massive shift in green color on BMPCC4K

Now, if only it made more sense....

Ahahahahha
+1

Very nice of you to remember that.
P4K has no comparison to a GH2. JB even posted images of it being used in high end TV. Still, there are a lot of folks complaining. Go figure...

The overarching concept is floating somewhere between Alan Moore’s take on Swamp Thing and a cannibalistic nematode labyrinth experiment, or an Adderall and prosciutto fueled Fountain, orbiting themes of porcine resurrection within a Jungian metaphor.

Just kidding, I have no idea. Funny thing is that it’s often categorized as an experimental film, however the world is so specifically defined it’s more genre.

Good Luck

I dont exactly remember the story, only remember feeling exactly as you described. Also, at the time, I wasnt distracted by the image. I remember it felt like just any other movie I was used to. Except the story was too strange for me.

This is an interesting discussion because I've always thought it was weird that more BMPCC 4K owners didn't talk about the yellow shift in foliage outside. Now, I guess it could be me somehow (I do use a vari-ND sometimes and I don't know if I did in this shot or what effect that would have in this case) and this is just a quick pull from a random shot - but when I was starting to learn the ins and outs of the camera outdoors I discovered I needed to hue-shift the greens or they just didn't look anything like what I was seeing in person. I expected the default film conversions from raw to do the color correction so I was surprised to need to do the hue shift which also looks a bit funky.

(the white balance is off and whatnot on this example - but it's just a real world example of what I was experiencing with green shift).

I will try one of the methods listed here with different color space conversion and see how it goes. I'm currently trying to learn the ACES approach but I am getting an awfully weird conversion bug on the video monitor calibration LUT and the timeline there too (which again could be me but doesn't seem right).

This is an interesting discussion because I've always thought it was weird that more BMPCC 4K owners didn't talk about the yellow shift in foliage outside. Now, I guess it could be me somehow

So this zombie thread rises.... Here we go, right out the window. RCM, levels adjustment, a moderate reduction in saturation, and a mild S-curve. No hue adjustments. Shot at 5600 degrees; no card (as you can see), so no balance. If you want no yellow tones at all, we could do that, but then the color we'll be left with won't be green.

green_enough_1.2.1.jpg (992.14 KiB) Viewed 8071 times

The thing is; I see what you're seeing on your image. I see what I'm seeing in my image. Obviously in your conditions greens were fine. I am not trying to be "right" in this forum (as I believe this is a forum for discussion of issues not for simply winning arguments). This happens to be an issue I have trouble with that is unanticipated. Greens aren't showing up green where I'm used to in digital capture.

I find that foliage doesn't come out as I would expect in moderately bright scenarios. So, based on this discussion I am guessing that it is a clipping issue. It reminds me of overexposing chrome film (where things shift yellow). Somehow the green is getting out of range in a way I am not used to. I'm learning this camera and in doing the learning I am finding that for whatever reason greens in foliage are really not well represented in my footage. Am I doing something wrong on capture? Sure - I well could be. But, I'm not doing anything extraordinary to stretch the files (as far as I know).

Now, the example I'm posting the hue shifted version isn't even equivalent to the actual lushness of the green (and it's a bit of a weird hue compared to real life which was more muted but decidedly less yellow). FWIW - the greens were more similar to your tree image above to the naked eye.

As for it being a zombie thread - better than starting a new one right?

bmpcc4k_yellowgreens2.jpg (453.5 KiB) Viewed 8048 times

[attachment=0]

Things that affect green color:
- Sensor
- Sensor glass
- Lens
- lens filter (IR/ND/80B)
- lens fading/flare due ambient light.
- foliage type. Some fresh plants and leaves are almost yellow, some are saturated vibrant green, but in reality we are not always differentiate this and see all plants as "plain green" probably due eye color adaptation.
- workflow/processing (input profile, output transform to Rec709, film emulation LUT or film emulation build-in into input profile by camera manufacturer)
- monitor type. On modern wide gamut monitors greens will always look reacher.

