massive shift in green color on BMPCC4K

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 7:47 pm

Eugenia Loli wrote:That's how I got the tree shot looking like it should be.


This thing has knotted itself into one big pretzel. You're working from somebody else's footage, shooting conditions unknown, there's apparently a difference of opinion as to whether it suffers from IR pollution, a condition typically all but impossible to fix, and your corrected shot, as seen from a jpeg, raises questions of its own, because it has its own very distinct color cast, which this viewer wouldn't associate with film or movie aesthetics....

I was going to start posting examples, which demonstrate none of the brown or yellow shifts, but we already have shots that Robert offered a few days ago, with access to the originals. None of his material display the ills attributed to that "surveilliance" sensor.

Is any of this real?
Offline
User avatar

Eugenia Loli

  • Posts: 207
  • Joined: Mon May 04, 2015 6:47 am
  • Location: Spokane, WA, USA

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 7:57 pm

>because it has its own very distinct color cast


No, it does not. There is no color cast anymore. Each color is separated. Unlike the default lut result which is just yellow. There was nothing wrong with that footage to shoot it down the way you do btw. I also found a backup of mine last night, and have tried my method with a few of my remaining clips from last year. I was able to get the desired result with my own footage too (which I've shot WITH an IR CUT filter).

>which this viewer wouldn't associate with film or movie aesthetics....


What you see in my corrected image, is the color range of what you should expect from actual film or Alexa when processed in a wide color space.
Collage artist, illustrator, filmmaker: https://vimeo.com/eugenia
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 8:00 pm


All done in ACES in REC709 output transform. I can see the extra red/magenta ( IR?) in the tree but the green as I said just needed a very small selective Hue shift. Pulled both the BRAW clips down to their lowest ISO settings and used scopes with a selection mask to match skintone but notices there is a coloured light going on and off but it's not a bad match considering it's different cameras in a mixed lighting situation on skintone. I don't see any fundamental problem here at all. I'd take the P4k over the S1 if absolute colour fidelity was the goal but the S1 if resolution, high iso/low light and DR were more important. No clear winner so keeping them both.
probably best password protected as it's got someone in it
s1hvsp4k
It's gone in 5mins.
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
Offline
User avatar

rick.lang

  • Posts: 13396
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 8:38 pm

The exterior looks good to me, John. The interior is an impossible situation due to the every changing light on the talent. I’d never use that clip.
Rick Lang
Offline
User avatar

Eugenia Loli

  • Posts: 207
  • Joined: Mon May 04, 2015 6:47 am
  • Location: Spokane, WA, USA

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 8:43 pm

All done in ACES in REC709 output transform.


This looks good, John. Wide color spaces for the win!
Collage artist, illustrator, filmmaker: https://vimeo.com/eugenia
Offline

John Brawley

  • Posts: 2362
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: London UK

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 10:59 pm

Eugenia Loli wrote:The BMPCC 4k does use a night surveillance sensor, nothing hearsay about that part.


Would love to see an attribution reference that actually can prove emphatically what you're claiming.

It sounds like the workflow problem was that you assumed the Resolve defaults would be the correct setting for you to work with the gamut of the camera the way you want to ?

Your problem is essentially solved even though it (apparently) has nothing to do with the alleged use of some inferior (to a Panasonic S1H ) night surveillance sensor ?

JB
John Brawley
Cinematographer
London UK
Offline

Dune00z

  • Posts: 319
  • Joined: Thu Oct 18, 2018 3:40 pm
  • Real Name: Duane Eues

Re: massive shift in green color on BMPCC4K

PostFri May 22, 2020 11:40 pm

John Brawley wrote:
Eugenia Loli wrote:The BMPCC 4k does use a night surveillance sensor, nothing hearsay about that part.


Would love to see an attribution reference that actually can prove emphatically what you're claiming.

It sounds like the workflow problem was that you assumed the Resolve defaults would be the correct setting for you to work with the gamut of the camera the way you want to ?

Your problem is essentially solved even though it (apparently) has nothing to do with the alleged use of some inferior (to a Panasonic S1H ) night surveillance sensor ?

JB


It isn't a sensor problem and I would also like to see proof that the sensor is a surveillance sensor. If so, its a fine one.

