Torn on High vs. Low Contrast and Usefulness of LUTs

Background: I shoot with a Blackmagic Pocket and have only ever used the FCPX color board for color correction. I have Slice X so I can track secondary corrections inside FCPX and have never learned Resolve because I don’t like the idea of round-tripping.

I downloaded the Color Finale trial to see what the “curves” fuss is about. Having played with curves, I like the additional control, and the end result is that I can achieve a higher contrast than using the FCPX Color Board. But I am torn on whether or not that is a good thing. All the S-curve is doing is compressing the shadows and highlights. Of course I get to decide how much compression I want by manipulating the curve (or the opacity slider). But there will always be more compression on any curve compared to leaving the curve as a straight line. So the image gets more “Pop” at the expense of losing detail in the shadows and highlights.

I also tried the included Blackmagic Film to Rec 709 LUT in Color Finale, which turned out to be the nearly the same as applying the native FCPX Blackmagic Film conversion to a clip under the “Info” panel. The result is a highly saturated, high-contrast, clipped image. LUTs are billed as being a “time saver” but the only way they save time I think is if you expose to the center of the histogram. If you expose to the right (as you should), the LUT just horribly clips the image and you end up wasting time by trying to fix the resulting image or be removing the LUT altogether. And yet from what I gather, both curves and LUTs are very popular.

So let me pose this very basic question:

Why are we buying expensive cameras with high Dynamic Range only to then spend considerable time in post compressing the DR to make the image look like it was shot in a DSLR? Would it not be better to just shoot with a DSLR? It would sure be quicker and cheaper. Yes you get more control on how the final image will look but if the standard for a good image is contrast or “Pop” why bother with high DR cameras?

LUTs should only be used when you shoot in Film mode, with a very flat image. If you do contrast and color adjustments in the camera itself, then don't use LUTs, just use a color grading tool. Very simple.

Shooting in a film mode, flat, you get the most detail in both highlights and shadows, more than having baked-in contrast. DSLR's and your Pocket are all on the same level in this respect. You're shooting in high dynamic range, or you're not. It is not dependent on your camera. The more filmic cams like RED and BMD simply allow for that technique to be taken further than "some" (not all) DSLRs.

Curves do give more control than a simple 3-way exposure/contrast tool (FCPX color board). Both do the same thing, just one is more detailed than the other.

I recommend doing the color grading tutorials at colorcentral.com to understand the technicalities in more detail.

Sadly, the unfinished flat log footage has become trendy a trendy "look". Just like shaky hand held DSLR footage with super shallow depth of field where the subject goes in and out of focus all the time. People seem to think it's a good thing.

If you are working with log footage, it's supposed to be graded. LUTs are just a short cut to that. Problem is that they are dependant on what comes in and as you say, exposure must be as intended for the LUT. Most big productions create their own, tailor made LUTsfor their work flow. Obviously LUTs need a consistant exposureto work best. It's not uncommon to have variations of the same LUT. For example, LUTs made to compensate also for overexposed footage. So if you decide to always over expose by 2 stops, you create a LUT that compensates for that.

I'm also using the BMPCC, shooting in ProRes Film and grading in FCPX and use Color Finale.
Yes, if you ETTR, the LUTs will generally look horrible since they're based on proper exposure.
Some people use LUTs just to have something that looks "normal" while editing before they do the Grade.
It's not a quick solution if you ETTR.
First you must go through and set exposure for each shot. If you ETTR for each shot than you do have to do this one shot at a time.
LUTs are a quick way to create a "look" once the exposure is correct. Then you tweak from there if you want.
If you have the time and the skill, there's nothing wrong in doing it all from scratch. Or, getting to a point where the LUT is added as a final tweak rather than a jumping off point or "quick fix" (which it won't be if you ETTR as you've discovered).

Curves can allow for some very subtle control in contrast. The common first step is the S cure contrast you're probably referring to. There's much more though. Sometimes you want to bring out (or bury) something in a very narrow range. Sometimes you're using it to adjust colors to create a variant on the "blockbuster" (orange/teal) look for example.

There's a big difference between the BMPCC and a DSLR If (a big if) you're doing a lot of grading. 10 bit 4:2:2 (ProRes) allows for a lot of pushing around without banding or noise being introduced. You can push even further if you use cDNG RAW (12 bit 4:4:4) but unfortunately FCPX doesn't support that (I wish it did).

