space of a particular codec.

It's weird to have a client sitting there looking at an awful image to match it to what they had on the day.

You need to set up two grades. Client grad and render grade but do not expect codecs to match the R3d file until a new redline and new qt codec are released.

For absolute accurate color accuracy you can use build 21 with clipfinder where r3d and qt match up really well.



David

 

[Damian Grady]

Thanks for the replies guys.

What I don't understand is why the R3D file in RocketCine-X/Redcine-X looks so far away from what we shot but the QT Proxy is much closer. If QT is striggling to interprate the colorspace and gammaspace I would understand more if the QT was highly inaccurate but it's the clips in Red's software that looks way out.

Are we doing something hideously wrong here?

 

[Nook Kim]

That's the question I had to ask myself. Don't quote me on this, but I came
to think that somewhere in the line when the guys developed the new color science, they may have tweaked the Redspace, etc.. to give us better resaults.
This would have been awefully great if Apple was quick to catch up by upgrading the QT codec, but it hasn't happened just yet.

I like what I get as a starting point with newer color and gamma options, especially the REC 709 is incredible to work with. It's just that most people tend to get accustomed to seeing the on-set look and get attached to it.
As a resault, when they see something different at the beginning of the grading session... "Uh... Ummm"

Red needs to take a break so the third parties have time to catch up... Not ;-)

 

[Noah Kadner]

You get that the RED RAW workflow means what you see on set is just one possible interpretation of the original sensor data- with a specific gamma/color space applied. And that you can recreate that in post by going into REDCINE or RED Alert! The QT Proxies are probably the least useful tool if you're looking for specific looks from the camera. Stick with the R3D files and grade them to your liking, then remake proxies or better- output a nice ProRes version instead.

Because really if you're going to be doing VFX/compositing you'll be doing it with something other than the R3Ds or the proxies. Typically either a ProREs 444 version or as a DPX or TIF sequence.

Noah

 

[CJ Roy]

At this point, is it even worth keeping the proxies? They haven't been updated since pre-build 20. Seems to just cause more confusion for editorial.

If they are worth keeping, can we start some sort of naming scheme for the various color science used? RedCS 1.0, 1.5, 2.0, etc... That way we can say the Camera is on REDCS 3.2, but the QT proxies are on REDCS 1.6, or something to that extent. Might cut down on the hours of explanation done each week on this.

The emotionally scarred faces of any producer looking over my shoulder while using Clipfinder should warrant some sort of vocabulary fix for this.

Thanks.