Monitor ICC with DaVinci Resolve

[MichaelEzra]

Since Photoshop works just fine with OS color management, I see no reason why it wouldn't work for video app. Clearly this must be BM marketing strategy requiring additional however inexpensive useless graphics card occupying PCIe slot. I would have been much happier customer paying more for Resolve but using that PCIe slot for SLI GPUs with more power to speed up rendering. Oh well..

[fredjeang2]

I think it's not so different than really understanding color management with Photoshop.  It's just that if a print isn't quite right, you just make another.  With movies there's often a lot more money on the line and not so easy to keep re-rendering the movie!  And once, the movie is distributed, it better be correct.

But still, it's not really so complicated once you get your head around it:)  Just look at the color management forum here...

Compared to color management in Photoshop?!?
Come on Bruce!
I have dozens of forum topics I could copy/paste here and nobody
Would understand anything on what's going on.
Aces for ex! Nobody fully understands how to make it work
Appart you and 5 or 6 other guys in the high-end LA.
Even in LGG, the one who know are very few.
Photo goes to print and web.
There is only 3 softwares worldwide used profesionaly
In still imagery. All runned by Adobe. PS, IS, ID. PS is both
The Resolve and the Nuke of photography.
Learning both Resolve and Nuke as well as most people know PS
Is an entire 4 or 6 years full time university course.
Not talking about systems like Mistika...another stratosphere.
Printing procedures are stable and universal.
What about DCPs? (And flavors, and if theaters do their job, and if the projector...)
What about the myriad of
Proprietary flat curves with their own special
Workflow? RLF, Slog, LogC, Pana...the list is infinite.
Now wide gammut Red with new curve, Rec709? Naaa...Rec2020...
ACEScc, ACESlinear, ACESproxy, ACEScct...what is going to be the new one?
What about the overloaded pac of codecs? What about roundtripping?
Ahhh...that's a good one, roundtripping between aps, when it works.
What about correct maths to do LUTs? 99% of the luts that circulate
Are wrong and crap. Who really know the maths to do that? The real pros.
It's not doing Technicolor simply using split channel as I saw in a Resolve tuto one day...
Completly wrong! But the dude who did that honestly thought he had something.
Just that he didn't have the real knowledge.
99% of the videos on Resolve on color that circulate are
Done by amateurs. When one goes in LGG, things change
Completly and it's way more complicated.
The list of technical habilities in motion post is enormous,
Way beyond the scope of a full mastering of PS.
And let's not even talk about the prod itself...
A good cam operator for ex is quite chalenging.
You have been in this business I guess all your working life,
Before digital and evolved naturally with years. I think this
Is why it's naturally easy for you. You gainned the expertise.
But a guy who jumps
In it knowdays has to learn it all in once. And that's IMO
Only possible in a specialisation.
A DP has to know things a director necesarlly not.
IMO, the key is to work with the right people for the right job.
Want a coffee?

[fredjeang2]

Since Photoshop works just fine with OS color management, I see no reason why it wouldn't work for video app. Clearly this must be BM marketing strategy requiring additional however inexpensive useless graphics card occupying PCIe slot. I would have been much happier customer paying more for Resolve but using that PCIe slot for SLI GPUs with more power to speed up rendering. Oh well..

I copy-paste a text from Steve Shaw

Unfortunately ICC profile compliance is rather a hit-and-miss affair.
It's one reason there are rarely used in the film and TV industry - it's very difficult to rely on them.
Calibration should really be display based, not associated with the image path, or the graphics program.
That is the only way to guarantee that any displayed images are correctly represented.

With Mac, ICC profile management is 'supposed' to be OS wide, but the reality seems rather more variable.
With Windows ICC management is program specific.

A real nightmare when it comes to colour accuracy.

For accurate colour management it is best to avoid any reliance on ICC profiles.

Steve

[MichaelEzra]

Strange, my NEC monitor takes internal calibration (display-based LUT) in addition to the corresponding ICC profile in OS / graphics card.
In ICC profiling color is measured by the calibration sensor from the display directly, why would it not be representative of the display?

I cannot complain on ICC accuracy using Photoshop. The program just needs to be ICC-aware and compliant, and as we know, most of them are not.

I suppose offloading color profile conversion workload from CPU/GPU to a dedicated hardware device, such as broadcast monitor, would make the internal computer resources more available for the demanding video processing workloads. May be this is the true underlying reason?

