Color Checker Chart for Premier Pro Plugin WOW!

[Offbeat Bryce Productions]

I don't know if anyone else here has seen this. MBR Color Corrector 2  It's a color checker chart plugin for Premiere Pro CC 2018 and lower. Just stumbled upon it. I tried the free trial and it works. You can even change the colors on the chart if you have a different one. This is going to be easier than having to export an XML file and import into Resolve.

Free version also only supports 8 bits per channel but I'm not sure what that means when Premiere automatically up converts to 32 bit.

But I also think buying this is worth it. 

[chrisw44157881]

good find!

[KiX]

Nice! I have been looking into something that does exactly that! was kinda jealous that Resolve can do that and premiere couldn't!

[mnewxcv]

this is great, I have been looking for this functionality for a long time. I am trying the free version since the links to buy the full version are not working right now.

[jamesdrakefilms]

awesome thank you!

[normanm33845938]

Five years later I read this post. Thank you so much. Why isn't it a feature of Premiere. (Like Davinci Resolve?)

[R Neil Haugen]

Most colorists don't use the auto-chart process in Resolve because it makes a number of typical errors. Mis-reads the chart, or reads a pixel or two of the edge, something like that.

I do this manually, and if the Pr devs could get a fully functional chip reader, I'd be thrilled.

Neil

[Guest]

Just found out about MBR Color Corrector 3.

Any experience? 

I am using lots of Sony-Cameras with 8bit cine2 color profile and slog3/cinegamut 10 bit (Fx3).

Can you use it to apply LUTs as well? 

Does it blow out your highlights or does it keep them?

Do you add your lumetri-correction before or after the effect?

Any change in playback performance with the effect applied? Stability?

I dont have a color checker yet.

As a solo shooter I need something tough and small that includes white for a quick WB.

So I guess colorchecker passport video 2 is my best bet.

Any experience? Version 2 isnt listed on the website of the plugin, is it supported?

https://calibrite.com/de/product/colorchecker-passport-video-2/

[R Neil Haugen]

Never heard of it before, was interesting to go read about. This is one of those processes that, after you do it a couple times, takes no more time really than using that tool would. 

Especially if you're working your own shot media ... as each camera's hue response will be rather consistent. Once you've got that camera's response corrected, you save that as a preset, drag/drop it onto all clips from that camera.

And when using a Passport vid-checker say, and of course, with saved presets per camera, once you've done a quick WB/shadow-cast and tonal adjustment per clip, you're pretty much done.

I've tried a couple things like unto the MBR plugin. I tended to get artifacts from them in Premiere, and in Resolve also. They tend to set the color chips out to the box near the limits of the Vectorscope, and I think that's too far. I never go past 1/2 way out.

You don't repeat that hue thing every time unless the lighting was way off and you really need to.

[Guest]

Just tried out the trial version 3 and it seems to work fine. Even the colour conversion works FAR better the the auto tone map and conversion in premiere. Colours still look natural, even without any chart correction applied. The only flaw: the effect slows down your machine massively. Even my MBP Max M2 with 32 GB slows down when scrubbing over HD Prores PROXIES when the effect is applied. So I will have to workout a good workflow to apply this correction in the end. I was hoping I could just create a sequence from my whole raw footage and apply the efect there from the color chart and then just edit the clips into my editing timeline. So I could have already consinstent colors that I just tweak a little in the end. Now I have to go into the final sequence and find the corresponding chart, apply the effects there... takes more time. OR is there a way to apply the effect and batch enable/disable only THIS effect temporarily?

[R Neil Haugen]

That's where my approach is actually faster. As for each camera, it's mostly a do it once thing, make a preset for each camera. Then just dump on all clips from that camera. After that is applied, at most, adjust the WB and maybe a couple tonal things, done.

And one Lumetri preset doesn't slow down Premiere at all.

I don't like wasting time, personally ...

[Guest]

Maybe you are right. I still dont know how much different lighting conditions affect color rendition of the chart and so I should just shoot the chart again to save some time. I will try the full version of the plugin and will see how good it performs - as I personally dont like the lumetri effect very much..

[R Neil Haugen]

The differences in color response are pretty much native to that specific camera/lens combination. Some lenses can make a barely noticeable shift, many ... not enough to bother with as you can't really tell for sure.

So changing the lighting changes exposure points and WB and sometimes shadow cast. But doesn't do anything to the basic color response of the camera.

It's why colorists are now so concerened with getting specific input transforms per camera. Which you can do in Resolve. But in Premiere, we would need to use chip charts for each camera as the "input transform".

