UM in Resolve using CST for Alexa color science

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

Gary Yost

  • Posts: 87
  • Joined: Fri Jan 17, 2014 2:03 pm

UM in Resolve using CST for Alexa color science

PostSun Jan 15, 2017 5:03 pm

I just want to share a tip that Jamie LeJeune turned me onto that I'm loving a LOT. If you use the Color Space Transform Effect on a node after doing some basic contrast adjustments, and select Blackmagic 4.6K Film v3 as the Input Color Space, BM Film Log as Input Gamma, ARRI Alexa as Output Color Space, and ARRI LogC as Output Gamma, the effect on skin tones and overall contrast is pretty amazing, getting me almost 90% of the way to a final grade in many cases. Here's an example on a short piece I made, interviewing my intern about his Rubik's Cube experience. (And one my first test of my homemade DIY Interrotron... thank you Errol Morris for inventing this!!) https://www.fastcodesign.com/1663105/errol-morriss-secret-weapon-for-unsettling-interviews-the-interrotron

Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostSun Jan 15, 2017 8:42 pm

That video is awesome Gary! I totally enjoyed it! Really great work!

For those interested in the grading technique, here's a link to download a .drx for import to the Resolve Stills or Powergrade Gallery:
https://drive.google.com/open?id=0B1bnfGXKHqpcb2pRZ0E3TmRVc3M

EDIT: Make sure to read to the bottom of the thread to see the caveats with this approach explained by JB and CaptainHook!

The first node is a Color Space Transform. Second node is a LAB node which allows adjustment of contrast, temp and tint using the RGB Offset controls (this could be swapped for a standard RGB node, I just prefer to make these adjustments in LAB). The third node is the Arri Alex Log C to REC709 LUT. Occaisionally, I add an additional node after the Arri LUT to make subtle adjustments to the deepest shadows or brightest highlights.

I also created a LUT that is virtually the same as the result of that set of nodes (without contrast or temp/tint adjustment) that can be loaded onto the Ursa Mini for accurate monitoring:
https://drive.google.com/open?id=0B1bnfGXKHqpcNGFSQ0F2R1hIbkk

I learned about this technique from Juan Salvo who detailed it in this video:
Last edited by Jamie LeJeune on Tue Jan 17, 2017 11:28 pm, edited 1 time in total.
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline
User avatar

rick.lang

  • Posts: 17173
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: UM in Resolve using CST for Alexa color science

PostSun Jan 15, 2017 9:50 pm

Juan is such a fine colourist. I think he was first to reply to my very first post on the Creative Cow four or five years ago.


Sent from my iPad using Tapatalk
Rick Lang
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 16, 2017 5:03 am

I would be a little bit wary with the Alexa LOG C working 100%.

It's a great cheat, I've done it, but as I've found it can also cause problems. Way back when we were grading the initial first demo footage on the 4.6K we went down a path of using it and it seemed to work great. However, we found there were problems that weren't immediately obvious, especially around super saturated colour and smooth gradients to clipping.

I say whatever works, BUT, when you apply the Arri LOG C transform, you're telling Resolve to expect pictures from a certain sensor and you're then feeding it pictures from another sensor. What can happen is that it creates grading / interpretation issues that you actually can't fix until you take that transform off.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 16, 2017 5:19 am

Thank you for the input and warning John. It is much appreciated.

Just to be clear, the method described above applies the Arri REC709 LUT only after using an OFX Color Space Transform node to change the sensor data from BMD 4.6K Film v3 Log to Arri Log C. Working this way the Arri LUT should be getting the input values that it was designed for.

I have noticed that this workflow can push highly saturated colors like vehicle tail lights into clipping more easily, but it easy to pull back the data before the LUT. In my own testing and grading of the 4.6K, completely clipped values seem to clip anyway no matter what I do in Resolve, Arri LUT or no Arri LUT. (See this recent thread for an example: https://forum.blackmagicdesign.com/viewtopic.php?f=2&t=54377&p=313098&hilit=LED#p313098)

ACES seems to have the same issue with the 4.6K footage too. With highly saturated blues near or past clipping I've found the color space transform + Arri LUT combo actually does a better job rolling off the top of the color channels than even working in ACES does. I'm not at all sure why that is the case.

Have you tested this with the v3 color or just v1?
Also, was the OFX Color Space Transform even available in Resolve when the Ursa Mini 4.6K test footage was released?
I'd be curious to know more about the situations where you ran into problems with the Color Space Transform + Arri LUT combination.

Many thanks!
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

CaptainHook

Blackmagic Design

  • Posts: 2054
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 16, 2017 11:21 pm

Jamie LeJeune wrote:Just to be clear, the method described above applies the Arri REC709 LUT only after using an OFX Color Space Transform node to change the sensor data from BMD 4.6K Film v3 Log to Arri Log C. Working this way the Arri LUT should be getting the input values that it was designed for.

Yes and no. It may be transformed into the appropriate gamut/gamma, but it still doesn't mean that individual colours are as the Arri 709 LUT was designed for or as it expects in terms of hue or saturation.

When you transform from 4.6K Film to Arri Log C, the transform doesn't know the spectral sensitivity or response curves of the 4.6K sensor or the Arri sensor. So it could be the Arri sensor may respond to say Red completely differently to the 4.6K sensor and the transform will NOT account for that. Arri Log C is designed around how the Arri sensor responds to wavelengths and 4.6K Film is designed around the 4.6K sensor obviously. Using the OFX plugin also doesn't take into account the WB settings which is needed to transform properly back into sensor RGB space. In RAW with RCM this is taken into account however. But then the mixing matrix used for each log space is specific to the sensor anyway, so giving them a different input (sensor) will again lead to somewhat unpredictable results in colour. The log curve might be "okay" however.

Long story short, the transform will not make 4.6K colours the same as Alexa or vice versa. More so, where colours end up is somewhat arbitrary. It could even be the case the transform makes the 4.6K colour LESS similar to Alexa colour and actually make it harder to match to Alexa footage if that's your goal.