The problem looks like user error and that is usually the case with these situations...
Offline
User avatar

Eugenia Loli

  • Posts: 207
  • Joined: Mon May 04, 2015 6:47 am
  • Location: Spokane, WA, USA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 1:46 am

IT IS NOT "my" user error. Jeez!!!! Why can't you understand that working in plain rec709 is the WRONG way to do it for BRAW footage SPECIFICALLY?!? This is how 99.9% of the users are using BRAW, and how everyone is taught to use it. EVERYONE is doing a "user error" at this point.

The green channel particularly is CLIPPING if you're working in plain rec709, so you need a much wider color space to capture all the color range. You see, digital sensors have more green colors than R or B. So they are very sensitive to it. Arri has a LOG format that can FIT their sensor's data, but Blackmagic does NOT. After all these v2, v3 and v4 versions they have, none can fit the whole dynamic range of the Sony sensors.

And if you try to fit the data by using luma and saturation mapping, these algorithms inside Resolve are VERY HARSH, so they clip the data EVEN MORE. If Blackmagic can't fix their color science format, then Resolve itself needs a SPECIAL algorithm to ease-in the whole data into the rec709 space. Currently, it doesn't FIT, that's why we see the color casts. Greens turn into yellows! Blues turn into cyans! And you're all familiar with the reds going berserk at night red lights! These are all symptoms of color clipping because the color science format is not good, AND because rec709 can't fit it in.

So to currently go around the problem, you have to either use a wider colorspace (and first transform the footage to something flatter), or if you must use rec709, you must change the colorspace/gamma within the RAW format and do a few more color transformations, as seen here. Part of the problem is in fact EXACTLY shown here:


As for the surveillance sensor argument, simply read which Sony sensor this one is, on the Sony catalog, based on its size. The information is out there. But let's just stay on topic here.
Collage artist, illustrator, filmmaker: https://vimeo.com/eugenia
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 2:34 am

Eugenia Loli wrote:[Part of the problem is in fact EXACTLY shown here:


In the video you cite, Juan Melara offers no support for your claims of a uniform color cast, thanks to supposedly inadequate rec. 709 color gamut space. And no independent proof, apart from a jsfilmz video frame, not provided here for examination, has been offered to support your theory that the traditional workflow produces footage with crude unacceptable color casts, or that uncorrected footage skews brown or yellow. At least one user here downloaded the jsfilmz footage and provided a graded sample at least as persuasive as yours, using standard rec. 709 space.

The claims Juan Melara did make, for recovering highlights and shadows and saturation in the highlights, were at least partially disavowed later. This is what he said on his website:

CaptainHook from Blackmagic Design has been in touch with a few corrections. First of all the BMD 4.6k Film curve is actually able to hold all of the dynamic range of the sensor. It just stores the brightest highlights beyond the standard range. This is so that the image occupies the full range of code values, making use of all the available bits in the container. This is actually smart for formats such as 10bit Prores 422HQ and even 12bit Prores 4444 as the Ursa Mini records. It's not something you would normally see in raw formats as raw gets decoded into a much higher precision 32bit float space, so distributing the data the way this curve does is unnecessary. Keeping the curve consistent between Prores and raw is likely the reason it behaves this way.

The saturation issue is apparently a bug in the way Resolve interprets the old pre V3 colour science. Apparently no one who has downloaded and worked with this footage or shot their own pre V3 footage noticed or reported it. It should be fixed with the current V3 colour science.
Offline
User avatar

Eugenia Loli

  • Posts: 207
  • Joined: Mon May 04, 2015 6:47 am
  • Location: Spokane, WA, USA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 2:47 am

At least one user here downloaded the jsfilmz footage and provided a graded sample at least as persuasive as yours, using standard rec. 709 space.


No, he didn't. He said that he could not make it green. The only user who could, worked in ACES, not in rec709.

As for Melara, I switched my workflow to his, and it works like a charm to have the SAME output as I had last night with my own rec2020 workflow. Transformations to larger color spaces are needed to retain the full color information without clipping. Plain rec709 (throwing a clip in the timeline and grading it) does NOT get rid of the various clipping casts.
Collage artist, illustrator, filmmaker: https://vimeo.com/eugenia
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 2:52 am

Eugenia Loli wrote:No, he didn't. He said that he could not make it green. The only user who could, worked in ACES, not in rec709.

