Accurate Colour & DSCLabs chart

I've been trying to create a LUT for the BMCC that accurately represents colour using the same technique described by Art Adams (link below):

LUTs (created with Zeiss ZF.2 100mm prime): https://www.dropbox.com/sh/1419xua7dpqjxop/oaouPvWd5R
Art Adams' instructions: http://provideocoalition.com/aadams/sto ... s_oneshot/

I’d appreciate opinions on this approach. I’m not trying to create a ‘look', but an accurate starting point. I hope this correction won't limit the scope for creating a look later, though I guess it would affect the use of LUTs used down the line such as Tom Majerski’s?? I also wonder why the colours are represented as they are by the BMCC.

while it would not be a good idea to apply this LUT and the one I recently released,

it would - if accurate - have a good usage as part of creating another lut or grade, as the first node I always create when working on a grade is one which adds contrast and saturation from the flat footage, so if this one works well - it will be very useful as a standardised part of the process.

I will have a look at it later on, thanks!

Cheers Tom

My LUTs just make changes in Hue Vs Hue (to align the colours to their corresponding vectors on the vectorscope) and Hue Vs Sat (to give even saturation of each of the colours on the DSCLabs chart), that's it, it's v simple.

By virtue of these changes, it increases contrast and saturation.

My hope was that I could apply this to all my footage, but unless shot with the same lens, it probably won't correct colours 100% accurately.

I may create a Hue Vs Hue and Hue Vs Sat correction LUT for each of my lenses or stick the chart to the clapper board for every scene.

Jim,

I may be way off but...

I think you are approaching this exactly right if your goal is accurate spectral reproduction to start grading. It is similar (I think) to using ACES camera profiles and ACES color space.

Every camera records color differently, and is graded differently.
http://www.fxguide.com/featured/the-art ... tal-color/

You are "equalizing" the BMC sensor to be "flat" if you were to think about this in audio terms. If you have speakers that have too much bass, then an eq in your master output can correct for this.

I also believe that any "looks" or presets that you add downstream will actually be better by using this initial step.

I can't imagine a situation that starting from the most accurate color would not help. Even a very stylized look will be easier to achieve with a balanced spectrum to begin with.

The only obstacle that I see is compensating for differing lighting sources. Every light source has a spectrum that is very different and will effect the image. You may need different LUT's for each of your typical lighting setups.

The same is probably true for filtering, as the image and color balance will be altered as your filter stack changes.

What I end up doing most of the time, chart every shot and correct for every shot. This is one of the ways that I've found helps to mix multiple cameras in the same edit.

Just my random thoughts, I may be wrong in many of my opinions.

Dave

Thanks Dave, that's really helpful. I was thinking if I accurately white balanced every scene and applied the LUT/correction that could work, but of course what you say about the spectrum of light from different sources means it's more complicated than that. I will use the chart on every scene. It's a tiny chart...that may pose some difficulties in practice using prime lenses.

(The scene for the correction was shoot in overcast day light and there was a speedbooster between the lens and the camera body).

I will use the chart on every scene. It's a tiny chart...that may pose some difficulties in practice using prime lenses.

Yep. While Custom WB or even AWB might be convenient and even usable sometimes, BM's minimalism forces us to be more involved in this critical step.

There is a larger version (about A4, 11" x 9.5") available here: http://www.dsclabs.com/front_box_series.htm

More un-researched and un-verified babble...

I think that the camera profiles that are used for the ACES system are determined on the sensor level. Or put another way, are independent of the lensing, lighting or filters.

So, they're may be a way to setup a LUT that "corrects" for the sensor regardless of the shooting environment.

It may be as simple as averaging out the results of corrections to many shots, or one correction to an "ideal" environment.

Maybe a back lit ChromaDuMonde chart with the spectral characteristics of a lens subtracted would work.
http://www.dsclabs.com/chromadumonde.htm

I used to do a ton of astrophotography and we used to do many, many corrective steps to our data before we ever worked on the aesthetics of an image.

We corrected for sensor noise, sensor defects, optical path distortion, vignettes and more all before the R, G, B and Luma channels were combined to make a color image.

Our goals were not only aesthetic, scientific accuracy was also part of the desired outcome.

But, I may be missing much of the big picture.

It seems like what you're trying to do is essentially what aces is. A way to take any raw or log camera footage and standardize it. Try using the IDT and ODT drop downs in resolve to see if it does what your looking for.

Thanks streetlightmedia and Dave I'll look into ACES

After finishing the feature I was shooting (The Renata Road) I got round to testing the ACES profile for BMCC and found my chart was not displayed with anything like accurate colour representation. I conclude that the only way to do it is to create a LUT for my BMCC and each lens I use.