Now, if you like the results is of course a different story and down to preference.
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostTue Jan 17, 2017 9:25 pm

CaptainHook wrote:Yes and no. It may be transformed into the appropriate gamut/gamma, but it still doesn't mean that individual colours are as the Arri 709 LUT was designed for or as it expects in terms of hue or saturation.

Long story short, the transform will not make 4.6K colours the same as Alexa or vice versa. More so, where colours end up is somewhat arbitrary. It could even be the case the transform makes the 4.6K colour LESS similar to Alexa colour and actually make it harder to match to Alexa footage if that's your goal.

Now, if you like the results is of course a different story and down to preference.


Thank you very much for the detailed reply CaptainHook!

I'm not trying to match an Alexa with this approach (but really do wish I had the budget that it was even an issue). In my grading of the 4.6K over the last few months, for a surprising amount of my footage the CST+Arri REC709 LUT yields better results to my eye than even ACES.

A related question: Why does BMD's own Ursa Mini 4.6K v3 to REC709 LUT appear not to be optimized for higher dynamic range scenes?
The CST+Arri LUT and ACES both roll off the highlights really nicely, while the standard v3 BMD LUT clips them quite harshly. Although the highlights can be pulled back in a node before applying the BMD LUT, the work it takes to reign in saturation and also ensure that the midtones aren't affected by the highlight range compression seems more trouble than it is worth to use the official BMD 4.6K v3 to REC709 LUT. RCM also appears to use an identical transform when the output color space is set to REC709.

Can you provide some context as to the reasons that BMD decided to go with a REC709 transform that is difficult to use for the high dynamic range scenes the 4.6K sensor was built to handle?

Please excuse my ignorance if these are issues that have been explained in detail before.

Many thanks!
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostTue Jan 17, 2017 9:39 pm

Jamie LeJeune wrote:Thank you for the input and warning John. It is much appreciated.

Just to be clear, the method described above applies the Arri REC709 LUT only after using an OFX Color Space Transform node to change the sensor data from BMD 4.6K Film v3 Log to Arri Log C.


The problem I believe is that this assumption does not work when the Alexa sensor generates a different colour matrix to the Ursa Mini.

You're applying a transform to Alexa footage that isn't coming from an Alexa. So that's where the errors come from.

So you will see errors. As mentioned in super saturated colours. I noticed it around deep LED red tail lights, around super blue NEON signs. But it also did some strange things around clipping. There would be a band of grey around the clipped highlights that wasn't very nice and couldn't be rekeyed or adjusted away until you took that arri transform out of the mix.

There are many who used to like the Blackmagic LUT for the Digital Bolex and got some nice looking pictures but they ended up doing their own LUT for the same reasons.

So use it if it works for you, but know that it's not optimised and it doesn't "make" your camera look like an Alexa. I get better results matching by NOT doing this because you're using the full range available, not jamming it through a mathematical transform it's not designed or optimised for.

It's a cheat, that superficially works, but you need to also know that it's not always going to work perfectly and you could do better by doing the legwork up front. It takes more setup time, but you'll get a better result.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostTue Jan 17, 2017 11:24 pm

John Brawley wrote:
Jamie LeJeune wrote:Thank you for the input and warning John. It is much appreciated.

Just to be clear, the method described above applies the Arri REC709 LUT only after using an OFX Color Space Transform node to change the sensor data from BMD 4.6K Film v3 Log to Arri Log C.


The problem I believe is that this assumption does not work when the Alexa sensor generates a different colour matrix to the Ursa Mini.

You're applying a transform to Alexa footage that isn't coming from an Alexa. So that's where the errors come from.

So you will see errors. As mentioned in super saturated colours. I noticed it around deep LED red tail lights, around super blue NEON signs. But it also did some strange things around clipping. There would be a band of grey around the clipped highlights that wasn't very nice and couldn't be rekeyed or adjusted away until you took that arri transform out of the mix.

There are many who used to like the Blackmagic LUT for the Digital Bolex and got some nice looking pictures but they ended up doing their own LUT for the same reasons.

So use it if it works for you, but know that it's not optimised and it doesn't "make" your camera look like an Alexa. I get better results matching by NOT doing this because you're using the full range available, not jamming it through a mathematical transform it's not designed or optimised for.

It's a cheat, that superficially works, but you need to also know that it's not always going to work perfectly and you could do better by doing the legwork up front. It takes more setup time, but you'll get a better result.

JB


Thank you for the detailed and clear answer JB. Yes, it makes total sense and matches the response from CaptainHook as well.

Although the results I've gotten have been surprisingly good for many shots, when I have a project that budgets for an Alexa (someday) I'll plan to shoot color charts and scenes with it next to my 4.6K to create some accurate grades without the color space transform.
Or, if you or CaptainHook have any you'd like to share... :D
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostWed Jan 18, 2017 4:12 am

Jamie LeJeune wrote:
Thank you for the detailed and clear answer JB. Yes, it makes total sense and matches the response from CaptainHook as well.

Although the results I've gotten have been surprisingly good for many shots, when I have a project that budgets for an Alexa (someday) I'll plan to shoot color charts and scenes with it next to my 4.6K to create some accurate grades without the color space transform.
Or, if you or CaptainHook have any you'd like to share... :D


I can't easily access any right now that wouldn't upset people.

I run into this often. Very experienced colourists do exactly the same thing you're doing. But, once you can wean yourself from using this approach you will have better skin tones, less super satuation.

I speak from painful personal experience. Just try building your grade from scratch. It will take longer and frustrate you more, but the end result will be better.

Blackmagic could do a lot more to shine a light on the best way of doing this and I know there's new refinements coming to help address some of this. I wish I was better at Resolve and I wish it was simpler in Resolve to get these better results so I can more directly show you, but Hook is 1000% right on this stuff.

He's being polite, but it's just silly to use an Arri transform on an Ursa Mini even if it seems to work. You're ending up limiting what you can do with the grade.

Someone once said to me it's like playing Russian roulette with your grading.

And I would say this.

This camera is a hell of a lot better when you stop trying to make it look like an Alexa, and just focus on making the best of it's native look.

I just show an entire series with them and it's so liberating to not have to be twisting the grade to match another camera all the time. And BM's innate "look" is very very nice. And a lot of that is Hooks DNA....