[digitaldog]

I copy-paste a text from Steve Shaw

Unfortunately ICC profile compliance is rather a hit-and-miss affair.

FUD (and without specifics). Who is Steve Shaw?

Calibration should really be display based, not associated with the image path, or the graphics program.

All actually

With Mac, ICC profile management is 'supposed' to be OS wide, but the reality seems rather more variable.

No specifics and yes, it's OS wide where it's needed (if not, where isn't it)?

For accurate colour management it is best to avoid any reliance on ICC profiles.

More FUD. If not relying on the ICC architecture, which has been used for over two decades, what then? 

[fredjeang2]

Who is Steve Shaw?

He is the CEO of Light Illusion. A UK company.

https://www.lightillusion.com

When this guy speaks, I listen. (Doesn't nean that he can't
Be wrong but it would be very strange in this particular topic).

[digitaldog]

He is both wrong and biased without providing a lick of data to back up his FUD on ICC CMS!

[fredjeang2]

He is both wrong and biased without providing a lick of data to back up his FUD on ICC CMS!

Multi million production houses with the best
Operators relly on this guy knowledge and company.
An example here (although dated)
http://www.btlnews.com/crafts/post-production/the-look-chooses-lightspace-cms-for-facility-wide-colour-management/
For a reason it will be.
He can not be completly wrong.
Nobody at those levels looses time and money.

[digitaldog]

Multi million production houses with the best
Operators relly on this guy knowledge and company.

Is that the best you've got, are you serious? Here's just one company that implements ICC color management in every product they print for their customers (I helped create some of those output profiles) and this company has provided ICC support in their software since the early 90's and this company is part of the ICC: Apple!
Last year their profits were 56 billion dollars. Hollywood was 11.2 billion (down). So your argument about money to back up his FUD as pasted, again without a lick of proof or explanation is rubbish!
Let's see you or him provide some colorimetric data to back up this sentence which is, as provided FUD: For accurate colour management it is best to avoid any reliance on ICC profiles.

He can not be completly wrong.

“There are two ways to be fooled. One is to believe what isn't true; the other is to refuse to believe what is true.”
― Søren Kierkegaard

[smthopr]

Compared to color management in Photoshop?!?
Come on Bruce!
I have dozens of forum topics I could copy/paste here and nobody
Would understand anything on what's going on.
Aces for ex! Nobody fully understands how to make it work
Appart you and 5 or 6 other guys in the high-end LA.
Even in LGG, the one who know are very few.
Photo goes to print and web.
There is only 3 softwares worldwide used profesionaly
In still imagery. All runned by Adobe. PS, IS, ID. PS is both
The Resolve and the Nuke of photography.
Learning both Resolve and Nuke as well as most people know PS
Is an entire 4 or 6 years full time university course.
Not talking about systems like Mistika...another stratosphere.
Printing procedures are stable and universal.
What about DCPs? (And flavors, and if theaters do their job, and if the projector...)
What about the myriad of
Proprietary flat curves with their own special
Workflow? RLF, Slog, LogC, Pana...the list is infinite.
Now wide gammut Red with new curve, Rec709? Naaa...Rec2020...
ACEScc, ACESlinear, ACESproxy, ACEScct...what is going to be the new one?
What about the overloaded pac of codecs? What about roundtripping?
Ahhh...that's a good one, roundtripping between aps, when it works.
What about correct maths to do LUTs? 99% of the luts that circulate
Are wrong and crap. Who really know the maths to do that? The real pros.
It's not doing Technicolor simply using split channel as I saw in a Resolve tuto one day...
Completly wrong! But the dude who did that honestly thought he had something.
Just that he didn't have the real knowledge.
99% of the videos on Resolve on color that circulate are
Done by amateurs. When one goes in LGG, things change
Completly and it's way more complicated.
The list of technical habilities in motion post is enormous,
Way beyond the scope of a full mastering of PS.
And let's not even talk about the prod itself...
A good cam operator for ex is quite chalenging.
You have been in this business I guess all your working life,
Before digital and evolved naturally with years. I think this
Is why it's naturally easy for you. You gainned the expertise.
But a guy who jumps
In it knowdays has to learn it all in once. And that's IMO
Only possible in a specialisation.
A DP has to know things a director necesarlly not.
IMO, the key is to work with the right people for the right job.
Want a coffee?

A pretty darn good description of the state of color management in the motion business!