I have used this a bunch, and been surprised that for most cameras the WB at time of shooting is irrelevant to the camera response. In other words, doesn't matter if the scene was lit 3100K or 5000k, same chip chart shot works.

One prosumer body did go a bit different between 'tungsten' and 'daylight' response, though still not too much.

[Guest]

I guess you are right.. as long as its just the WB. If there is any color cast by artificial lights for example you have to work more.. So its not much more work when I have to perform a new WB to just film the chart for two seconds. If matching is done with one click, I dont mind doing it a few times more in a documentary - and i think I will even save time in the long run. 

[R Neil Haugen]

I know which cameras are used for most of what I work with. I've "charted" them. I just bulk apply the camera chart preset to the camera's clips in the bin.

Then for most of what I do, heavily inteviews or short scenes of scenery or interior b-roll, all that I do additionally is check the WB and shadow neutrals. The basic hue (color) tonal responses are mapped decently close by the charted preset.

This means that most of my work doesn't even really need any "shotmatching" work. It's close enough just a quick touch of one or two controls moving down the sequence takes the 'visual bumps' down to not enough to worry about.

And ... I'm pretty darn sensitive to, and picky about, color/tonal things ...  😉

[Guest]

You seem very convinced about your workflow and I wouldnt want to ague its the quickest way. 😉 I am still not sure if the plugin can give some more accurate results as it corrects ALL color patches available on a color chart, and on some charts the have lots of patches especially for skin tones. So I would think that matching just the six primaries on a vectorscope could cause some issues in the skin tones when camera colours are shifted..

[R Neil Haugen]

With a chip chart and Hue v Hue, you aren't making global changes. Just moving specific hues a few degrees one side or the other, getting a more clean color response. Aligning each part better. But you're only moving that hue, and just a few degrees either side.

So you can't screw up some tones by moving another. Just doesn't work that way. Not at all like say grabbing the center of the color wheel and pushing.

I've spent a truly MASSIVE number of hours with my Elements panel, in both Premiere and Resolve over the last decade.

I've been taught by some of the best, and I teach this now to pro colorists.

Resolve has enough different and tightly targeted controls you can shotmatch full-on if you want. It's what we always used to do.

But if you can find or create an IDT or OETF equivalent, that adjusts the files from a camera to more match your goal, why would you still choose to fight every clip?

Resolve gives you a ton of camera/media IDTs to try. Premiere doesn't.

This is the equivalent. And I do mean, for most uses, make once, reuse over and over.

You'll trim WB and some if the tonal shape maybe. But have a vastly more unified timeline quickly.

[Guest]

I dont question your authority on this field 😉

It can just take very long sometimes to match skin tone so I would be glad it the tool can help me there. 

So with a few additional fields maybe most COLOUR corrections is already done and I can just concentrate on exposure.

[R Neil Haugen]

I think you're assuming I mean to do the full routine on the chip charts for every shot. Every time.

Not at frickin all!

Do it once good for any camera, done. Reapply, don't redo.

[Guest]

No, I meant, if I correct with the plugin and skin-tone patches, I dont have to match skin tones later on. Matching color primaries is no guarantee skin tones will match between cameras.

[R Neil Haugen]

Once you've mapped the hue response of the camera, applying that to the camera must also align skin tones relative to WB. That's what this does, align all hues closer.

The camera has a built-in response pattern between sensor filters (physical properties) and the mathematical process used to manufacture color from luminance data cross matched across the filter pattern as it's debayered.

That cannot change. The chip chart process I suggest recaps the built-in response to what you prefer. Hence it’s no longer a "variable".

[Guest]

If you calibrate a display for example I think theres a reason why it samples and scans hundreds of patches..

[R Neil Haugen]

Certainly ... if you're trying to maintain an absolute measurement value from dark through light, the entire scale, right? That's what you need for matching an absolute standard.

Are you checking the skin tone at 15 different levels from near black to near white? I somehow ... doubt that. Because if your monitor is correctly calibrated, you shouldn't need to.

You never ever do the type of changes of hues for say skin tones that are required for monitor calibration. Do not conflate the two.

Your eyes ... and mine ... are nice relative tools but absolute fails on aboslute measurements. You're expecting that you need an absolute tone ... which has been proven over and over again not to exist.  In any practical reality.

For example, requiring that screen values for trademarked colors match precisely when they hold the can of pop to the screen, which has been done many times.

Think about that ... first, no other screen will ever show exactly what that screen showed to begin with! Screens vary immensely between them, even "identical" models calibrated by the same gear on the same computer system will not be visually identical.