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

rick.lang

  • Posts: 17173
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: UM in Resolve using CST for Alexa color science

PostWed Jan 18, 2017 6:18 pm

John Brawley wrote:... Blackmagic could do a lot more to shine a light on the best way of doing this and I know there's new refinements coming to help address some of this. I wish I was better at Resolve and I wish it was simpler in Resolve to get these better results so I can more directly show you, but Hook is 1000% right on this stuff...

BM's innate "look" is very very nice. And a lot of that is Hooks DNA....

JB


Thanks for this post.


Sent from my iPad using Tapatalk
Rick Lang
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 1:05 am

Thank you John. I fully trust in your experience and I appreciate your advice to build a grade from scratch.

I realize that you and Hook are correct. As I look my grades with and without the transform it is true that CST + Arri LUT adds a noticeable amount of noise to the 4.6K signal (especially in the blue channel) that isn't apparent if I grade without them.

I haven't graded enough Alexa footage to even be able to claim I know what the Alexa should look like, but I have figured out what I liked about the effect that the CST + Arri LUT had on the 4.6K image and have been able to distill it down to a few hue curve adjustments -- no LUT no color space transform needed.

The 4.6K v3 color makes lovely skin tones, but to my eye it also bends blue toward magenta in such a way that (at least on the images from my 4.6K) I too often seem to be fighting magenta tints in the shadows and highlights and often ended up dragging the tint of the whole scene too far toward green to compensate. The effect of the color space transform plus Arri REC709LUT that I liked was that it pulled purple-blue just enough toward true blue to eliminate the tint I was seeing in the shadows and highlights. I've found (thanks to your encouragement) that with Resolve's hue curves I can do essentially the same thing, but without the problems introduced by the CST + Arri LUT. These hue curves may not be good for all scenes, but it's a start and it if does cause issues it can be adjusted as needed, unlike a LUT.

Once I've had more time to test it and add an appropriate contrast adjustment, I'll post a .drx file and a matching monitoring LUT to this thread for anyone who is curious to test it out.

And I'm very much looking forward to the "refinements" you mentioned. Can't wait for their release!
Last edited by Jamie LeJeune on Thu Jan 19, 2017 9:03 pm, edited 1 time in total.
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

James Parker

  • Posts: 184
  • Joined: Sun Jul 13, 2014 3:38 pm

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 1:34 am

Are you shooting everything at zero tint on all white balances. I started that from day 1 of 4.0beta and haven't gone back since.


Sent from my iPhone using Tapatalk
J.Parker
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 2:00 am

James Parker wrote:Are you shooting everything at zero tint on all white balances. I started that from day 1 of 4.0beta and haven't gone back since.


The ideal tint setting for any situation depends on the light source (and reflected light) and your particular 4.6K sensor. I've used the exact same lens with the exact same temp and tint settings on two Ursa Mini 4.6K shooting the same chart in the same light and when I pulled the images into Resolve one sensor had a green tint while the other had a magenta tint.

In any case, I'm not talking about something that can be tuned out using the overall tint control. What I mean is that on the 4.6K v3 color, a blue chip from a color chart leans more toward magenta than it did on the 4.6K v1 color. See this thread for an extreme example of that difference with LED light:
https://forum.blackmagicdesign.com/viewtopic.php?f=2&t=54377
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

James Parker

  • Posts: 184
  • Joined: Sun Jul 13, 2014 3:38 pm

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 2:43 am

I'll give it a look


Sent from my iPhone using Tapatalk
J.Parker
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 3:09 am

Jamie LeJeune wrote: I've used the exact same lens with the exact same temp and tint settings on two Ursa Mini 4.6K shooting the same chart in the same light and when I pulled the images into Resolve one sensor had a green tint while the other had a magenta tint.



Yeah I've found there can sometimes be quite a bit of variation. I had three cameras on my last show. When we checked the white balance alignment, one was +10, second was - 30 and the third was -23 to make them all the same. Those offsets where the same up and down the WB presets.

(Alexa's do this do, but maybe not as much variation)

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 6:53 am

It's really good to hear that the offsets hold with different kelvin settings!
I didn't get that far as we just had to match two cameras for a single interview setup and tested only daylight. They were not too far off, about a dozen points between them.
I've never had multiple bodies of the same model camera by any manufacturer match out the box, so I expected that they would be different.

Have you noticed any change in the native tint of a particular 4.6K after recalibrating on the new firmware?
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 3:31 pm

CaptainHook wrote:
Jamie LeJeune wrote:Just to be clear, the method described above applies the Arri REC709 LUT only after using an OFX Color Space Transform node to change the sensor data from BMD 4.6K Film v3 Log to Arri Log C. Working this way the Arri LUT should be getting the input values that it was designed for.

Yes and no. It may be transformed into the appropriate gamut/gamma, but it still doesn't mean that individual colours are as the Arri 709 LUT was designed for or as it expects in terms of hue or saturation.

When you transform from 4.6K Film to Arri Log C, the transform doesn't know the spectral sensitivity or response curves of the 4.6K sensor or the Arri sensor. So it could be the Arri sensor may respond to say Red completely differently to the 4.6K sensor and the transform will NOT account for that. Arri Log C is designed around how the Arri sensor responds to wavelengths and 4.6K Film is designed around the 4.6K sensor obviously. Using the OFX plugin also doesn't take into account the WB settings which is needed to transform properly back into sensor RGB space. In RAW with RCM this is taken into account however. But then the mixing matrix used for each log space is specific to the sensor anyway, so giving them a different input (sensor) will again lead to somewhat unpredictable results in colour. The log curve might be "okay" however.

Long story short, the transform will not make 4.6K colours the same as Alexa or vice versa. More so, where colours end up is somewhat arbitrary. It could even be the case the transform makes the 4.6K colour LESS similar to Alexa colour and actually make it harder to match to Alexa footage if that's your goal.