But you forgot to mention HDR display which is a whole new can of worms (as we say in English:)

And even the "standard" REC709 doesn't have a true gamma standard and that creates a lot of confusion for people.

And "video/broadcast" levels vs. "data" levels!

I've seen the biggest and most expensive post houses make mistakes sometimes ... and they always blame the client!!!!

[smthopr]

Is that the best you've got, are you serious? Here's just one company that implements ICC color management in every product they print for their customers (I helped create some of those output profiles) and this company has provided ICC support in their software since the early 90's and this company is part of the ICC: Apple!
Last year their profits were 56 billion dollars. Hollywood was 11.2 billion (down). So your argument about money to back up his FUD as pasted, again without a lick of proof or explanation is rubbish!
Let's see you or him provide some colorimetric data to back up this sentence which is, as provided FUD: For accurate colour management it is best to avoid any reliance on ICC profiles.
“There are two ways to be fooled. One is to believe what isn't true; the other is to refuse to believe what is true.”
― Søren Kierkegaard

Andrew, Steve Shaw is one of the most expert people in the display calibration business for motion pictures.  He has written many posts with the data to back them up, but you'll need to search for them.  I think he sometimes argues for a level of color accuracy that is beyond necessary to promote his products, but he is extremely knowledgable.

I think most of this has to do with motion delivery always being delivered to a standard, of which there used to be only a few standards so it made good sense to set the display to the standard and not deal with color management at the OS level or even the application level.  Now, they are creating new HDR display standards every other month and it's becoming really complicated.  The dynamic range of these new standards is all over the map and I think they are so different that automated mathematical transforms don't work effectively.  Just like in Photoshop, when one is printing the same image to low contrast matte paper vs. glossy vs. a high contrast emissive display, one needs to do a new color correction (using soft proof) for each final product.

And just to be a little provocative here...  Even when doing still photography in photoshop, for print output, I think the REAL working space is the display, not the underlying photoshop working space.

There is a new movement in motion to use ACES  (  http://www.oscars.org/science-technology/sci-tech-projects/aces  ) color management, which is the motion picture academy's attempt to create a photoshop like .icc color management system for motion pictures.  It works well... until it doesn't.  I think this is because in ACES the color space transform and the output contrast curves are built into the same transform.  This means the contrast curve is baked in and difficult sometimes to get out of.  Otherwise, it works ok, but assumes a standard display of some sort.

And so, for motion, the best practice is to create a 3D LUT to bring the display into standard compliance.  We often work with multiple displays (for client viewing) so it's best that each display hit the same standard  from a single signal from the video card.  Also applying the 3D LUT in software is possible, but it slows down the processing which can become a hurdle when we need real time playback of color correction of 4k+ material.  It's not much, but it's sometimes the difference between realtime playback and not quite realtime.  I'm a small time guy compared to tent pole movies, and I have my 3D display LUT transform done in software and I notice the effect on playback.  I will soon need a LUTbox to place between my graphics card and the display.

Sorry to mumble on here.  I know it's a forum for still photographers, but since some are now getting into motion, and motion post production, I thought I'd give a little idea of the complexities and differences with still photography color management.

[fredjeang2]

.  It works well... until it doesn't...

I've experienced that!

[digitaldog]

Bruce, everything you've stated about this fellow may be accurate but the statement as pasted is rubbish. Maybe taken out of context. Maybe there is an anti ICC agenda. There's so far, not a lick of colorimetric data to backup his claim which was made extremely broad brush. And thus, until proven, it is IMHO just a load of FUD.
Are the same ICC tools and workflow used in still appropriate for video/film? Don't know. But again, if not, that wasn't a fair or correct or colorimtrically correct statement and hence my comment. What is best practices for one industry may not be best for another; that doesn't make the previous workflow a hit-and-miss affair. The comment: "With Mac, ICC profile management is 'supposed' to be OS wide, but the reality seems rather more variable" is equally bogus even when using seems in the sentence. Where in the OS, when working with images, is exactly it variable? Variable in what way?

[digitaldog]

There is a new movement in motion to use ACES  (  http://www.oscars.org/science-technology/sci-tech-projects/aces  ) color management, which is the motion picture academy's attempt to create a photoshop like .icc color management system for motion pictures.

Interesting link but I see nothing about an ICC color management architecture or anything linked to Photoshop per se.

Even when doing still photography in photoshop, for print output, I think the REAL working space is the display, not the underlying photoshop working space.