I can't assess what's going on from your jpegs, but there are a few possibilities: IR pollution (prevailing shooting conditions), IR pollution+ color casts (ND filters) and most likely, post-production errors.

If you look further up in the thread, you'll see another example I posted of greens, in bright sunlight. I made a point of avoiding bright sunlight this time, so there wouldn't be objections to naturally occurring hues shifts or bright conditions taken as gamut clipping.

I can't assess what's going on from your jpegs, but there are a few possibilities: IR pollution (prevailing shooting conditions), IR pollution+ color casts (ND filters) and most likely, post-production errors.

Yeah, that's all very plausible. If only I could figure out which of the above it is. In terms of post-production errors - these are minimally touched. The climbers have a primary wheels change of [-0.03, 0.04, 0.93] in the 'default' with Raw set to 3900 WB, 35 Tint, ISO 200, no exp adj. No luts. Shot in braw.

So, if it's a post issue then I'm missing a fundamental which is concerning. The end result is that greens in foliage aren't what I'm expecting.

I've been shooting stills in this location for over a decade with everything from small format film to medium format digital and haven't had the leaves look like this (using the same tools, lenses, etc) across the board (Canon film and digital, Sony digital, Hasselblad film and digital, etc etc). Thus the confusion for me.

Nobody reads it, but the manual does offer color grading basics. Are you even normalizing the footage? In your case, I would read the relevant sections on color management (ACEs or Resolve Color Management) so at least you start out in the correct color space. Your stuff looks pretty logy.

Well, for sake of this discussion I did remove most of the adjustments in the footage above. I have read a portion of the manual (basics) as well as watched the training videos (shocking I’m sure - I even took notes). I’m just moving into ACES workflow as we speak so that is still relatively new to me. This issue (and footage) predates my moving toward ACES.

Still, I will revisit the manual. (That said while I think I can easily do a proactively corrective procedure - on injest I expect a profile to have a certain degree of accuracy).

I'm assuming you're shooting BMPCC 4K braw. Upload a frame of it for download here and behold the mess people make of it. But at least you'll get some sense of what's possible, with something you've already shot.

In my life I don’t think I’ve ever posted to a forum saying - here’s a picture tell me what I’m doing wrong. But, I guess we can feed the beast. Thanks for your thoughts. I’ll come back to this when I have a moment. (And yes, braw)

Jim, in some ways this thread bordered on trolling the camera so it may make sense to explore it in a new area where discussion could be more reasonable.

Default Davinci YRGB with REC 709 2.4, is a good option to explore grading options without necessarily jumping to ACES or another color space transform. There are some good examples on here about simply using a color space transform too within the default workspace to a color space of your choice to get a different result.

Juan Melara's powergrade is a good reference and I recently bought it to test out. You can dig into the nodes he provides and make your own custom adjustments. I would encourage you to try it and also see what you can do with the same clip moving the color wheels around and especially the offset wheel without necessarily making white balance / tint adjustments. There is more control between green / red vs the slider tint adjustment of green / magenta.

Still, for me there is a clear difference when handling reds and greens between V3 and V4 color science. I posted some differences below from the URSA Mini 4.6K where all adjustments are the same expect V3 to V4. I think you can get where you want with both, but the starting points have changed.

URSA Mini V3

ursa_mini_v3.jpg (889.33 KiB) Viewed 7982 times

URSA Mini V4

ursa_mini_v4.jpg (780.49 KiB) Viewed 7982 times

Here's another example between V3 and V4 with the URSA Mini 4.6K

Ursa Mini 4.6K V3 Color Science

portrait_v3.jpg (531.73 KiB) Viewed 7975 times

Ursa Mini 4.6K V4 Color Science

portrait_v4.jpg (467.05 KiB) Viewed 7975 times

I my case I think of them like different film stocks and was waiting to see if there was input from Blackmagic about whether they thought that the V4 color science should even be applied to the original URSA Mini 4.6K. As I'm using 2 of the original URSA Minis and now the pocket 4k I'm finding V4 color science to behave mostly the same way pushing down saturation in both reds and greens.