Now, if you like the results is of course a different story and down to preference.


i disagree. the color in a file once debayered are simply values that represent a gamut color space identifies by the RGB primaries and white point plus a tonal mapping that is linked to the original dynamic range of the camera.

each camera has a white paper (published or not) that describe how those values are spread and what they correspond in scene referred liner light captured.

the resolve ofx for the most part does:

invert the tonal range back to liner light
map the color gamut to xyz
map the xyz to the desired net colour gamut
apply the new tonal range curve.

unless the bm camera is vastly superior in terms of gamut/tonal range tab an alexa (highly unlikely) the map works just fine.

that kind of transformation is the cornerstone of ACES.
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

CaptainHook

Blackmagic Design

  • Posts: 2054
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 9:01 pm

waltervolpatto wrote:i disagree.

So you are saying the OFX plugin or ACES will match the colour of the 4.6K to the Alexa?
What i said above is it won't, and I've tried both.
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 10:38 pm

CaptainHook wrote:
waltervolpatto wrote:i disagree.

So you are saying the OFX plugin or ACES will match the colour of the 4.6K to the Alexa?
What i said above is it won't, and I've tried both.


the math will convert the color space to be used under the Alexa/log workflow and it will match under a wide range of lighting situation.

if a camera sensor has a "different" spectral response, well, that is s characteristic of the camera, or how the manufacture decide to interpret the raw and transform it to a colour space suitable for color pipeline.

is the same between red and alexa, however, proper lighting and exposure of the sensor will yield almost indistinguishable images once the ofx is applied.

and yes. i do that for my work. no "magic lut" necessary.
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline
User avatar

Dmytro Shijan

  • Posts: 1760
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 19, 2017 11:51 pm

I also can confirm that there is no universal gamut or log curve for all cameras. LUTs or color space transform node convert camera gamut/color space to working gamut/color space mathematically correct, but when you start to grade it the colors and gamma will behave very different with different combinations of camera to timeline gamut/color spaces.
Some color spaces and log curves are more optimized for camera sensors response, some build as more universal solutions.
Here is example when i develop scanned negative films converted from scanner input profile: Sony S-gamut3cine feels abnormally warm, Rec2020 shifts yellows too green, Sony S-gamut feels ok. Same difference occurs with log curves.
BMMCC/BMMSC Rigs Collection https://bmmccrigs.tumblr.com
My custom made accessories for BMMCC/BMMSC https://lavky.com/radioproektor/
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 20, 2017 1:43 am

waltervolpatto wrote:
if a camera sensor has a "different" spectral response, well, that is s characteristic of the camera, or how the manufacture decide to interpret the raw and transform it to a colour space suitable for color pipeline.



I think we're getting beyond the point.

I know and have seen when you apply a REC 709 Alexa LUT to Blackmagic Ursa Mini 4.6K raw footage errors that can't be corrected without removing the LUT.

If the way the sensor performs is different to what the LUT is expecting then we have a problem no ?

That's what we're discussing. I have seen this occur on many occasions in post facilities big and small.

It's easily provable. You can download the freely available shots for yourself. Even the proponent of this approach concedes going back to his footage that they are seeing a difference.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 20, 2017 4:00 am

Just to share what I saw, below are three stills from Ursa Mini 4.6K footage captured to HD ProRes444 12bit.
It's an awful shot that was never meant for anything, but I chose it because it includes a lot of blue items that illustrate well the issue I finally noticed after JB pointed it out.

UM46_Original.jpg
UM46_Original.jpg (408.07 KiB) Viewed 15393 times

UM46_CST_plus_Arri_LUT.jpg
UM46_CST_plus_Arri_LUT.jpg (485.35 KiB) Viewed 15393 times

UM46_Graded_to_Match.jpg
UM46_Graded_to_Match.jpg (493.17 KiB) Viewed 15385 times


In the matching grade, beyond setting contrast and saturation, I adjusted the hue v hue and hue v sat curves to get as close as possible to the look of the Color Space Transdform + Arri REC709 LUT combo. The real noticable difference is the luminance of blue. The only way I could see to match that exactly would have been to adjust the hue v lum curve, but that added a lot of noise to the image so I didn't do it. Interestingly, the Color Space Transdform + Arri REC709 LUT combo also adds noise to the image. So, it seems to me that the Color Space Transdform + Arri REC709 LUT combo is doing things to the image that I'd otherwise not like and would not choose to do in a manual grade.

Is this the problem with the Color Space Transdform + Arri REC709 LUT combo that you are describing JB?

If you'd like to see it motion so that the noise is apparent, here's a link to download the original file and graded clips, plus .drx files of the grades to load into Resolve to test on other Ursa Mini 4.6K Film v3 files.

https://drive.google.com/open?id=0B1bnfGXKHqpcdzhHdnM4WGZtZjA
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 23, 2017 5:56 am

Just to share what I saw, below are three stills from Ursa Mini 4.6K footage captured to HD ProRes444 12bit.
It's an awful shot that was never meant for anything, but I chose it because it includes a lot of blue items that illustrate well the issue I finally noticed after JB pointed it out.
In the matching grade, beyond setting contrast and saturation, I adjusted the hue v hue and hue v sat curves to get as close as possible to the look of the Color Space Transdform + Arri REC709 LUT combo. The real noticable difference is the luminance of blue. The only way I could see to match that exactly would have been to adjust the hue v lum curve, but that added a lot of noise to the image so I didn't do it. Interestingly, the Color Space Transdform + Arri REC709 LUT combo also adds noise to the image. So, it seems to me that the Color Space Transdform + Arri REC709 LUT combo is doing things to the image that I'd otherwise not like and would not choose to do in a manual grade.
Is this the problem with the Color Space Transdform + Arri REC709 LUT combo that you are describing JB?
If you'd like to see it motion so that the noise is apparent, here's a link to download the original file and graded clips, plus .drx files of the grades to load into Resolve to test on other Ursa Mini 4.6K Film v3 files.



Ok, i finally have some time to look at the shot. nice raining day both outdoor here in LA and in the image.

no evident problems in the raw. a bit on the contrast side, no clip of sort, no breaking of color, on the saturated side: I'm not used to the tonal distribution of the Ursa, so I assume that it is fine.