A couple of things:

• Blackmagic Pocket is a high dynamic range video camera. It produces video that is closer to an Alexa than to a Cannon or Panasonic DSLR.
• I always shoot in film mode, I thought that was implied. I shoot ProRes though, so I do need to select the white balance. If FCPX supported RAW directly from the camera I would occasionally shoot RAW.
• I have done the tutorials and am experienced at color grading.
• We are seeing different trends. But yes, I agree flat footage in a finished video is bad.

I agree that curves are not only about the S shape and I’ve played with many other shapes. I have not needed to bring out or bury something in a narrow range, but what cseeman is saying makes perfect sense. However in doing so you would compress the detail in other areas, but I guess that just doesn’t matter as long as we are placing the emphasis what is most important to the “story of the image”. I like the additional control of the curves, but since their main purpose is to add contrast at the expense of detail in other areas, they have made me question why not just shoot with a DSLR? I guess more control will always be the answer.

Have there been any rumors of Apple having plans for curves natively in FCPX? I’m torn on whether or not to purchase Color Finale only for the curves. It would suck if I purchase it only to have Apple add curves to the app a few months down the road. I’m not a fan of the Color Finale floating window and after using the color wheels I don’t see why all the fuss about them being better than the color board. Rectangle vs. circle, meh. Pushing opposite color vs. pulling a color, meh. At the moment I prefer the color board simply because I can move the pucks with the keyboard. Also FCPX has 4 saturation sliders while Color Finale has only 1. Not that I care much, but it was unexpected to discover that the highly praised color wheels allow less control than the Color Board.

For secondary corrections I’m not completely sold on Vector Tool in Color Finale. I need to experiment with it more, but I think most of the time a shape mask will give better results (admittedly they take a lot longer to set up). But those curves… I think I’m going to miss them when my Color Finale trial expires. Apple please give us curves inside the inspector!

As to shooting with a DSLR, if you can get the control with picture profiles in camera so you need very little grading that would most certainly work. Consider that the AJA Cion doesn't really have a flat/film profile and Sony cameras have lots of picture controls as well. The biggest problem with DSLRs is the fragility of the codec (unless you're using an external recorder). If you have to make changes you can be severely limited.

Sometimes I think LUTs, for many people, are simply picture profiles except added in post. There's a lot more potential with a good recording codec.

BTW I agree about wheels vs "the board." i think it's just a matter of taste and convenience.
Curves are the real power though if you want that kind of fine control.
I do wish Color Finale had individual saturation control for each Wheel as well. I believe I gave Denver Riddle that feedback.

The strength of the Vector Tool is the ability to impact colors throughout. Perhaps flesh tones of several actors or maybe a background that would not otherwise lend itself to easy masking. Specifically they're an alternative to HSL type masking. The problem I find is that while most HSL masking tools allow you to see the mask areas, the Vector Tool does not. I'd love to hear more about his preference for Vector Tool over HSL masking.


Regarding the floating window, this is all about Apple's FCPX GUI constraints. As you might not, many other color plugins overall the image and many find that disconcerting. Since there's no other way to alter the interface by adding a tab or panel, you're left with a floating window which, at least, you can move out of the way.

Riddle said there's going to be an update to Color Finale this week. He's adding the ability to balance to an XRite color chart much as one can do in Resolve.

El Cuajinais wrote: Why are we buying expensive cameras with high Dynamic Range only to then spend considerable time in post compressing the DR to make the image look like it was shot in a DSLR? Would it not be better to just shoot with a DSLR? It would sure be quicker and cheaper. Yes you get more control on how the final image will look but if the standard for a good image is contrast or “Pop” why bother with high DR cameras?

Regarding your question, it's mistaken to think that there's a "standard for a good image." The point of shooting in a flat profile that maximizes your dynamic range is so you can have flexibility and finer control over your look in post. A lot of times you might want a muted, desaturated, low-contrast look. Or maybe something in between. Or really, whatever you want it to look like.

Also, I'm pretty sure most people don't want their image to "look like it was shot in a DSLR." Not sure what you mean by this. DSLRs became popular years ago because it was a cheaper way of getting a good video image using decent lenses, images that more closely approximated film cameras and higher cost HD cameras.

High dynamic range better helps to ensure you don't clip highlights or have crushed blacks while shooting.
What you do with that detail in post depends on your aesthetic goals. Personally I like keeping as much detail as possible unless the scene dictates otherwise.

If one thinks crushed blacks and clipped highlights are "popular" perhaps that's because a lot of DSLR shooters are stuck with that. They can't pull out details that are lost in the initial recording.

All the more reason why curves are so important. You can isolate parts so you can do things like crush the very bottom of the blacks yet, still pull out details just above that, something you really can't do well with an 8 bit 4:2:0 codec.