Here's another example between V3 and V4 with the URSA Mini 4.6K

portrait_v3.jpg

portrait_v4.jpg

Hey Ryan,

Thanks for posting these. It's great to have a person's face in there because we all know what skin tone SHOULD look like...

I see the greens look more vibrant, but I also look at the woman skin tone CHANGES between these images. Would you say you prefer one skin tone over the other here ? Also the colour of her shirt changes a lot too...

JB

Ryan Earl, can you post please those DNG frames? There is not so many examples from Ursa to learn difference between color spaces.

I prefer V4 skin tones, olive skin is olive to my eye and without the excess pink / red that can feel like plastic.

Color Science 3 URSA Mini 4.6K Zeiss 50mm CP2

v3-portrait-tungsten.jpg (574.22 KiB) Viewed 7915 times

Color Science 4 URSA Mini 4.6K Zeiss 50mm CP2

v4-tungsten-portrait.jpg (530.99 KiB) Viewed 7915 times

Color Science Gen 3 Ursa Mini 4.6K Zeiss 32mm CP2

truck-gen-3.jpg (614.37 KiB) Viewed 7911 times

Color Science Gen 4 Ursa Mini 4.6K Zeiss 32mm CP2

truck-gen-4.jpg (555.92 KiB) Viewed 7911 times

Here is probably where you can see the biggest difference in pinks / reds, this is a Xeen 35mm with Firecrest .3 ND

Color Science V3 Ursa Mini 4.6K Xeen 35mm

v3-portrait-brick.jpg (820.64 KiB) Viewed 7907 times

Color Science V4 Ursa Mini 4.6K Xeen 35mm

v4-portrait-brick.jpg (787.95 KiB) Viewed 7907 times

Here's another example between V3 and V4 with the URSA Mini 4.6K

The attachment portrait_v3.jpg is no longer available

The attachment portrait_v4.jpg is no longer available

Hey Ryan,

Thanks for posting these. It's great to have a person's face in there because we all know what skin tone SHOULD look like...

I see the greens look more vibrant, but I also look at the woman skin tone CHANGES between these images. Would you say you prefer one skin tone over the other here ? Also the colour of her shirt changes a lot too...

JB

I've been shooting stills in this location for over a decade with everything from small format film to medium format digital and haven't had the leaves look like this (using the same tools, lenses, etc) across the board (Canon film and digital, Sony digital, Hasselblad film and digital, etc etc). Thus the confusion for me.

I feel like pocket 4k skin tones have been the best so far "out of the box," it's a lot like working with portra film, even the texture of the noise.

Pocket 4K No offset

pocket4K-no-offset.jpg (961.09 KiB) Viewed 7919 times

Pocket 4K Cooler

pocket4k-cooler.jpg (988.95 KiB) Viewed 7919 times

Pocket 4K Warmer

pocket-4K-warmer.jpg (975.13 KiB) Viewed 7919 times

Backyard Covid Tests - iPhone SE 2020

IMG_0192.jpg (859.33 KiB) Viewed 7883 times

This is the background for my initial portraits above taken with an iPhone SE 2020. At least in my burned out grass there are yellows and browns.

Ryan, interesting results.

Did you find this iphone picture as a better presentation of what you saw with your own eyes vs a classic Rec709 V4 transfer?

Looking at both shots, the last iphone and the V3 “puddle”. I believe the color science should be somewhere between V3 and V4. At this point they both seem a bit extreme towards both ends of the spectra.

Ryan, interesting results.

Did you find this iphone picture as a better presentation of what you saw with your own eyes vs a classic Rec709 V4 transfer?

Looking at both shots, the last iphone and the V3 “puddle”. I believe the color science should be somewhere between V3 and V4. At this point they both seem a bit extreme towards both ends of the spectra.

Yeah, you can get in between, I prefer to start with V4 and I don't feel like I'm fighting to remove color casts or get a natural skin tone.