First test, do the transformation from [blackmagic design film v3/film log] to [709 gamma 2.4].

the result is pleasing, I'm noticing a clumps of color in the cyan side (straight line) and it is the cyan in the window. very high and saturated. It looks like that particular shade of cyan is hitting the boundaries of the color space. I open the log image (offset) by 90, looking at the waveform: indeed the math create two lobs in the R and B channel that go under [0] line. nothing unexpected for pure math.

I add another color transformation to see if it is just a 709 limit or the data is really clipped somehow: [709 gamma 2.4]to [blackmagic design film v3/film log] and i get the original image with no clip/crush/loss. The math works (as expected).

BM LUT; signal log straight to BM LUT: this will map the tonal response as for the recommended BM workflow. No clipping, nice tonal curve, good colors, nice toe and shoulder, no noise in the image.

This, is the BM color science at it's minimum:
sensor capture light
[AD] --> raw container (or log container, whichever is used in this case)
apply the recommended tonal curve from BM.

Nothing strange, this is how BM decided that the camera should look and respond to light plus how the final image should be mapped to the display device, both mathematically and artistically (the toe/shoulder part.)

Now, i will transform the BM to the alexa using the color transform and to 709 as for the other workflow:

[blackmagic design film v3/film log] to [alexa wide/log]+ another node+[alexa wide/log] to [709 gamma 2.4].

This will look identical to the [blackmagic design film v3/film log] to [709 gamma 2.4] done earlier.

In fact if i add the [709 gamma 2.4]to [blackmagic design film v3/film log] from there I get the original image.

Then i will do what the original idea was:
[blackmagic design film v3/film log] to [alexa wide/log]+ color corrector+alexa lut (the standard provided by arri): the image looks nice and smooth, no issue. The workflow works. I will call this [alexa lut] for short

Now, does the [alexa] lut match the [BM lut]? No. and it should not be a surprise.

If you decide to use the [ALEXA LUT] workflow, what you're trying to say is: if i have an alexa camera and i point it to the image, using the standard ALEXA color science (and how they decide to map it to the final device) it will look like that.
If someone has an alexa and a ursa, you can try to put it side by side and do the double check, when both are used at the optima setting, similar lenses, you can do either:

Ursa->color->UrsaLUT
and you will use the BM device representation, in this case you will want
ALEXA->color trasform to URSA log-> color-Ursa LUT
will make the ALEA behave like an ursa.

The opposite if you want to mimic the alexa workflow:
ALEXA->color->alexa lut
Ursa->color trasform to ALEXA log-> color-alexa lut.
and both will behave as for the Alexa color science.

w.
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 23, 2017 5:58 am

BTW, thanks for providing the Image for testing!!
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostTue Jan 24, 2017 12:39 am

waltervolpatto wrote:BTW, thanks for providing the Image for testing!!


You're welcome. Thank you for your detailed investigation!

waltervolpatto wrote:
BM LUT; signal log straight to BM LUT: this will map the tonal response as for the recommended BM workflow. No clipping, nice tonal curve, good colors, nice toe and shoulder, no noise in the image.

Nothing strange, this is how BM decided that the camera should look and respond to light plus how the final image should be mapped to the display device, both mathematically and artistically (the toe/shoulder part.)


Which BM LUT are you referring to here? The included BMD LUT for 4.6K Film v3 to REC709?
When I apply that LUT, the result is pretty garish with severely clipped and oversaturated highlights. Here's what it looks like applied to the test image:
BMD_LUT.jpg
BMD_LUT.jpg (400.76 KiB) Viewed 15261 times
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

Juan Salvo

  • Posts: 343
  • Joined: Wed Aug 22, 2012 4:22 am

Re: UM in Resolve using CST for Alexa color science

PostThu Jan 18, 2018 11:23 pm

I don't want to totally rehash this thread, I just saw the comment on my tut from Jamie which brought me here. From my skim of the discussion, I just want to say that my view is that of Walter. The characteristics of the sensor are folded into the image, they don't affect the way the colorimetry is transformed from one colorspace to another. Will the two images always behave exactly the same? No. As John pointed out you may hit walls, particularly in color beyond the gamut of the target space. But that's a limit of the camera not the color science. You'd have the same issues using BMDs LUTs. In fact, ARRI's gamut mapping in their LUT is quite elegant. My demo was also not just meant to apply to converting to ARRI space, but could be used for others. I'm glad you guys found it useful.
Offline

CaptainHook

Blackmagic Design

  • Posts: 2054
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 1:28 am

Juan Salvo wrote:The characteristics of the sensor are folded into the image, they don't affect the way the colorimetry is transformed from one colorspace to another.

I've said it before, but the colourspace transform plugin assumes a 6000K illuminant. For lower colour temps like 3200K, the matrices used will not be as accurate. In my testing comparing the transforms using correct matrices the difference is large enough that I would caution using the plugin at temps lower than 5000K unless an accurate transform is not of concern (for most colourists it seems its not). But the spectral response of different sensors under various illuminants in relation to the colourspace transform plugin is a factor. Like how the idea that ACES can make all cameras behave the same despite spectral response variation is flawed. Walter makes a fair assumption and does great work, but all the colour scientists I've talked to including Arri's head colour scientist who did the Alexa colour say the same thing and I will defer to their opinion on this.

You essentially only have 6 coefficients in a matrix you can use to map the spectral response of a sensor to training data (since the others must be changed to sum to 1), so no matter how good the fit or the weightings you do, how a sensor responds to a colour under a specific illuminant will very likely not map to the same XYZ/scene value as another sensors spectral response will - even using the same approach to calculate a fit. So the colours from two different sensors will not have the same XYZ/scene value, and not reproduce the same colour transformed to the same target space. They may get "close", but there will be variation and that variation will also change across illuminant white points. CST always assumes 6000K.

If you use the CST plugin to transform BMDFilm 4.6K footage shot under 3200K for example to Alexa, and then apply the Alexa LUT - the result will not match the Alexa shot in same conditions with the same processing - especially not as well as it could using the appropriate matrices. But do you like it? That's another story.

Juan Salvo wrote:As John pointed out you may hit walls, particularly in color beyond the gamut of the target space. But that's a limit of the camera not the color science.