As for Melara, I switched my workflow to his, and it works like a charm to have the SAME output as I had last night with my own rec2020 workflow. Transformations to larger color spaces are needed to retain the full color information without clipping. Plain rec709 (throwing a clip in the timeline and grading it) does NOT get rid of the various clipping casts.



He used ACEs, within a rec. 709 workflow. And anything which can be done in ACEs this way can be done in numerous other traditional ways, including RCM or Davinci YRGB.

I can't answer for the other user (who attributed his failure to IR pollution, not gamut clipping), and haven't been able to download that material. The site says too many downloads.

Lots of people here own cameras, but nobody seems able to provide any samples of supposedly skewed color casts, or clipped color casts, other than this apparently immortal example from jszfilm. Odd! And even that supposedly clipped color is magically rejuvenated by ACEs, in your own estimation.

This whole thing is easy to prove or disprove; I don't know why we're arguing. Link to a braw frame, free of IR pollution, and a grade you believe is only possible with your method. If you're right, we will have all learned something.
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 4:34 am

Eugenia, your example with "fixed" tree color posted here viewtopic.php?f=2&t=114189&view=unread#p629930 is not a fix at all. You just shifted all colors to blue. Yellow colors became greener, but same time distant green trees became almost black. It could be due WB/Tint adjustment or due it could be some mistake in color spaces transforms when you select Rec2020 in BRAW settings but use BMDfilmv4 as input color space node.

Also these two nodes do the same thing. Use only one if them and use only in the last node, but not both.
- Add the arri rec709 lut to the third node.
- At the last node, do another CST to Rec709, with sat/luma mapping
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 5:10 am

Pocket4K Footage example with tree is brownish for sure.
It is same brownish in YRGB BMDfilm or in YRGB Rec2020 or in ACES project (unless you apply in-out color spaces that didn't match one to other)
We have no idea how it was shoot. Maybe some specific lenses produce IR pollution.
Maybe that tree is brownish in reality. Distant far trees on are way greener.
Сlaim that BMPCC 4k use a night surveillance sensor sounds like passive aggressive :mrgreen:
If someone missed this, topic starter Fixed problem with Color Checker correction plugin in After Effects. viewtopic.php?f=2&t=114189#p629421 So as it was suggested from beginning - use IR filter, use ColorChecker correction.
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

Dune00z

  • Posts: 319
  • Joined: Thu Oct 18, 2018 3:40 pm
  • Real Name: Duane Eues

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 5:37 am

My mistake.

The user had to change their method from one that doesn't work, which resulted in color gamut errors, to one that does work.

Call it whatever you like.

Good luck.
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 8:03 am

Pocket4K ProRes Example from BM website. Processed in same way as my other examples. Color Temp adjusted to -100 and some Noise Reduction applied. No any film LUTs of ColorChecker correction applied. Nice shades of green colors everywhere:
Image

And the same with Fuji Provia100 Film emulation LUT:
Image
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 8:47 am

Same time in another side by side sources shared by JSFILMZ and processed with same workflow and white balanced to same patch we see things like this. Weak greens and magenta cast in Pocket 4K compare to ARRI.
And i can't match those shoots even with quick automatic color checker correction.
Image
Image
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 10:37 am

Here is render from another test sources. Same project setup, both balanced by same grey patch. Even when image was captured to DNG, Pocket 4K produced same purple-magenta tint in shadows. Ursa G2 shifts to green-yellow side. And the worst thing - we don't know is this due different lenses used in test, or due different sensors.
Image
Image
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

Robert Niessner

  • Posts: 2954
  • Joined: Thu Feb 21, 2013 9:51 am
  • Location: Graz, Austria

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 4:27 pm

Dmitry, one problem in the last example is that the main light has hit the color checker in different ways. You can clearly see this in the black patch. This can throw the whole automatic thing off the track.
Robert Niessner
LAUFBILDkommission
Graz / Austria
---------------------------------------------------------------------
Blackmagic Camera Blog (German):
http://laufbildkommission.wordpress.com

Read the blog in English via Google Translate:
http://tinyurl.com/pjf6a3m
Offline

John Brawley

  • Posts: 2362
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: London UK

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 4:30 pm

Robert Niessner wrote:Dmitry, one problem in the last example is that the main light has hit the color checker in different ways. You can clearly see this in the black patch. This can throw the whole automatic thing off the track.


Indeed. That's why you have to have the colour checker on a stand.

I like to make a light trap as a black references as well.