The iPhone can be a quick check where you can look right away and say "I can use this later as a reference or no I can't." It's not the most accurate but I think more so than not. Where I have the picture of the green background out of the car window, intuitively I might think it should have more blue if I don't have another reference point, but taking some extra time to look and then snapshot the iPhone, I think it has more yellow along with how color science V4 is interpreting it.

iPhone SE 2020

IMG_0199.jpg (1009.84 KiB) Viewed 7824 times

I tried to equalize the saturation below to show only color differences:

Ursa Mini SLR Magic 50mm V3 Saturation 75

gen3-sat75_1.116.1.jpg (631.59 KiB) Viewed 7824 times

URSA Mini SLR Magic 50mm V4 Saturation 100

gen4-100_1.117.1.jpg (646.95 KiB) Viewed 7824 times

Yeah, you can get in between, I prefer to start with V4 and I don't feel like I'm fighting to remove color casts or get a natural skin tone.

In your samples, v4 is apparently preferable to v3, but how exactly are you handling these images in post?

Up above, you cite saturation of 75 for V3, and 100 for V4. They values would be gargantuan (for Resolve saturation settings) if you were using either RCM or ACEs, or Resolve normalizing LUTs.

Could it be that what you're seeing is an artifact of the method (whatever it is) and not the cameras? Just asking....

Yeah, you can get in between, I prefer to start with V4 and I don't feel like I'm fighting to remove color casts or get a natural skin tone.

In your samples, v4 is apparently preferable to v3, but how exactly are you handling these images in post?

Up above, you cite saturation of 75 for V3, and 100 for V4. They values would be gargantuan (for Resolve saturation settings) if you were using either RCM or ACEs, or Resolve normalizing LUTs.

Could it be that what you're seeing is an artifact of the method (whatever it is) and not the cameras? Just asking....

Here all the previous portrait examples are in default YRGB and no luts, just working the wheels and adding in saturation and contrast.

I'm more willing to post a DNG without a face in it for Dmitry as I did earlier in the thread that you can take through RCM or ACES which could illustrate V3 to V4 and the reds and greens.

Below I have color temp the same from the URSA Mini, both 65% saturation, a contrast adjustment in RCM with REC 709 2.4 for project and output. Let me know if you want me to post screenshots of the exact development settings.

URSA Mini V3 Saturation 65

ursa_mini_v3.jpg (534.53 KiB) Viewed 7790 times

Ursa Mini V4 Saturation 65

ursa_mini_v4.jpg (984.52 KiB) Viewed 7790 times

URSA Mini V4 Sat 90

ursa_mini_v4_saturation_90_1.3.3.jpg (545.52 KiB) Viewed 7787 times

URSA Mini V4 Sat 90 Hue 47.4

ursa_mini_v4_hue_47_sat90.jpg (538.94 KiB) Viewed 7787 times

It's not wrong to want more saturation is it? Even working with RCM or ACES? This time I saturated the V4 color to 90% and it reproduces the skewed green to yellow and the water bottle skews violet to blue. I adjusted hue to 47.4 and the color was brought back to attempt to match V3.

It's not wrong to want more saturation is it? Even working with RCM or ACES? This time I saturated the V4 color to 90% and it reproduces the skewed green to yellow and the water bottle skews violet to blue. I adjusted hue to 47.4 and the color was brought back to attempt to match V3.

Though it's possible to get results the way you're doing it, my own view, fwiw, is that most people are far better off leaving the conversion math to Resolve, either through LUTS (pretty good math) or Resolve Color Management (better math). That way you know you're not grading against initial normalization errors from the outset. You may be able to push and pull it into some manner acceptable image doing it your way, but by then, the damage may be done.

At least, that's my experience of it. Professional graders do it both ways. Some prefer manual, but many others rely on LUTs and color management.

If you use RCM or Resolve-supplied LUTs, I'd be very surprised if you find the need to raise saturation. It almost always goes the other way (lowered). ACEs conversion tends to be less saturated from the outset, among other differences, so there you might or might not.

It's not wrong to want more saturation is it? Even working with RCM or ACES? This time I saturated the V4 color to 90% and it reproduces the skewed green to yellow and the water bottle skews violet to blue. I adjusted hue to 47.4 and the color was brought back to attempt to match V3.