If a colour is beyond the target space but available from the camera, then its a limit of the target space and the transform into it. Why do you say it's the camera limit if the transform is not happening in camera? ACES has that issue currently with Alexa etc and the issue is NOT the camera(s).

http://acescentral.com/t/colour-artefac ... aces/520/1

"Are there any ‘gotchas’ when using ARRI cameras in an ACES workflow? Do they have workarounds?"
http://www.arri.com/camera/alexa_mini/l ... with_aces/
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 3:54 am

've said it before, but the colourspace transform plugin assumes a 6000K illuminant.


Are you talking about the resolve ofx? Because i don't think so.
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

Juan Salvo

  • Posts: 343
  • Joined: Wed Aug 22, 2012 4:22 am

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 4:10 am

Like how the idea that ACES can make all cameras behave the same despite spectral response variation is flawed.

I'm not saying that, and Walter isn't saying that. What I am saying and what is the case is one can use the transforms to have the image behave in the color controls the same way. And given that a portion of the ARRI look.. the ARRI photometric matrix, is folded in to the LUT, that can be carried over as well. Again, I'm not saying and Walter isn't saying that makes the sensors limits a non-issue. Using ARRI luts on GoPro won't make the image 12bits.

If a colour is beyond the target space but available from the camera, then its a limit of the target space and the transform into it. Why do you say it's the camera limit if the transform is not happening in camera?

Well it depends on the colorimetry of the source and target color space as well. But the attenuations of the specific color filters in the cameras CFA have a kind of decisive impact on how one handles out of gamut colors, even more so as they approach or exceed the sensors.

The highlight clipping/fringing exhibited in ARRI's example is a part of the ARRI Photometric transform, which again, is folded into their lut. Many other cameras exhibit the same issues. I've laid off ACES as much as I can, since RCM does almost all of what aces would be doing for me, with far more control. But I believe the later versions of their RRT actually adopted a very similar solution to that problem and now behaves rather similarly.
Offline

Juan Salvo

  • Posts: 343
  • Joined: Wed Aug 22, 2012 4:22 am

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 4:18 am

waltervolpatto wrote:
've said it before, but the colourspace transform plugin assumes a 6000K illuminant.


Are you talking about the resolve ofx? Because i don't think so.

I don't think so either. I suspect Captain was comparing RAW media with log media, where the image had been normalized to 6000K vs one where it could be normalized to the target temp. I imagine that in that rather narrow use case, you might see a difference. Of course you could just normalize the RAW footage to 6000K as well and then transform it, in which case it would behave very similarly.

The CFA's color densites, can and will cause difference in the camera performance at extremes. That is some cameras just have denser filters than others, and will capture extreme saturations with more or less fidelity. An Alexa shooting a car's tail light vs a RED Epic shooting the same scene, the alexa will capture more detail and have better roll off than the Epic, because the Alexa's CFA is rather thick compared to the thinner CFA of the Epic. If we shot a scene under a red lamp, at say 1700K, and then normalized that, the Alexa would retain far more detail than the Epic. This is simply because the cameras are physically different, has nothing to do with the CST.

But I'm sure you already know all that Walter. :)
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 5:40 pm

yes, give me an alexa any day.....
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

Howard Roll

  • Posts: 2506
  • Joined: Fri Jun 03, 2016 7:50 am

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 8:02 pm

The irony of this conversation, while the kid in the video is literally performing a 3x3x3 CST, is priceless.
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 8:07 pm

With the greatest of respect,

I see two colourists arguing with one of the lead developers of the Blackmagic camera team.

I’ve heard this “thinner” CFA story before, but I’ve never ever seen it empirically proven anywhere. Other than being an old wives tale, is there anything that shows conclusively that this is the case ? That Alexa has a “thicker” CFA and RED doesn’t ?

I have seen some evidence from manufactures like Digital Bolex that a more “pure” CFA (not thicker) is one of the reasons their camera had reduced sensitivity but a better spectral response. Or maybe it’s just the legendary Kodak Sesnor.

Alexa has been ISO 800 since day one, and RED until very recently was ISO 320, though they often say to expose at 800 to protect the highlights. That’s the opposite outcome if Alexa has a thicker CFA to justify the view that it’s colour is “better”.

Sony also have been know to adjust (and market ) their CFA purity / chromacity on their higher end cameras (wide gamut) vs their lower end cameras. See F55 / F65.

I call BS on this thicker CFA theory. Unless someone can point me to actual empirical evidence other than hearsay ? It’s a distraction otherwise to describe something that’s not known.

I think what Hook was saying was that the CST assumes that your subject is lit by a 6000K light source.

His view comes from his first hand knowledge of discussions with his cited source, Arri themselves.

He also deals daily with the Resolve team. Are we really doubting this opinion ?

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 9:05 pm

John Brawley wrote:With the greatest of respect,

I see two colourists arguing with one of the lead developers of the Blackmagic camera team.

I’ve heard this “thinner” CFA story before, but I’ve never ever seen it empirically proven anywhere. Other than being an old wives tale, is there anything that shows conclusively that this is the case ? That Alexa has a “thicker” CFA and RED doesn’t ?

I have seen some evidence from manufactures like Digital Bolex that a more “pure” CFA (not thicker) is one of the reasons their camera had reduced sensitivity but a better spectral response. Or maybe it’s just the legendary Kodak Sesnor.

Alexa has been ISO 800 since day one, and RED until very recently was ISO 320, though they often say to expose at 800 to protect the highlights. That’s the opposite outcome if Alexa has a thicker CFA to justify the view that it’s colour is “better”.

Sony also have been know to adjust (and market ) their CFA purity / chromacity on their higher end cameras (wide gamut) vs their lower end cameras. See F55 / F65.

I call BS on this thicker CFA theory. Unless someone can point me to actual empirical evidence other than hearsay ? It’s a distraction otherwise to describe something that’s not known.

I think what Hook was saying was that the CST assumes that your subject is lit by a 6000K light source.

His view comes from his first hand knowledge of discussions with his cited source, Arri themselves.

He also deals daily with the Resolve team. Are we really doubting this opinion ?

JB


if I can make a light joke, i cannot trust anything that come from someone called "CapitanHook"... no more than something that came from anybody called "PeterPan", or "SnowWhite"....