You can also use one of these.

https://www.bhphotovideo.com/c/product/ ... rCube.html

JB
John Brawley
Cinematographer
London UK
Offline

John Brawley

  • Posts: 2362
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: London UK

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 4:43 pm

Eugenia Loli wrote:
As for the surveillance sensor argument, simply read which Sony sensor this one is, on the Sony catalog, based on its size. The information is out there. But let's just stay on topic here.


No Eugenia.

You said

"And while the S1H uses also a Sony sensor, it's using a newer, actual dSLR sensor, and not a night surveillance sensor (that the BMPCC 4k uses) that happens to be good-enough to be soldiered into cameras."

So you're making a few claims there, and really it's to support an argument about colour discretion that you have NOW SOLVED by other means.

You're making a claim about the build quality, somehow inferring it's "soldiered-in" and is an inferior sensor. I happen to disagree with you, and it sure isn't what I understand of what's going on here, so I'm asking you to simply show in a credible way that isn't " I read it on the internet" proof.

So, either, it's irrelevant that it's a "night surveillance sensor" or you were just trying to throw out an insult for why you couldn't get the result you wanted which has nothing to do with if it's a night vision surveillance sensor (whatever that is !)

You haven't been able to show how you KNOW that it's a night surveillance sensor and the S1H isn't ?

I'm not very knowledgeable about night surveillance cameras, but I'm pretty sure most surveillance cameras have much much smaller sensor sizes.

What you claim is ridiculous, not credible and can't be verified.

You're making an assertion about why you're getting a certain result that you don't like and offering this "evidence" as to why.

Except that it's not the case because it turns out to be about the way one works in Resolve. You even used my footage from a different camera shot with a different codec and a different sensor to make a point about it.

You can keep making claims, but I'm going to ask you to simply back up what you say. Prove it instead of just writing it.

So far you haven't been able to show your issues very clearly. You can't even show with your own footage, you're relying on others to do so.

JB
John Brawley
Cinematographer
London UK
Offline

Ryan Earl

  • Posts: 115
  • Joined: Sun Jan 17, 2016 6:56 pm

Re: massive shift in green color on BMPCC4K

PostSat May 23, 2020 11:09 pm

Pocket4K-ACES11-REC709-24-5000K-.jpg
Pocket 4K ACES CCT 1.1 REC709 5000K 0 TINT 65 SAT
Pocket4K-ACES11-REC709-24-5000K-.jpg (308.68 KiB) Viewed 591 times


I have a Ryobi power washer, and the wand no longer works :cry: I took the picture with a Pocket 4K this afternoon and a Nikon 50mm 1.4G AF and developed it with ACES in Resolve with input, timeline and output set all to REC 709 2.4. The output card is a Blackmagic Intensity 4K to a REC 709 2.4 Monitor. I then loaded the still I exported into Photoshop and combined it with the Ryobi drill screenshot from someone's earlier post.

Rainy Saturday here in Philadelphia and I've been trying to digest the methodology suggested in this thread for color managing images taken with the Pocket 4K.

You generally (really all the time) want to match the output transform in the project settings, REC 709 2.4 to REC 709 2.4 of your monitor. If you have a REC 2020 monitor then REC 2020 for timeline and output. Otherwise how will you see what you are working with?

I think where there is some contention is whether or not you should perform an input transform on top of the Pocket 4K. In my case I did select "Blackmagic Design Film" for the ACES input transform.

I can upload the original Blackmagic RAW file or DNG if anyone wants to play with it with other settings.
Offline
User avatar

Dmitry Shijan

  • Posts: 1365
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: massive shift in green color on BMPCC4K

PostSun May 24, 2020 6:58 am

Ryan Earl wrote:developed it with ACES in Resolve with input, timeline and output set all to REC 709 2.4.

Lucky for you for BRAW in ACES Resolve always ignore project input setting and set proper BMDfilm input color space and gamma under the hood. Use REC 709 2.4 as input for wide gamut log source is a big mistake.

Ryan Earl wrote:You generally (really all the time) want to match the output transform in the project settings, REC 709 2.4 to REC 709 2.4 of your monitor. If you have a REC 2020 monitor then REC 2020 for timeline and output. Otherwise how will you see what you are working with?