Though it's possible to get results the way you're doing it, my own view, fwiw, is that most people are far better off leaving the conversion math to Resolve, either through LUTS (pretty good math) or Resolve Color Management (better math). That way you know you're not grading against initial normalization errors from the outset. You may be able to push and pull it into some manner acceptable image doing it your way, but by then, the damage may be done.

At least, that's my experience of it. Professional graders do it both ways. Some prefer manual, but many others rely on LUTs and color management.

If you use RCM or Resolve-supplied LUTs, I'd be very surprised if you find the need to raise saturation. It almost always goes the other way (lowered). ACEs conversion tends to be less saturated from the outset, among other differences, so there you might or might not.

I can reproduce the behavior of the color science v3 and v4 with the color space transform tool and RCM. Here I've made no adjustments to the images beyond the CST.

CST URSA Mini 4.6K Gen 3 with Gen 3 in Raw Tab

CST_GEN3-GEN3-REC709.jpg (556.12 KiB) Viewed 7741 times

CST Blackmagic Design Wide Gamut Gen 4 with Gen 4 in Raw Tab

CST_WIDEGAMUTGEN4-GEN4.jpg (551.37 KiB) Viewed 7741 times

CST Blackmagic Design Wide Gamut Gen 4 with Gen 3 in Raw Tab

CST-WIDEGEN4-GEN3.jpg (622.36 KiB) Viewed 7741 times

If you were to develop the cDNG in Photoshop RAW you would find Adobe Color, Adobe Standard, Adobe Portrait, Adobe Vivid. I think V3 and V4 are acting like Adobe Vivid vs Adobe Portrait, then the pocket 4K following the look of portrait.

Yeah, you can get in between, I prefer to start with V4 and I don't feel like I'm fighting to remove color casts or get a natural skin tone.

Yes, that is also my conclusion regarding V3 and V4. There are some obvious drawbacks regarding lush green colors with V4 (out of the box), but it’s not that hard to bring back if you want natural green. But all this is about natural colors… like Rec709 transfer.

Ironically. Modern colorist don’t focus on Rec709 like they used to. Natural look is not really common in most productions. Today its more about the artistic aspect where you want to put a fingerprint on the work and go to extremes that was not "allowed" some years ago. The (film look) is a dead subject IMO. The V4 is focused on skin tones. It’s supposed to be a better base point for the colorist regarding that. But all the colors are waiting to be picked out regardless. Looking back to the era of analog film. You choose a filmstock… That was the look! The only real change you made in post was contrast, temperature and to some degree exposure. That’s not the case with V3 or V4 for that matter. We have total control and can do much more. Like emulating any filmstock with V3 or V4 regardless.

The topic was about trouble with green color regarding the BMPCC4K, but we know by now that it’s possible to get those colors out V3 or V4 regardless. Its falsified as a problem IMO. It’s only a problem If you want, out of the box, natural color straight to REC709.

But let’s look at it from a colorist perspective. You want to achieve (The Look) most efficient. What is the best basis? So the question I would ask at this point. What is the most efficient base to achieve ANY grade when it comes to the BMPCC4K and Davinci resolve? This is a tricky question with many answers, few more right than the other.

I'm more willing to post a DNG

I'm still a bit puzzled by your results. If you can upload a single .dng frame, something with grass (and ideally face), I'd be interested in putting it through accustomed workflows.

I'm still a bit puzzled by your results. If you can upload a single .dng frame, something with grass (and ideally face), I'd be interested in putting it through accustomed workflows.

I can see what additional images I have and if you want one of the frames from what I already posted I'll ask my "model" for permission.

With covid-19 restrictions still in a no-work - no-entertainment zone in my area I have some time to construct a new scene. I'm sure I have grass and some objects between home and studio. The face is the tricky part, I'm not keen to be in the shot, I can see if a family member or local actor would be willing to be in the scene.

I can include the URSA Mini and the pocket 4K along with 2 different lenses like Zeiss CP2 and SLR Magic APO. I use Firecrest NDs which would influence color too, so maybe let me know if you want the ND off.