Next, we are horrid at absolute values. Tests have been done showing people specific hues. Then they are to pick the matching one from a set of close images. They never ever actually choose the matching image. All humans tested this way pick a color closer to the 'center' of the wheel ... in other words, less specifically differentiated and saturated, in value! ... and a bit darker in tone ... than the one they were shown.

We have an expectation for things like grass, bricks, sky, and skintones. But our actual preferences do not at all align with the actual measured color. Grass isn't actually nearly as green as we think. Sky isn't blue, it's way cyan. And  the full range of  skin tones vary dramatically between people within the same ethnic background.

To show what people "think" is correct, is to understand the above, and give them what they "want", not reality! An early lesson for colorists to learn, understand, and use.

Yea, I've been through hours and hours of training, discussions, and of course, I've got a huge stack of books on color matching from the days of film  on through all current digital processes. I had my own full wetlab for color negative materials for over 20 years, btw ... I was doing color tests and calibration processes carefully read on densitometers and plotted on curves on paper for YEARS. Long before digital imaging came into use.

We were one of the fist professional photography businesses in our state to go fully digital, back in 2003. A year or two before we should have, in hind sight. But we survived.

And the last decade plus, I've been concentrated in working with and for pro colorists. Totally. As noted, I teach pro colorists. For pay.

Back to your issues ...  match the upper midtone hue response pattern of any camera, even with only the one set of six color and maybe 4-5 tonal ones. Do it once. Reapply as needed. Match WB and tonal curves between cameras.

And in practical terms, your skin tones work will look great. Relative to what people expect and normally get, on any screen.

And Dado, like many other colorists, are people I know and have talked with. I could give chapter and verse about why some colorists think Dado's colorlab is useful, and some who are more than a bit skeptical. I've talked with Dado about it, he's demonstrated the whole thing to me in person. We had a great time, over an hour together on it. I find it intriguing in concept. But I've not used it personally.

And yes, I can talk with Dado and Cullen and MTO and other colorists going down rabbit holes on color science and know what they're talking about. Because I've spent hours upon hours in color rabbit holes. For YEARS. 

For me, it's all fun stuff. 

[Guest]

I was wrong: Playback performance seems to be fine. Just applied the effect to 1,5 hour timeline with Prores HD proxies. So I think for me its the easiest way to just shoot the chart when I change WB, apply the effect to the timeline part and ready to edit. In the end theres only very subtle finetuning needed.

[R Neil Haugen]

There's a local video producer that I've talked with. Does mostly corporate website 'splainers and socials. "The two S's. " is how he puts it.

He's got to run a tight shop, buying no extra gear, and spending not one minute extra time, to make a decent living, right?

So he'd heard me mention that I'd read, that many lower cost bi-temp LED light units are less 'spiky', more even color, on their 'tungsten' 3200K LEDs than on their 'daylight' 5500k LEDs. So he tested in the studio.  Set the lights, did a WB in-cam, shot the chip charts. Yea, his chip charts under T lighting needed less adjustment to neutralize hues.

So he changed his setup for interviews, and for "office-pass shot" lighting, which are all set to camera and light tungsten settings.

Then he thought ... he likes simple, working without having to think. Changing camera setup from Tungsten to Daylight ... too much chance for him to forget.

And he still had his Tiffen filters from his first few years, shooting film. So he left the camera set to tungsten, went outside, slapped an 85B on the lens, did a WB and shot a chip chart.

It pretty closely matched the Tungsten lighting chip chart.

So ... his 3 Canon camcorders got all set to Tungsten and left there. His lights are all set for max Tungsten output. And he uses the 85 series filters 'outside'. 

He shifted to BlackMagic pocket 4ks a year or two back, does the same thing. Spends almost no time in post with color. And I must admit, his stuff looks pretty solid. He does always do a WB with either the Kodak Grey Card white side, in studio, or the Xrite color checker white card out of studio. And shoots the card while doing so, so he can check his WB quickly in post.

And that's about all the change he ever makes ... a quick adjustment of WB from the card, applied to the following footage, bang. Done.

Oh ... and each camera has a bit of colored gaffer's tape on it. ALL media cards for that camera have a patch of same colored paint on them. So he knows without having to check on uploading which camera made the shots. Puts them in folders by camera. Each camera has it's own WB shot every setup. And each camera has its own Lumetri preset for hues, as I've listed before.

Saves him a ton of time, and he hates doing color work with a purple passion. But his stuff ... is pretty well done.