You can think that, but the entire idea behind RGB working space is to divorce the display from how we edit our images! The display is often an intermediary device. It's gamut is severely limited compared to many, many output devices. The idea behind synthetic RGB editing spaces is to keep users from editing based on a single emissive output device:
http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf

[fredjeang2]

For anyone interested to know a little more about Steve, there is this interesting interview at NAB that also points towards the concerns on HDR displays Bruce just emitted.
https://www.youtube.com/watch?v=6OWs75BLCAA

[smthopr]

Interesting link but I see nothing about an ICC color management architecture or anything linked to Photoshop per se. You can think that, but the entire idea behind RGB working space is to divorce the display from how we edit our images! The display is often an intermediary device. It's gamut is severely limited compared to many, many output devices. The idea behind synthetic RGB editing spaces is to keep users from editing based on a single emissive output device:
http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf

I completely understand the idea of synthetic RGB working spaces, but in reality, I'm editing what I can see, not what I can't.  Hence my opinion.

For motion picture work, I'm color correcting on a display set to REC709 (very close to sRGB).  The cinema standard is P3 which is a bit bigger (like Adobe RGB).  I work in the smaller space as the transform to P3 is accurate with no out of gamut colors.  Yes, I'm limiting the colorspace, but very little of our work doesn't really fit into REC709.  If I color corrected to P3, there could be out or gamut colors in the conversion to REC709 for TV delivery.  Rather than deal with a separate color grading pass to fix those issues, I work in the smallest standard.  Things get more complicated in print as there are colors that the display can produce, that can't be printed, and colors that can be printed, that the display can't produce.  But, when we soft proof, we're still making corrections based on the limits of the display as that's all we can see   Perhaps some colors that we can't see on the display will get printed, but we won't find out till we make the print...

[digitaldog]

I completely understand the idea of synthetic RGB working spaces, but in reality, I'm editing what I can see, not what I can't.  Hence my opinion.

Your display cannot show you (you cannot see) what's in all your data, data you can and many do use for output to something other than a limited gamut emissive display. No printer can produce all of sRGB or any RGB working space or display space. There simply is a disconnect between RGB working spaces and output color spaces. This limitation may be of little concern if your only output is to another emissive display system (of matching color gamut). That simply isn't the case in the still/photo world. We (many of) make prints. Or have our image reproduced on a press. The display is just an intermediate output in the chain.

I work in the smaller space as the transform to P3 is accurate with no out of gamut colors.

Depending on the color gamut of the source. Those of us capturing raw data; not the case.

Yes, I'm limiting the colorspace, but very little of our work doesn't really fit into REC709

For your industry (film), today sure. For photo work, not at all. Not at all difficult to find many real world images** from raw that greatly exceed REC709, DCI-P3, Adobe RGB (1998) etc. For those of us that wish to capture and retain all the color we can, and output all the colors to any number of devices, none of those working space are sufficiently large enough to contain the color data. Another reason why raw converters like those from Adobe use much larger color spaces for processing.
I realize that the needs of the photo industry and the needs of the film industry, the limitations of color gamut differ. NONE of the above changes nor defends Steve's text on ICC color management as posted.


** Everything you thought you wanted to know about color gamut

A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.

High resolution: http://digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q

[fredjeang2]

as posted.

I didn't add nor rest a line. It's the complete post as it. But be aware that when people writte in internet they not always take the time to devellop. it is obviously a resumed vision of his view, without entering into details. But again, with the person's pedigree, it's a higly trustable source in his field.

[digitaldog]

For anyone interested to know a little more about Steve, there is this interesting interview at NAB that also points towards the concerns on HDR displays Bruce just emitted.
https://www.youtube.com/watch?v=6OWs75BLCAA

Well that's 12 minutes of my life I'll never get back   
Nothing in the video from Steve that has anything to do with the subject of color management, nothing to back up the pasted claims about the so called 'issues' of ICC color management.

[digitaldog]

I didn't add nor rest a line. It's the complete post as it.

And you expect other's to believe it? As you did*?
Be aware that when people write on the Internet, it's often FUD and bogus.

But again, with the person's pedigree, it's a higly trustable source in his field.

I see no reason to trust what he wrote because it's FUD, without a lick of data to back it up, from someone presumably with an agenda.
“*Anyone who has the power to make you believe absurdities has the power to make you commit injustices.” Voltaire,