-w.
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

Juan Salvo

  • Posts: 343
  • Joined: Wed Aug 22, 2012 4:22 am

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 9:45 pm

John Brawley wrote:that a more “pure” CFA (not thicker) is one of the reasons their camera had reduced sensitivity but a better spectral response. Or maybe it’s just the legendary Kodak Sesnor.

Tomato, tomato. (I said one of those in a british and in an an american accent).

We're describing the same thing, the "thickness" or "purity" describes how narrow the spectral attenuation of the filter is. We're all saying the same thing.

John Brawley wrote:Alexa has been ISO 800 since day one, and RED until very recently was ISO 320

There's so much being conflated here.

1 thing I want to clarify. I pointed out Alexa and RED Epic (and specifically Epic) because those two are just an easily available example with a large difference. But same could apply to number of other cameras. The newer RED cameras, which have improve photosite sensity now use thicker CFAs. There's a very strong correlation between those two things.

John Brawley wrote:I call BS on this thicker CFA theory. Unless someone can point me to actual empirical evidence other than hearsay ? It’s a distraction otherwise to describe something that’s not known.

Ok, we've done tests in house that show this, others have too. But honestly not that worried about winning over the confidence of one person. Feel free to call BS. This is like flat earth theory now.
Offline
User avatar

Jamie LeJeune

  • Posts: 2012
  • Joined: Tue Feb 26, 2013 4:33 am
  • Location: San Francisco

Re: UM in Resolve using CST for Alexa color science

PostFri Jan 19, 2018 10:07 pm

I'm thoroughly enjoying the discussion. Thank you all for taking time for this.

Please forgive my ignorance, but so far this is what I understand:

1. The Color Space Transform in Resolve assumes a 6000K input. I assume it could have been built to adjust for different inputs set in the OFX panel, but for whatever reason BMD didn't build it that way and designed it to default to 6000K input.

2. That being the case, if one were to use the CST as Juan showed in his video, the most correct way to use it with raw footage would be to set the temp to 6000K in the raw tab, apply the CST to get to Alexa C-Log and AWG on the first node, in the second node balance the color using LGG and then in a third node apply the REC709 LUT. With plain log footage, this isn't possible and one just has to accept the lack of accuracy in the transform for shots that weren't captured at 6000K.

3. The other limitation to using the CST are the limits imposed by the different Color Filter Array on each camera. Some cameras may capture colors outside the gamut of the CST. Or, is the other way around?

Is this somewhat accurate? Or have I totally misunderstood what you are all saying here?
www.cinedocs.com
http://www.imdb.com/name/nm4601572/
Offline

Luca Di Gioacchino

  • Posts: 160
  • Joined: Sat Oct 11, 2014 4:23 pm
  • Location: Montreal, Canada

Re: UM in Resolve using CST for Alexa color science

PostSat Jan 20, 2018 12:40 am

Sorry to go off topic, but I thought we were supposed to use our real names on the forums (*coughCaptainHookcough*) ;)
Offline

CaptainHook

Blackmagic Design

  • Posts: 2054
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: UM in Resolve using CST for Alexa color science

PostSat Jan 20, 2018 12:53 am

waltervolpatto wrote:Are you talking about the resolve ofx? Because i don't think so.

Yes, the ResolveFX plugin. The guy who wrote the code told me and showed me the code, so I'm inclined to believe him. If you use RAW with RCM then it's not based on this fixed assumption, but with the ResolveFX plugin it is.

QE (quantum efficiency) and spectral response is affected by CFAs and micro lens design along with pixel size and many other interacting elements that you must design as a system (usually balancing trade-offs), yes. But when I speak of spectral response in relation to the topic here, I'm talking less about wavelengths at the extremes of most display gamuts and more about the shape and sensitivity of the crossovers between the LMS wavelengths, or how those colours in-between red/blue/green are captured and how they can be mixed to reproduce the "in-between" colours. No two sensors respond to all colours the same way, and the type of transforms used are not accurate enough to account for the differences.

Juan Salvo wrote:how one handles out of gamut colors, even more so as they approach or exceed the sensors.

Yes, how one handles it is the issue. The camera is only the issue if it clips it away at some point or clips it in its own transforms baked into ProRes for example, but with RAW it's down to how you deal with it, so not a camera issue since you can always convert out of a manufacturers log/wide gamut into a wider space if a colour clips in their provided gamut. A typical sensors (and CFA etc) spectral response is wider than the visible spectrum so all colours we can see are captured. Most camera wide gamuts I've looked at cover the fully blue monochromatic light sources, red is less common and green even less but generally these blue lights are even covered in ProRes. But these fully monochromatic light sources are an issue for us in involved in image capture/display in more ways than this, and are more and more common.

Jamie LeJeune wrote:Is this somewhat accurate? Or have I totally misunderstood what you are all saying here?

I would say you have understood everything pretty well which is great as you can choose to use these tools or Juans technique from a more informed perspective which was my goal here. Just with #3 again it's not really out of gamut colours I was talking about in regards to spectral response, more to the point no two different sensors see colours the same way and the transforms back to XYZ/scene provided by all manufacturers cannot fully account for those differences. So you will never get two cameras reproducing all colours the same by using those types of transforms. They could be "close enough" for most though. In the end whatever works for you is the decider.

Years of working one way with a particular workflow based on one type of input can be hard to change, and most will fight to stay that way and produce evidence as to why it's "better", instead of just different. At the same time, learning and developing a larger/wider toolbox can be expensive in cost and time, so the choice for many could indeed be to stick to what they know and try to change anything different into what they are used to. It's human nature.

Luca Di Gioacchino wrote:Sorry to go off topic, but I thought we were supposed to use our real names on the forums (*coughCaptainHookcough*) ;)

Real names are about being accountable and knowing who you're dealing with. It's pretty obvious I'm a Blackmagic employee and you will find more about me with this name than any other. Many outside of the internet also only know me as "Hook" and have known me for 10-15 years.