I can accommodate additional requests or restrictions as far as objects in the scene or the lighting. In a new test where there are direct comparisons you could find the most benefit when reviewing the images.

What is the most efficient base to achieve ANY grade when it comes to the BMPCC4K and Davinci resolve? This is a tricky question with many answers, few more right than the other.

And for stills work Adobe doesn't appear to have resolved a single base. I've found some interviews from Adobe in the past about how and why they settled on "Adobe Color" which I can try and find. With corporate head shots I will take the Pentax 645Z images and replace the camera name and brand meta info with "Ricoh645Z" to import to Capture One for their RAW development workflow over Adobe's.

In my case, something you've already posted would be helpful, for comparison to your results.

This link has 2 of the images I posted earlier and 2 new ones.

https://we.tl/t-AJSgEhmnwW

Ryan, there is not so huge difference between Gen 3 and 4. Extra saturated greens in your examples because RAW and input color spaces are not match (as you described). Technically this is wrong, but visually assigning wrong color spaces may result interesting effects and shifted colors. Just make sure you understand what you do.

Ryan, there is not so huge difference between Gen 3 and 4. Extra saturated greens in your examples because RAW and input color spaces are not match (as you described). Technically this is wrong, but visually assigning wrong color spaces may result interesting effects and shifted colors. Just make sure you understand what you do.

If you take the image I included with the color chart and just drop it into Davinci YRGB Color Managed, REC709 leave it set to GEN 3 in the RAW tab then add only saturation. Take the same image and do the GEN3 CST to REC709 with GEN 3 in the RAW tab I find the colors needed to get bright greens to represent the objects in the picture shift away and become less accurate.

Just putting URSA Mini 4.6K GEN 3 in the timeline and adding saturation, it's accurate to my eye in the bright green objects and the red violet spray bottle.

If you add the CST with correct settings, it's not as accurate in the bright greens and violets and it's closer the color you get with GEN 4 and GEN 4 on the pocket 4k. A slight hue shift is required to bring it back or using the color checker tool.

So then my question is, is it wrong to simply work with GEN 3 in the timeline without a LUT and without a CST especially if I can output more accurate color for product imagery or objects with bright colors?

GEN 3 in the Timeline - No Lut - No CST- Saturation Only

DavinciYRGB_Color_Managed_GEN3_NO_LUT_NO_CST_SAT80_1.126.1.jpg (999.6 KiB) Viewed 7634 times

GEN 3 in the Timeline -CST 4.6K GEN 3 to GEN 3 RAW TAB

DavinciYRGB_Color_Managed_GEN3_CST_GEN3_1.127.1.jpg (536.84 KiB) Viewed 7634 times

CST 4.6K GEN 3 to GEN 3

GEN3-CST4.6KGEN3-GEN3-RAW.jpg (687.59 KiB) Viewed 7624 times

GEN 3 No Lut No CST Saturation Adjustment

GEN3-NO_CST.jpg (743.61 KiB) Viewed 7624 times

Here is the result where GEN3 greens in the timeline are not shifted yellow / brown when you are not using a CST. Then they are after adding the CST. You have to work it back to the vibrant green after performing the CST.

Then below an image with DNG link of my dad hanging out in my studio with the URSA MINI 4.6K and SLR Magic 50mm. The skin tones are more realistic to me after doing the CST in V4.

https://we.tl/t-uAqcqx1ueV

URSA Mini 4.6K V4 SLR MAGIC 50mm

portrait-v4.jpg (596.87 KiB) Viewed 7624 times

I would look again at your workflow. This is the backyard shot, with RCM and v3, and saturation lowered from the default of 50 to 30. Sorry about the bars, incorrect export settings.

2_1.2.1 copy.jpg (544.07 KiB) Viewed 7620 times

It is technically wrong simply increasing saturation to fit wide color space to small color space. Color primaries and white point arranged in very different way in each color space. Plain saturation increase may give you more vibrant greens, but same time may produce hue shifts and low saturated areas in other colors. This method may only provide some artistic "look" but same time it may provide color clipping artifacts, increased noise and probably posterization effects.