Color space transform, color management, and technical LUTs do the thing. So you manipulate image in wide color space and log gamma but see result compressed inside Rec709 color space in realtime.
Check my workflow here viewtopic.php?f=21&t=65149#p537852

Ryan Earl wrote:I can upload the original Blackmagic RAW file or DNG if anyone wants to play with it with other settings.

Yes, please! As you can see daylight always reproduce colors way better than LED lights.
P.S. I will also shoot some examples with BMMCC later as well.
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

Paul Leeming

  • Posts: 9
  • Joined: Sun Mar 08, 2020 9:19 pm
  • Real Name: Paul Leeming

Re: massive shift in green color on BMPCC4K

PostSun May 24, 2020 7:00 pm

Ivan, just seeing this post. Here's my Leeming LUT Athena for BMPCC4K on your raw PNG file you posted, plus I adjusted white balance off your X-Rite and lifted the exposure a bit. The goal with my LUTs is to get a DeltaE < 1 and my BMPCC4K one measures 0.68 so it's essentially imperceptible to the human eye!

mylut_1.75.1.jpg
mylut_1.75.1.jpg (940.78 KiB) Viewed 506 times


Hope this helps!

Cheers,

Paul :)

Edit: would be good to have the BRAW as the PNG is low quality which leads to the banding seen in my example.
Paul Leeming
Director / Cinematographer / Colourist
https://www.LeemingLUTPro.com
https://www.visceralpsyche.com
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 4:21 am

Eugenia Loli wrote:
At least one user here downloaded the jsfilmz footage and provided a graded sample at least as persuasive as yours, using standard rec. 709 space.


No, he didn't. He said that he could not make it green. The only user who could, worked in ACES, not in rec709.

As for Melara, I switched my workflow to his, and it works like a charm to have the SAME output as I had last night with my own rec2020 workflow. Transformations to larger color spaces are needed to retain the full color information without clipping. Plain rec709 (throwing a clip in the timeline and grading it) does NOT get rid of the various clipping casts.

So, for someone like myself, who doesn't pretend to be anywhere close to the level of any of you guys, what would you suggest be the best workflow to get better colors from the 4k?
https://www.imdb.com/name/nm0972296/
Offline
User avatar

carlomacchiavello

  • Posts: 1037
  • Joined: Tue Aug 28, 2012 6:04 pm
  • Location: italy

massive shift in green color on BMPCC4K

PostTue May 26, 2020 5:28 am

Eugenia Loli wrote:, as in the Fuji cameras, that ALSO use Sony sensors), but fixes the greens.


Color is matter of taste, I respect the taste
but... Fuji developed their sensors from twenty and more years, from super ccd to actual x-trans sensors, not use Sony tech.

I can tell you that I used every kind of brand and every brand had their look, no one is neutral, canon is unbalanced on magenta on skin tone and not only, Panasonic add pink on white, reduce blu on skin, Sony give me too much green and more... Kodachrome (which some one tell that is emulated from canon, big no no, add magenta on skin tone, yellow often had brown, green over some saturation is green under some saturation go to brown, and more) had a look, every tool had a look (today called color science).
If we had a device no standard to see (monitor, tv, smartphone/tablet display) we can try to neutralise a bit all but everything we do will be see in different way from different people, if you add the fact that most of people enable evil function that add yellow and blu in different hour of day, and all engeneer try to destroy you picture in tv with other evil function on gamma and color...
Every camera had a look, like people in past search a look in a brand of film strip.
Also in the print environment when you print art catalogue you fight with color and never you reproduce real color, most of great printer arrive near, with additional color to classic cmyk, but print is an art that can be linear (if printing process could be constant, also different printing pressure change color), video is so weak in reproduction and not controllable that it’s very difficult, like you discover, to “manage” correctly,
This is the reason behind the existence of work like DiT, colorist, and more in a video/movie production.

Inviato dal mio iPhone utilizzando Tapatalk
Last edited by carlomacchiavello on Tue May 26, 2020 6:40 am, edited 2 times in total.
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 5:34 am

Henchman wrote:
Eugenia Loli wrote:
At least one user here downloaded the jsfilmz footage and provided a graded sample at least as persuasive as yours, using standard rec. 709 space.


No, he didn't. He said that he could not make it green. The only user who could, worked in ACES, not in rec709.