You can find these links under my username next to my posts if you really want to know more about me ;) (profiles are public):
https://facebook.com/CaptainH00K/
http://twitter.com/captainh00k/
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostSat Jan 20, 2018 2:45 am


I would say you have understood everything pretty well which is great as you can choose to use these tools or Juans technique from a more informed perspective which was my goal here. Just with #3 again it's not really out of gamut colours I was talking about in regards to spectral response, more to the point no two different sensors see colours the same way and the transforms back to XYZ/scene provided by all manufacturers cannot fully account for those differences. So you will never get two cameras reproducing all colours the same by using those types of transforms. They could be "close enough" for most though. In the end whatever works for you is the decider.

Years of working one way with a particular workflow based on one type of input can be hard to change, and most will fight to stay that way and produce evidence as to why it's "better", instead of just different. At the same time, learning and developing a larger/wider toolbox can be expensive in cost and time, so the choice for many could indeed be to stick to what they know and try to change anything different into what they are used to. It's human nature.

Luca Di Gioacchino wrote:
Sorry to go off topic, but I thought we were supposed to use our real names on the forums (*coughCaptainHookcough*) ;)

Real names are about being accountable and knowing who you're dealing with. It's pretty obvious I'm a Blackmagic employee and you will find more about me with this name than any other. Many outside of the internet also only know me as "Hook" and have known me for 10-15 years.



Does not matter how they called you in the last 15 years, nor that you're a blackmagic representative.

But BECAUSE you're a blackmagic representative and out of respect of everybody else in this forum, you really should follow the rule that your employee setup for everybody: even Peter use the full name, not a random title..
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: UM in Resolve using CST for Alexa color science

PostSat Jan 20, 2018 3:16 am

Juan Salvo wrote:
We're describing the same thing, the "thickness" or "purity" describes how narrow the spectral attenuation of the filter is. We're all saying the same thing.


I think it’s important to be very clear about what’s being discussed.

You’ve twice disputed the point that hooks making about the underlying maths. I think I’m going with his take simply because Ive been in this situation before with him on the very material you use in your demo.

You have a great eye for being a colourist but saying this issue is to do with the “thickness” of the CFA means something to the many that follow you.

While the design of the CFA can greatly affect the end gamut available, it’s a combination of many variables that include the CFA design (which in itself has many variables) that produce that result. It’s a throw away excuse in this scenario that is also seemingly wrong.

You’re reverse engineering the problem when it’s a different kind of issue from what I can understand.

Also, it’s easily confused with another issue that’s been written about greatly, the sensor stack thickness.


John Brawley wrote:Alexa has been ISO 800 since day one, and RED until very recently was ISO 320


Juan Salvo wrote: But same could apply to number of other cameras. The newer RED cameras, which have improve photosite sensity now use thicker CFAs. There's a very strong correlation between those two things.


I think you’ll find it’s a lot more than them changing the thickness of their CFA’s.

Juan Salvo wrote:
John Brawley wrote:I call BS on this thicker CFA theory. Unless someone can point me to actual empirical evidence other than hearsay ? It’s a distraction otherwise to describe something that’s not known.

Ok, we've done tests in house that show this, others have too. But honestly not that worried about winning over the confidence of one person. Feel free to call BS. This is like flat earth theory now.


I’d love to see your empirical tests showing the same sensor with different thickness CFAs. Please prove me wrong.

BMD changed their CFA design on the 4.6K sensor from the original BMCC sensor even though they share the same sensor linage. But the CFA tweak brings many other changes as well. You can’t just point to one variable like this and say that’s why Alexa is so different or why the OFX results vary. Especially when you’re disputing something that seems to not be what you assumed.

waltervolpatto wrote:But BECAUSE you're a blackmagic representative and out of respect of everybody else in this forum, you really should follow the rule that your employee setup for everybody: even Peter use the full name, not a random title..


Walter a little respect for the BMD camera equivalent to Peter or Rohit here, Hook. He’s been here for years under this name. And was known under this name amongst camera folk before he joined BMD. Everyone in the Cinematography forum knows who Hook is. You’ve just maybe not come across him as much and it’s a shame you’ve dismissed his insights based on his name. He really does know what he’s talking about.

Unlike me who’s just trying to keep up.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

waltervolpatto

  • Posts: 10502
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: UM in Resolve using CST for Alexa color science

PostSat Jan 20, 2018 4:24 am


Walter a little respect for the BMD camera equivalent to Peter or Rohit here, Hook. He’s been here for years under this name. And was known under this name amongst camera folk before he joined BMD. Everyone in the Cinematography forum knows who Hook is. You’ve just maybe not come across him as much and it’s a shame you’ve dismissed his insights based on his name. He really does know what he’s talking about.


Regardless of who he/ she is.
There is a rule.
Either the rule is valid for everyone, or we can arbitrarily ask for waiving.
I put my name in all of my posts, it is respectful for him/ her to do the same.
I don't dismiss his/ her knowledge, just pointing out the breaking of a rule.
Tomorrow you can have mister_arri show up and everybody knows him, and who will tell him to change his name from mister_arri to his real name when exceptions are allowed/ tolerated?
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline
User avatar

rick.lang

  • Posts: 17173
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 22, 2018 5:00 am

Peace and love, Walter.

Sometimes there are things that are beyond our understanding but that we abide. What gives mass? How can a photon be a particle and a wave? If energy equals mass times the speed of light squared, what really makes the speed of light what it is?

I just add CaptainH00k to my long list of wonders and don’t doubt there’s a good reason that I’ll likely never understand, but I accept that there’s a good reason why it is what it is.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

Steven Abrams

  • Posts: 275
  • Joined: Tue Feb 24, 2015 12:43 am
  • Location: LA La Land

Re: UM in Resolve using CST for Alexa color science

PostMon Jan 22, 2018 9:20 am

You know how when you go into a hotel, or a restaurant, or basically anywhere and you're not allowed behind the counter or to go into certain rooms/places, but the staff are? Because there's different rules for the guests than there are the hosts/staff? Yeah.

And then there's when someone's pride can't handle the way a discussion has gone because their last 3 posts are nothing but personal and off topic. Gotta love fragile hollywood ego's.

Return to Cinematography

Who is online

Users browsing this forum: carlomacchiavello, paweltkaczyk and 96 guests