Oh yes sure. We have worked with Company 3 for several years, either delivering “linear_srgb” exr files or “linear_ap0” (ACES2065-1) exr files. They never asked for anything though, we did take the decision on our own.
But I think my point is : working space and delivery space do not have necessarily to be the same. Just look at an ACES workflow for instance. You most likely render in ACEScg but interchange on ACES2065-1.
Chris
Ok. So what if the device isn’t BT.2020, which as we all know, is a shred of a niche so much as to be a statistical error.
What happens to those values on a MacBook Pro Display P3 device? What about a common sRGB display? Or a generic DCI-P3 theatre?
These are the very real problem of using wide gamuts, and I’m curious what you think is happening?
The color grading is the last step in production. You have a master, let’s say ACES. The colourist will output many different versions, depending on how it is going to be shown. The REC2020 will be for UDH broadcast and 4k blurays. The rec709 for HD and DVDs. The DCI-P3 is for digital cinema projector. If the colorist need to output for UHD, he will not make the color grading on his computer monitor but on a UHD tv. In the time of NTSC, we use to call then sh?$ boxes, because professional monitors would be give you the same results as a regular TV. And NTSC is know as Never Twice the Same Color. At Real by Fake, when we mastered The Morning Show for Apple, is was done on a huge UHD. That’s the only way you will have the right results. If you need to do the colouring for cinema, you do it in a projection room. I’ve seen it at Technicolor.
You would not mix a 5.1 Dolby Surround soundtrack on iPhone headsets. You need an approved Dolby system.
So it’s unlikely you would see a Rec2020 on your computer monitor.
I know this is off-topic, I apologize, but I’m curious : how can you know for sure whether he’s syndicated ?
What I am trying to address is far more practical, which is what people actually look at, during authorship and working. Focus on my question, please.
For feature films I would say that color grading is actually the first step in a production. The look development of the “show lut” would normally happen at the beginning of a show before any production starts, and would involve the director of photography and the colorist.
This show lut would then be used as “the window that we look through to see our image data,” at every step of production, from on-set review, through vfx, to DI and finishing.
Working on images without that window, or with a different window, is a recipe for disaster. The window influences all creative decisions made, on set and in vfx and is an essential component of the imaging pipeline (especially working scene-referred).
Yes but you know that at the end of the process it goes back to color grading. You can’t prepare all the different versions rec2020, rec709 etc at the beginning of production.
I will not focus on your question because I feel like you just want to challenge every single sentence I write and it’s getting tiring. I hope you don’t do this with the directors and the DOPs on set. I don’t know what your end game is. I don’t know if you are trying to prove yourself something. Frankly, I don’t care. I’m just done with that tread.
Unionized is how we refer to it here – but key grips are IATSE, and if he’s worked on shows like Deadpool and Jurassic World I am pretty sure those were union shows. Since he’s in Canada I assume he’s working on the IATSE Canada jurisdiction - https://canada.iatse.net
Either way…much better off than most VFX artists who unfortunately don’t have a union (yet).
The way I see it is that this is a highly complex topic, not necessarily one that many creatives are primed to understand since it deals with abstractions as well as some complex math and science that will cure insomnia for many artists.
We are all trying to understand the material, and are getting information where we can find it (i.e. online, and for some of us at work from people whom we view as experts). However it is important to understand that this knowledge is highly fluid, and while it’s necessary that we all strive to increase our knowledge, it is also important to understand that we might attempt to communicate the information that we understand as being accurate at any given point – but not try to misguide others for our own amusement.
…
What I do think is of critical importance however is to continue these types of conversations in public forums such as these, in a respectful and friendly manner – we are all after the same goal here, we are all on the same side!
I have learned so much just in this past week, and I would please urge everyone to dial the tone back a hair so that we may continue a thread that will likely become something that is referenced for years to come.
Fair. Deleted.
Seconded. Best I can tell, BlenderBob is coming from a sincere, collaborative place, and bullying him out of the thread does no good for the rest of us.
Discord and 4chan have the screaming emo know-it-all thing locked down. Heck, I’ve noticed lately that even Reddit has become more respectful that what I’m seeing in this thread.