I would look again at your workflow. This is the backyard shot, with RCM and v3, and saturation lowered from the default of 50 to 30. Sorry about the bars, incorrect export settings.

But. . . That's my point, where people are struggling with green is in GEN 4 not GEN 3. If you do the CST in GEN 3 to GEN 3 it does yellow greens slightly but not to the extent in GEN 4 to where someone would struggle to bring it back.

The only reason I chose gen 3 is because 4 isn't supported in RCM. But I did try it with LUTs, both v3 and v4. There's not a whole lot of difference, or least none that can't easily be compensated for in the grade. It's your workflow which is wrong -- unless you like the results you're getting.

The only reason I chose gen 3 is because 4 isn't supported in RCM. But I did try it with LUTs, both v3 and v4. There's not a whole lot of difference. It's your workflow which is wrong.

If you apply your settings to the table top still life that I included I'd like to see the result. As I posted above when I follow your settings the greens skew yellow. I need a hue shift.

I'll be more convinced if you were to take on the scene that includes the color checker. Of course we are not in the same room looking at the scene, but the color checker tool will make an adjustment that I believe corrects the skew within your workflow.

And if you simply drop the clip into the timeline without a CST in GEN 3 you'll find the greens in that image don't require shifting to be green. I do like the result I get there without CST or LUT.

Proper film emulation LUTs gives you COLORS.
ColorChecker Correction provides better greens.
WORKFLOW: viewtopic.php?f=21&t=65149&p=543725#p537852
LUTs: viewtopic.php?f=21&t=77553&p=637455#p569634
Examples:



CST node transform only provides you boring basic look:

Dmitry, I think she would be happy with the Velvia version of herself here, it looks really great!

Turns out I shot the perfect example of this, salad!

Shot with both a UMP and Pocket4k, prores with extended video lut applied in post. A very default setup.


The salad basically came out yellow, it looked wilted. It was very east to adjust with hue vs hue, which didn't seem to affect skin tones. (note the blue saturation boost in the bowl is from additional color work, not caused by the green HvH correction.)

Never noticed the average greens in Gen4 until this salad shoot, would be great to have it patched up by BMD though.

Comparison.jpg (639.67 KiB) Viewed 7602 times

Proper film emulation LUTs gives you COLORS.

CST node transform only provides you boring basic look:

As Mike is showing in the salad, hue vs hue is still a tool that could be applied to the log image in the greens behind my "models" face to hold the skin tone while creating a custom look.

Turns out I shot the perfect example of this, salad!

Before we go any further, is it safe to assume you had no filters in front of the lens, except maybe IR cut? Or that you'll try again with an IR cut ?

A frame would be helpful, to at least confirm that post-production isn't doing it. Jsfilmz also had actual yellow/brownish greens, but the only apparent explanation was use of NDs, because his results couldn't be replicated by this shooter. Green was/is green.

Turns out I shot the perfect example of this, salad!

Before we go any further, is it safe to assume you had no filters in front of the lens, except maybe IR cut? Or that you'll try again with an IR cut ?

A frame would be helpful, to at least confirm that post-production isn't doing it. Jsfilmz also had actual yellow/brownish greens, but the only apparent explanation was use of NDs, because his results couldn't be replicated by this shooter. Green was/is green.

I think that is disingenuous John. You've had to alter your previous comment to include "at least none that can't easily be compensated for in the grade."

Good question, The UMP has a Rawlite permanently installed which has an IR cut and the pocket does not. Both displayed the same symptoms though

No other filters were used, shot with the same glass sigma 18-35 on the pocket, 50-100 on the ursa, shot with LED's in a cool room, not much heat to generate IR pollution.

I think previous examples clearly show its a difference between gen3-4 color science, its seems BMD went a little overboard trying to correct skin tones and nuked the greens in the process.

Here are the pro-res files, it was a mulitcam shoot, these are the same take for each camera.

https://www.dropbox.com/sh/jqf5xum8r6c5 ... lNtYa?dl=0