Is Color Timing/Grading Scientifically Accurate?

[Jimmy Pierce]

First off, I am not a colorist. I think I do an okay job keeping skin tone natural and not crushing shadows too much while still achieving a look and have done so for several years. However, I don't know how to read a histogram, vector-scope or waveform monitor, so I have no idea whether what I'm doing to the image is scientifically accurate, especially when the "look" sometimes doesn't transfer over to other mediums like YouTube or even my consumer grade TV.

I will say it's much easier to breath life into an image that has low contract and saturation.

So that raises the question, is color grading scientifically accurate? When you watch an action movie where blue is the predominant color, or you see a sunset that looks like a bomb went off in the air and the sky caught on fire. Or when someone's skin looks pinkish, or yellowish, or reddish, or greenish or....you get the idea. They would be an alien from outer space or somebody with a skin disease if you saw them walking down the street, but in the context of the film, it's fine.

But does it pass the vector-scope, histogram or waveform monitor exam? Even the bleach contrast look, how come it transfers so well to different mediums and doesn't get lost in the shadows and what does that look like on a color sensitive monitor?

 

[morgan_moore]

Id define grading as moving the image away from the scientifically accurate for artistic purpose.

Id say the colour process is 'normalisation' where one would make the image scientifically accurate, so for example grey would read at 100,100,100 with a colour checker.

All the colours on a gregtag card have values that can be alligned with numeric RGB values.

Once the image is normalised it would be graded.

The grade may render grey at 90,110,90 for example (a bit green) because the scene is in a morgue and the director wants it to look green and spooky.

Normalisation. Grading.

 

[Samuel H]

Many people call those color correction and color grading, though color normalization and color timing and other terms are used too. In any case, that's just how most people do it: first make every shot look correct, and make them all look similar, then apply a look on top of that. Secondary corrections (with masks and/or color keys) can happen both in the first and the last stages.

 

[morgan_moore]

..so "Grading" is moving the image away from scientific accuracy.

 

[David W. Jones]

I'll add legalization to the list. Keeping the video within it's intended color space for delivery.

 

[Josh Bass]

do you guys think that at the lower end of the spectrum, where you may only grade/color correct once even though distributing to different media (dvd/fest screenings/computer viewing/etc), its better to stay within ntsc (or pal/whatever) legal specs with colors/luminance to give things the best chance to look right on most screens?

 

[David W. Jones]

Short answer... Depends!

I will say I have seen my fair share of film fest movies which I couldn't sit through, due to entrants not understanding DCP XYZ colorspace workflow.

 

[Jimmy Pierce]

That must make the post process interesting if color space from one medium doesn't transfer to another medium. How much adjusting has to be done to the "grade" to achieve the same look across the board. Would you need to adjust the color space for a DCP differently than Blu-Ray? Is greyish green just going to look grey on one of those formats if you don't?

 

[morgan_moore]

You should be grading on the delivery spec.. Or learn aces which is a wide space designed to be output to many smaller spaces.. I've not really studied it but I think that is the jist of aces

 

[j1clark@ucsd.edu]

That must make the post process interesting if color space from one medium doesn't transfer to another medium. How much adjusting has to be done to the "grade" to achieve the same look across the board. Would you need to adjust the color space for a DCP differently than Blu-Ray? Is greyish green just going to look grey on one of those formats if you don't?

Bluray conforms to the Rec. 709 color space, which is different than the DCP XYZ space.

Here's an illustration showing the various gamuts...(got it from a thread that apparently is touting the Sony F65...).

 

[j1clark@ucsd.edu]

So that raises the question, is color grading scientifically accurate?

Color grading uses certain results of science, but for the most part most films are not produced for their forensic truth value.

Even when one is using imaging for science, very often color will be transformed, into 'false' or 'pseudo' colors (pseudo color has a specific meaning under the class of 'false' color methods...).

In any case, the process of color work involves removing unwanted color bias, aka white balance, or color cast removal, color noise, blue channel issues, etc. and of course there are creative color processes, which introduce, or augment, all these things in the final image look... but under the control of the colorist.

 

[David Jimerson]

Because it came up and wasn't really answered, a basic conceptual distinction:

Color correction: fixing problems.

Color grading: adjusting color for artistic reasons.

Color timing: literally the time which elapses while developing color film; different times give different color grades, so if you time it a certain way, you can get a certain grade.

There are other terms which refer to the same things, and a lot of it can overlap, but those are the basic concepts.

And it's good that this is in the cinematography section, because it's not always done in post. You can grade your footage on set, in camera, too. We have a little bit on this in Lighting for Film and Television.

 

[Jimmy Pierce]

Bluray conforms to the Rec. 709 color space, which is different than the DCP XYZ space.

Here's an illustration showing the various gamuts...(got it from a thread that apparently is touting the Sony F65...).

Is this illustration the gamut of one color or several? Does the yellow circled area represent the "range" of color produced by the F65? I'll do some research too because I'm sure there is more to that explanation than and I should probably understand a bit more to understand the explanation, haha. Education is fun. :2vrolijk_08:


I posted this in the cinematography section because I see timing/grading as more like post-cinematography than post-production. Editors are not colorists and in shooting film and the process of color timing, my understanding is the DP & director is part of that process.

I see so many attempts to "grade" footage, it's almost become an arts and crafts exhibit, and I often wonder to what extend does this look serves a purpose other than propelling the story and evoking emotion.

So how about contrast and saturation?

For example

 

[Aharon Rothschild]

Grading is a natural extension of cinematography. The DP crafts a look with sensor or film, lenses, lighting, wardrobe, and set. This is really what separates documentary filmmaking from narrative filmmaking. A film's look can be considered successful to the extent that is sells the story. See the Hobit in high frame rate as a negative example of using a high reality format for a fantasy story vs Hugo which uses a very purple mood color to successfully change the mood of the film.
A narrative colorist will tell you that the only accurate and acceptable standard is the director's mind image for the film. In that sense waveforms vectorsvope etc. are used as a way to keep shots consistent within a scene and as a very useful way to maintain the look of a film over what may be a one to three week coloring process.
Where accuracy to real world is actually a issue is with advertising when the object shown must match to a specific hue. The way to maintain that accuracy is to shoot a DSC labs color chart that has targets that will line up with targets in the vectorsvope when accurate. Some will use the macbeth or other charts where there are color chips with know values that can then be matched to the colors on the chart either with software provided by macbeth or in davinci resolve. These charts are intended to profile a particular camera and lens for color still photography and have issues (as all things do) when applied to cinematography. See this excellent article by Art Adams http://www.provideocoalition.com/what-good-is-a-macbeth-colorchecker-chart