Color Grading....HELP! Dragon and MX

[Christoffer Glans]

If it was possible to get the same result with an MX through grading so that it looks like Dragon, why bother with getting the Dragon sensor. This is where things get hard, but I would hire an experienced colorist to do this job, it might be educational for you to match them up if you just started grading, but if you're on a deadline you might not get much sleep over this.

If you had any color chart shots, like the x-rite, it's possible to match up cameras through an automated process in Resolve. That might also be the way to go in future projects if you use the same setup.

 

[Scot Yount]

Christoffer,
I am ordering x-rite shortly for the very reason you point out. In reality I have just started grading, and while people say, just get a colorist (and I will if the project pays enough) I have been able to teach myself a great deal in this life, by just doing and I want to learn as much as I can. I am not aiming to become a colorist (I will never have that skill) but rather, a director who shoots and grades when needed. This forum has been fantastic for that reason. I know that the more I know about something that I need someone else to do, the better our collaboration.

 

[Marc Wielage]

If you had any color chart shots, like the x-rite, it's possible to match up cameras through an automated process in Resolve. That might also be the way to go in future projects if you use the same setup.

The automated Color Match process is not perfect. Here's a short video from Gray Marshall showing how it can fail under certain lighting conditions:

My own experience is that getting a decent colorist to match cameras and keep dailies consistent is a far better choice. Monitors, eyeballs, scopes, and good color charts can get you there.

 

[Scot Yount]

Marc,
I find your input on these boards always very helpful. I have learned a lot from you by just reading your posts. That's why I keep coming back here.

 

[Brett Harrison]

Try applying DC2 to the MX footage

 

[Scott Crawley]

A couple more minutes with the charts for one last go; I believe these are technically much better.

DRX and jpg files are here so you can import the grades into Resolve. https://www.dropbox.com/sh/s4rb6bpm87uzc9a/AABzFv3azZiBSPnhhdmYWZ-qa?dl=0

 

[Jake Wilganowski]

These look magenta to me

 

[Scott Crawley]

The automated Color Match process is not perfect. Here's a short video from Gray Marshall showing how it can fail under certain lighting conditions:

My own experience is that getting a decent colorist to match cameras and keep dailies consistent is a far better choice. Monitors, eyeballs, scopes, and good color charts can get you there.

It seems pretty clear that his chart was reflecting colors from a variety of contaminants around him: the red wall, cars in the alley (One of which we can see.), and probably another wall. If those contaminants were not present, I'm sure Resolve would do a better job.

I have had a failure rate of about 35-40% with the new tool and those shots were mostly acquired indoors where the set would have been better controlled than his alley. 60-65% of the time it either got me very close, or brought me to a reasonably good starting point.

 

[Scott Crawley]

These look magenta to me

I see what you mean. Hmmm... when I bring magenta closer to target his entire shirt and many mid greys go very green though... Back to the drawing board I guess.

Are you seeing it in both the MX and Dragon corrections or just the MX, because the Dragon looks pretty near balls on on my scopes.

Is this better do you think? I had to make additional adjustments for those green mids. That could be due to the LED, but I fear it is an indication that I am not making my corrections in the most correct fashion.

 

[Chris McKechnie]

Here's a quick grade I did for ya...just used Color Finale and FCPX. Not a colorist, but I think it's pretty close.

 

[Scott Crawley]

I had an attempt or two looking similar to yours. I found that in the MX shot less yellow in the reds and less blue in the whites were an improvements. Try it?

Your Dragon shot looks good. The lighter skin tones on the back of the hand look better than mine. Mine seemed a touch more yellow. It also showed in my pasta, which looked a touch orange to me.

Here's a quick grade I did for ya...just used Color Finale and FCPX. Not a colorist, but I think it's pretty close.

 

[Hrvoje Simic]

 

[Marc Wielage]

It seems pretty clear that his chart was reflecting colors from a variety of contaminants around him: the red wall, cars in the alley (One of which we can see.), and probably another wall. If those contaminants were not present, I'm sure Resolve would do a better job.

I have had a failure rate of about 35-40% with the new tool and those shots were mostly acquired indoors where the set would have been better controlled than his alley. 60-65% of the time it either got me very close, or brought me to a reasonably good starting point.

Possibly, but: it works absolutely every time with a skilled human being. My point is that these automatic tools are not necessary if you know what you're doing. I'm still a huge, huge fan of at least shooting color charts, particularly on A&B camera shoots, and absolutely when the cameras are different models or different vintages. (And I really like DSC's charts and the new XRite ColorChecker Video chart.)

I'll typically come up with 6 or 7 different "starting points" built by hand as preset looks, then drop them in depending on lighting, exposure, and location conditions. For example, a Day Exterior might get "Look #1." A night interior might get "Look #7." Where things go off the rails is mixed lighting or situations where the exposure changes, like a handheld shot that goes from an F11 exterior and then walks inside to an F2.2 living room. No chart will save you on this. Craftsmanship and skill will do it. And in thousands of cases, I did it by eye and by scopes without any charts. You get a feel for it after a while.