As for Melara, I switched my workflow to his, and it works like a charm to have the SAME output as I had last night with my own rec2020 workflow. Transformations to larger color spaces are needed to retain the full color information without clipping. Plain rec709 (throwing a clip in the timeline and grading it) does NOT get rid of the various clipping casts.

So, for someone like myself, who doesn't pretend to be anywhere close to the level of any of you guys, what would you suggest be the best workflow to get better colors from the 4k?

ACES
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:29 am

John Griffin wrote:ACES


That's it?
https://www.imdb.com/name/nm0972296/
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:51 am

And where do you set the aces setting?
https://www.imdb.com/name/nm0972296/
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 7:12 am

And where do you set the aces setting?
https://www.imdb.com/name/nm0972296/
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 8:50 am

Henchman wrote:
John Griffin wrote:ACES


That's it?

yes
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 8:59 am

Henchman wrote:And where do you set the aces setting?

Project settings - Color management - color science - set to ACEScc or ACEScct. Then there is a bit of a learning curve but plenty of info out there.
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 12:53 pm

Henchman wrote:
Eugenia Loli wrote:
At least one user here downloaded the jsfilmz footage and provided a graded sample at least as persuasive as yours, using standard rec. 709 space.


No, he didn't. He said that he could not make it green. The only user who could, worked in ACES, not in rec709.

As for Melara, I switched my workflow to his, and it works like a charm to have the SAME output as I had last night with my own rec2020 workflow. Transformations to larger color spaces are needed to retain the full color information without clipping. Plain rec709 (throwing a clip in the timeline and grading it) does NOT get rid of the various clipping casts.


So, for someone like myself, who doesn't pretend to be anywhere close to the level of any of you guys, what would you suggest be the best workflow to get better colors from the 4k?


I wouldn't take anyone here too seriously -- no professional colorist among us, AFAIK. But with that in mind, foliage greens in the jszfilms footage are uniformly rendered as yellow/browns, something this camera absolutely doesn't do. To my eye, it doesn't look like IR pollution. So who knows what he did or had in front of the lens.

Despite the state of the footage, you can still use the traditional tools, like keys and power windows, to make the greens indistinguishable from those of the S1H. And in this case it takes very little effort. Use ACEs if you want, but there's no need to.

This other idea that you need an exotic workflow to get true color or ideal grades is still fairly rife on the internet, but I don't believe this example makes the case for it.

With minimal adjustments and saturation left high (it's even higher, with RCM) for comparison:

braw_1.4.1.jpg
braw_1.4.1.jpg (805.79 KiB) Viewed 311 times


s1h_1.3.1.jpg
s1h_1.3.1.jpg (749.55 KiB) Viewed 311 times
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 2:52 pm

I tried Eugenia a "Melara" approach last night on some clips that I had to put a lot of work in to, to look right. I achieved much better results, much faster with less corrections needed. The results looked much more natural.
https://www.imdb.com/name/nm0972296/
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 2:54 pm

If you're satisfied, great. But in themselves, these statements don't mean anything, without knowing what did you before, how you're doing it, how you're viewing it, what "natural"means, etc.

And in this field, you can't even take for granted that folks understand that log is a recording and not an exhibition format.

The internet.....
Last edited by John Paines on Tue May 26, 2020 3:09 pm, edited 3 times in total.
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 2:59 pm

John Paines wrote:If you're satisfied, great. But in themselves, these statements don't mean anything, without knowing what did you before, how you're doing it, how you're viewing it, what "natural"means, etc.

The internet.....


Well, what it comes down to is that for me personally, I am far happier with the results that I ended up with, that were much easier achieved with this workflow.
And that's all that matters to me.
https://www.imdb.com/name/nm0972296/
Offline
User avatar

rick.lang

  • Posts: 13396
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:02 pm

The professional colourist’s are in the Resolve forum and there’s been lots of discussion there which can be reviewed including more complex solutions.

There are many solutions as discussed here. Online tutorials are often too quick and too narrow. This is non-destructive change so it seems it does take some harmless experimentation to grade a given clip or project but there isn’t a single magic bullet that solves all issues.

No matter how simple or complicated the work in Resolve may be, the end results must be coloured for the delivery target devices. If you’re only producing output for one target, such as YouTube, then you only have to perfect one grade. If you have a half dozen targets, you have to use more deliverables suited to each. Or you do one for the most important target defined by the client and the others be damned. Realistically the other delivery devices’ playback will almost always be acceptable. I’m colouring for HDTV and if I was ever shooting for cinema, I’d need to do another grade.
Rick Lang
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:10 pm

Henchman wrote:
Well, what it comes down to is that for me personally, I am far happier with the results that I ended up with, that were much easier achieved with this workflow.
And that's all that matters to me.


