General DSLR color grading theory

B

boulder

Well-known member
It was first thought that grading an image shot flat gives more lattitude...sort of emulating the RED raw concept. But these arn't RED cameras. When pushed too hard the image bands, artifacts and image degradation rear their ugly head. I don't believe any latitude is gained but someone might point to tests that prove me wrong. I no longer shoot flat and shoot closer to what I want my final image to look like (as long as it's not too extreme). The reason I bring up this dead horse is that I still see a lot of articles telling users to shoot flat and I think we are misleading people and causing more harm than good.

My question...is there a consensus on this or is there still two camps? If there is still conflicting theories it might be in everyones best interest to prove this out once and for all and make everyone understand that YES you do or NO you don't gain much or any lattitude.

Of course this will all change if Canon or some other DSLR camera maker outputs something closer to raw...but with these heavy compressed codecs I have extreme doubts.
 
Neutral or Faithful

Neutral or Faithful

I tried superflat and like you was disappointed with the grading. Now I shoot most with Faithful or Neutral and am able to make subtle grading changes in FCP/Color that work for me.
 
This is the video I believe is the root cause of some of the misinformation. I don't think this video is entirely accurate and is hardly thorough in its methodology and I think it has been accepted too quickly as the truth.

As always I may be wrong ;)

 
My 2c: so much of the admitted, graded, superflat footage I've seen looks washed out and, well, too flat, plus skin tones seem off a bit. Personal opinion, without being remotely expert on the technical side, is that it's a specific look that works for specific shots, but I don't like it for a general look. Maybe if I saw it projected and not on an LCD it would look better, but every time I see a video that's a superflat example, I have a hard time reconciling what my eyes see with the claims of proponents.

That video is one of the few examples of superflat that I thought looked halfway decent graded as a non-flat style. What's the problem with it?
 
My biggest issue is that he never really proves the point that he is actually increasing the dynamic range. He is giving examples of nice grading but not showing us that he has gained X-stops with or without degredation. I'm just saying this has been accepted as truth without anything to back it up.

Again I may be wrong and I just wanted to bring up the point to make sure we arn't telling people to do something that is counter-productive.
 
The old WYSIWYG is true for these cameras. I tried superflat once and wont do it again. It introduces more problems. Just tone down sharpness, contrast and chroma a tad.
 
Superflat and other flat settings record deeper highlight and shadow information, but they do so at the expense of mid tones. I quit shooting flat as well, especially people / faces suffered from it being dull and flat and missing detail, post-processing didn't fix that for me.

WYSIWYG is not quite true though, almost all vDSLRs record slightly more detail in highlights and shadows than they appear to have straight out of cam, you can recover them in post (nothing to recover when shooting flat though).
 
Thanks for this thread! I had been thinking about this also. I was shooting superflat, as well as the custom settings everyone suggested from the start. However, while grading I realized I am still such a novice at it that I wondered if it might not be better to try some of the other settings. I mean... if its not a looks preset I don't really know what to do, so I doubt I really need superflat. Thoughts?
 
Burritoclock said:
Thanks for this thread! I had been thinking about this also. I was shooting superflat, as well as the custom settings everyone suggested from the start. However, while grading I realized I am still such a novice at it that I wondered if it might not be better to try some of the other settings. I mean... if its not a looks preset I don't really know what to do, so I doubt I really need superflat. Thoughts?
Click to expand...

As others have stated in this thread the concept of shooting superflat in general is questionable. I shoot generally in Faithfull or a custom AGFA look and set the contrast and sharpness down a tad.

One way to start learning is to start with presets and tweak from there as I usually find presets such as Magic Bullet, NewBlu etc too drastic for my liking.
 
I think most of the examples in that video look terrible before AND after grading! I would eat my arm off before acquiring footage that looks like it's been shot through a tupperware lid.

The only convincing example is the one of the guy's head. But the reason the "STANDARD" style is blown out is (I guess) probably because contrast is at the middle notch.

Contrast is the only control that reliably adjusts the actual black and white points. It is the main determinant of dynamic range.

An inverted S curve just moves everything that is in between the black and white points. So it might appear to offer more dynamic range as it moves values that are at, let's say for the sake of argument, 250-254 down to 220-224. So you can see that fine highlight detail more easily now. Of course the image as a whole now looks dreadful.

In my testing, I found that superflat styles do not afford any more detail in the shadows or highlights that could not be recovered by applying an inverted S curve in post to footage derived from Neutral picture style with contrast at the minimum.

So there is no reason to screw your footage up in the first place. You can safely screw it up when you get home and on to your computer!

Now with that said, there is one style that I use, based on one uploaded by oli over on cinema5d. Instead of going near the gamma curve in the PSE, simply eyedropper pure black and define the largest range you can for it and then raise its luminance by 4 levels. This will lift your blacks a bit and get them out of the noisiest area. You need to pull them back down in post.]

It's attached if you want to play with it. I'd be interested to get your thoughts on this.
 

Attachments

I prefer to use Picture Styles to achieve the look I want. IMO, in many cases, color grading results in a degraded final product.
 
Hmm, I've only recently got a 7D and I've shot flat and standard, I think I may try the other picture styles, because flat is easy to completely lose detail, even though people do it for exactly the opposite reason.
 
Have said this for a while now. Shooting Neutral is plenty enough to capture highlight detail. The rest is done in front of camera, aka controlling light.

It's not RED or RAW... treating it as so seems very backward.
 
You must log in or register to reply here.
Back
Top