For the record, I think RedCine-X Pro is a fantastic tool for dailies and for on-set viewing. But a terrible tool to actually try to correct images for even a rough final. The need for windows, secondaries, and keyframes can't be underestimated. I've had many, many cases where we had to use everything we had to try to salvage dailies so that the DP wouldn't get fired from a job. If the dailies look great, if the editor has a great temp correction to use for the edit, everybody is happy. And they can use that as a guide for the final. Even for a cooking show.

 

[Scot Yount]

Marc,
Do you have a suggestion for an entry level monitor for grading? Also, what do you recommend in the same scenario for scopes and a control surface?

 

[Marc Wielage]

Their ain't much, unfortunately. True grading monitors are neither cheap nor affordable. Some users like the Eizo CG-series, but it depends on which specific model. I think the Flanders/FSI monitors are surprisingly good, but they do suffer the usual problems from LCD when you're a little off to one side. Head-on, they can be fine. The prices basically start at about $2K and go up from there.

A lot depends on how well you can calibrate the monitor, and that's a whole issue unto itself.

Control surfaces are a lot simpler. When it comes to Resolve, you can get the Tangent Wave (which I have used and don't like), the Tangent Elements (which is pretty good but not cheap), and the Avid MC Color. I personally use the MC Color, but it's limited and has its quirks. I combine it with an XKeys-80 keypad off to one side, which helps manage a lot of the menus and modes within Resolve, and I also reprogram the user keyboard within Resolve for a few things that aren't normally accessible. Some people like the J.L. Cooper Eclipse, but it was about $5K when it first came out and it's kind of fallen out of favor. The little Chinese-built OxygenTek panel is very basic, but it will at least give you access to the basic Lift/Gamma/Gain trackballs and is very good for the money ($350). Tangent is about to introduce a similar panel in the same price range.

The full-size daVinci Panels are excellent, and I used them for many years, but they're $30K, which is not too affordable.

This is all strictly my opinion. Choosing these is like buying a car or buying clothes or lenses -- it's a fairly subjective area.

 

[Hrvoje Simic]

Control surfaces are a lot simpler. When it comes to Resolve, you can get the Tangent Wave (which I have used and don't like), the Tangent Elements (which is pretty good but not cheap), and the Avid MC Color. I personally use the MC Color, but it's limited and has its quirks. I combine it with an XKeys-80 keypad off to one side, which helps manage a lot of the menus and modes within Resolve, and I also reprogram the user keyboard within Resolve for a few things that aren't normally accessible. Some people like the J.L. Cooper Eclipse, but it was about $5K when it first came out and it's kind of fallen out of favor. The little Chinese-built OxygenTek panel is very basic, but it will at least give you access to the basic Lift/Gamma/Gain trackballs and is very good for the money ($350). Tangent is about to introduce a similar panel in the same price range.

The full-size daVinci Panels are excellent, and I used them for many years, but they're $30K, which is not too affordable.

Yeah, MC Color feels nice, too bad for the options and material. At least someone in Euphonix cared about the visual appearance of a tool for folks caring for visuals. Hoping someone makes something in between this and full size panel. As in 30-40 % larger and more capable.
I have somewhere the design I made years back. $30K is a waste of money for a controller if one is not chasing high budget project deadlines daily. Not a fan of Element design and ergonomics. It can be done much better.

Current grading surface offer is like having to choose between a Smart, Jetta and a pick-up truck.

 

[Scot Yount]

Their ain't much, unfortunately. True grading monitors are neither cheap nor affordable. Some users like the Eizo CG-series, but it depends on which specific model. I think the Flanders/FSI monitors are surprisingly good, but they do suffer the usual problems from LCD when you're a little off to one side. Head-on, they can be fine. The prices basically start at about $2K and go up from there.

A lot depends on how well you can calibrate the monitor, and that's a whole issue unto itself.

Control surfaces are a lot simpler. When it comes to Resolve, you can get the Tangent Wave (which I have used and don't like), the Tangent Elements (which is pretty good but not cheap), and the Avid MC Color. I personally use the MC Color, but it's limited and has its quirks. I combine it with an XKeys-80 keypad off to one side, which helps manage a lot of the menus and modes within Resolve, and I also reprogram the user keyboard within Resolve for a few things that aren't normally accessible. Some people like the J.L. Cooper Eclipse, but it was about $5K when it first came out and it's kind of fallen out of favor. The little Chinese-built OxygenTek panel is very basic, but it will at least give you access to the basic Lift/Gamma/Gain trackballs and is very good for the money ($350). Tangent is about to introduce a similar panel in the same price range.

The full-size daVinci Panels are excellent, and I used them for many years, but they're $30K, which is not too affordable.

This is all strictly my opinion. Choosing these is like buying a car or buying clothes or lenses -- it's a fairly subjective area.

Marc,
Thanks for taking the time to write this up and reply. Much appreciated.