So industry standards don't mean anything? If a beginning sound guy made an equivalent outlier claim to you, would you rejoice or take it seriously?
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:21 pm

John Paines wrote:
Henchman wrote:
Well, what it comes down to is that for me personally, I am far happier with the results that I ended up with, that were much easier achieved with this workflow.
And that's all that matters to me.


So industry standards don't mean anything? If a beginning sound guy made this claim to you, would you rejoice?


Actually, if a beginning sound guy asked me for my workflow and template, I'd give it to him. As I have done numerous times. I expect he would do what I have done in my career, which is use from it what works for him, and not use what doesn't.
There are a million ways to skin a cat, and as I always say tongue in cheek, when talking workflow with fellow sound guys "there's the wrong way, and there's my way".


In this case, I am taking a workflow, suggested by someone with way more experience than myself, created by someone with way more experience than myself, and trying it.
I found it to be a much better starting point for me, than my original approach.

Really, that's all.
https://www.imdb.com/name/nm0972296/
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:27 pm

Henchman wrote:In this case, I am taking a workflow, suggested by someone with way more experience than myself, created by someone with way more experience than myself, and trying it.
I found it to be a much better starting point for me, than my original approach.


I know I should let it go, but nobody in this thread is a well established professional colorist. This includes Juan Melara. Some have barely done anything at all, judging from what's been presented.

I doubt you'd send aspiring sound guys to amateurs, for advice, and least of all, for unconventional workflows.
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:32 pm

John Paines wrote:
I know I should let it go, but nobody in this thread is a well established professional colorist. This includes Juan Melara.

I doubt you'd send sound guys to amateurs, for advice.


No, I'd give them my advice.
Nobody sent me anywhere.
I simply looked at the approach, listened to the reasoning behind the approach which made sense. And tried it. It worked for me.
It gets me from point A to point B faster, with results that I found more pleasing.
Isn't that the only thing that should matter!
https://www.imdb.com/name/nm0972296/
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 3:44 pm

Henchman wrote:It gets me from point A to point B faster, with results that I found more pleasing.
Isn't that the only thing that should matter!


If you're really asking, my answer is "no". What matters is understanding what you're doing, so you have a basis to make progress and adjust as needed.

It's surprising to hear this "good enough for me" approach from an industry guy, but maybe you don't care about the visual much, I don't know.
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 4:05 pm

John Paines wrote:
Henchman wrote:It gets me from point A to point B faster, with results that I found more pleasing.
Isn't that the only thing that should matter!


If you're really asking, my answer is "no". What matters is understanding what you're doing, so you have a basis to make progress and adjust as needed.

It's surprising to hear this "good enough for me" approach from an industry guy, but maybe you don't care about the visual much, I don't know.


Seeing as I am only doing my own personal projects, and am focused on being able to get the visual I want, then how I get there is pretty irrelevant.
https://www.imdb.com/name/nm0972296/
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 4:24 pm

John Paines wrote:
If you're really asking, my answer is "no". What matters is understanding what you're doing, so you have a basis to make progress and adjust as needed.

It's surprising to hear this "good enough for me" approach from an industry guy, but maybe you don't care about the visual much, I don't know.


When trying his suggestion, what did you find wrong about it?
https://www.imdb.com/name/nm0972296/
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 5:49 pm

Henchman wrote:When trying his suggestion, what did you find wrong about it?


Melara's method, as best I understand it, is a means of preparing log, in preparation of a grade. But it's not an actual grade in itself. The difficulty is that, without that final grade, there's no way to compare his results to an actual grade achieved with RCM, ACEs or normalizing LUTs. How crucial is it, to manipulate log, in the way he proposes, in advance of the grade? No idea....

So I'm waiting for somebody to show me that results obtained with Melara's method can't be more easily achieved with the more straightforward approach, as designed by BMD engineers, for their software and their cameras.

Eugenia's example fails at this in my view. You can get exactly that result, which (sorry) I don't believe is good in itself, any number of ways, without going through multiple transforms, side-effects unknown. It's not even clear to me if her version is presented as a grade or simply altered log, awaiting a grade (which is what it looks like).
Offline

Henchman

  • Posts: 281
  • Joined: Fri Sep 21, 2018 12:43 am
  • Location: Los Angeles
  • Real Name: Mark Hensley

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:03 pm

John Paines wrote:
Henchman wrote:When trying his suggestion, what did you find wrong about it?


Melara's method, as best I understand it, is a means of preparing log, in preparation of a grade. But it's not an actual grade in itself. The difficulty is that, without that final grade, there's no way to compare his results to an actual grade achieved with RCM, ACEs or normalizing LUTs. How crucial is it, to manipulate log, in the way he proposes, in advance of the grade? No idea....

So I'm waiting for somebody to show me that results obtained with Melara's method can't be more easily achieved with the more straightforward approach, as designed by BMD engineers, for their software and their cameras.

Eugenia's example fails at this in my view. You can get exactly that result, which (sorry) I don't believe is good in itself, any number of ways, without going through multiple transforms, side-effects unknown. It's not even clear to me if her version is presented as a grade or simply altered log, awaiting a grade (which is what it looks like).

So, what you're saying is that you're critical of a method, without actually having tried it.
https://www.imdb.com/name/nm0972296/
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:33 pm

ACES and a small selective hue vs hue shift got the shots matched in minutes - it just doesn’t need to be complicated and this is pretty ordinary and easy footage. I took the footage down as it had a person in it and it may not have had permission for use but I can post just the tree footage again if anyone wants to see it? Pro colourist workflow is as much about speed, flexibility and usability in a shared environment as it is about getting the best possible image so just because there are no pro colourists here doesn’t mean much.
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
Offline

Ryan Earl

  • Posts: 115
  • Joined: Sun Jan 17, 2016 6:56 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:42 pm

John Griffin wrote:ACES and a small selective hue vs hue shift got the shots matched in minutes - it just doesn’t need to be complicated and this is pretty ordinary and easy footage. I took the footage down as it had a person in it and it may not have had permission for use but I can post just the tree footage again if anyone wants to see it? Pro colourist workflow is as much about speed, flexibility and usability in a shared environment as it is about getting the best possible image so just because there are no pro colourists here doesn’t mean much.


+1
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 6:59 pm

Henchman wrote:So, what you're saying is that you're critical of a method, without actually having tried it.


I have no view on the method, one way or the other at this point. Detailed analysis could take days and may require more chops than I possess.

What I'm saying is the method is being promoted without evidence of efficacy, by people who haven't demonstrated basic competence in the barest essentials. And/or by persons claiming defects in traditional grading which reflect ignorance, not the methods themselves.
Offline

John Paines

  • Posts: 3565
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 7:03 pm

John Griffin wrote:ACES and a small selective hue vs hue shift got the shots matched in minutes - it just doesn’t need to be complicated and this is pretty ordinary and easy footage. I took the footage down as it had a person in it and it may not have had permission for use but I can post just the tree footage again if anyone wants to see it? Pro colourist workflow is as much about speed, flexibility and usability in a shared environment as it is about getting the best possible image so just because there are no pro colourists here doesn’t mean much.


ACEs is dandy, but why insist on it? Correcting the tree footage is the same, regardless of color management, LUTs, etc.
Offline

John Griffin

  • Posts: 577
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: massive shift in green color on BMPCC4K

PostTue May 26, 2020 7:18 pm

John Paines wrote:
John Griffin wrote:ACES and a small selective hue vs hue shift got the shots matched in minutes - it just doesn’t need to be complicated and this is pretty ordinary and easy footage. I took the footage down as it had a person in it and it may not have had permission for use but I can post just the tree footage again if anyone wants to see it? Pro colourist workflow is as much about speed, flexibility and usability in a shared environment as it is about getting the best possible image so just because there are no pro colourists here doesn’t mean much.


ACEs is dandy, but why insist on it? Correcting the tree footage is the same, regardless of color management, LUTs, etc.

Because it's specifically designed to fix the issue we have here which is getting everyone so worked ie matching multiple camera types on the same timeline.
Which is what I said, and even though this is what I was talking about, there is not too much comparison between your actions at the moment compared to some not around here.
PreviousNext

Return to Cinematography

Who is online

Users browsing this forum: Denny Smith, Harold Tomlinson and 29 guests