Not to overstep my place, but for the sake of the community, can we maybe confine the edgelord malarkey to the skate park and keep this place respectful and supportive?
Small World… and sorry for the off topic 
, but tell JB (John Brawley) if you’re on the sets when Season 3 starts up, that Tim said “Hi”.
I got to hangout with him on set a few times when he was the DP on “Queen of the South” Season 2 here in Dallas. Super cool dude if you haven’t got a chance to meet.
damnn i need to learn compositing and color grading/correction stuff, its really complex!!!
Come over to the Blackmagic forums and you can tell him yourself. He’s always hanging out there and participating on the various conversations! 
That’s actually how I first became familiar with JB, but through the older now defunct BMD forum. A solid dude working at the highest level and still active on the forums helping junior shooters learn their craft.
Sorry for the derail. 
I haven’t worked in film production at all but I’ve read that VFX peeps get frustrated by this process. Where the production LUT (look) is determined earlier but grading and direction crush the details down into the blacks to get a Hi-con style that “builds” upon the earlier LUT.
And I guess this is what Troy is getting at. How, in VFX dev, do you asses the image when its being worked on in linear? It must be getting a transfer curve applied so that you can see it represented on the sRGB monitor. In this case Blender is applying Filmic transfer curve so that the sRGB display can view it?
Should this Linear to Display transfer function get applied to baked imagery like jpeg or png? Well if the next app that displays that image can’t interpret the source image correctly (pre-filmic) then it becomes WYSIWYG, no metadata required.
But then I see @BlenderBob’s point, if you work with this transfer function that behaves like a LUT to the user, how do you know that your careful decisions will be retained in the next app? I guess they would all have to execute Filmic the same way that Blender does?
There are two separate but intertwined factors at play with the way an image is properly displayed – Gamma curve and Color Space. One of the main reasons why many people will bring their EXR’s into another app like AE or Resolve is that it will look incorrect due to EXR having a Linear Gamma while PNG’s and JPEGS have an sRGB (Gamma 2.4) or rec709 (Gamma 2.2) baked into the image. Actually technically speaking AE will apply a built-in transform upon importing an EXR automatically, but if then eXtractor (equivalent to Nuke’s Shuffle) is applied to the image to select which render layer to use, then the EXR will be brought back into Linear Gamma.
The Color Space refers to the size of the triangle within the CIE horseshoe color shape that almost everyone is familiar with by now. The main problem the way I understand it is that, as that triangle gets bigger and bigger from say sRGB to rec2020 or ACES, the way the color values are translated becomes unreliable and yield undesirable results such as clipping of similar colors into one blob.
Morning guys,
I hope you are doing alright !
It is not exactly like that but almost. From this article :
The sRGB Electro-Optical Transfer Function (EOTF) is a slight tweaking of the simple Gamma 2.2 function.
I think you have mixed up sRGB and Rec.1886. Please note the “slight tweaking” part : an sRGB EOTF is not exactly a Gamma 2.2. The Rec.709 OETF is actually a camera encoding, almost equivalent to a Gamma at 1.95.
I can only recommend this excellent video on this topic.
Not the color space, but rather the primaries (which are the vertices of the triangle) or the gamut (which is the triangle itself). The CIE horse shoe shape is a chromaticity diagram.
Exactly. I have not seen yet any color management workflow capable of displaying wide gamut values properly.
Sorry but I disagree. Is Troy a pain in the arse ? Most certainly ! Are his questions necessary ? 100% yes ! There is no endgame with Troy and he is not trying to prove himself anything. He wants to empower image makers and extend our/his knowledge.
I won’t deny that his questions have caused me several headaches. But in the long-term ? Absolutely worth it ! I could not have helped this thread for instance without him hammering stuff with much much patience over the last year.
Chris
Thank you for the corrections, I’m still half asleep here! 
The idea that there is still no color management workflow capable of displaying wide gamut values properly is troubling given how exponentially important color management is becoming. Despite all the hate they get for their “walled garden” Apple so far seems to be one of the few companies to take this issue seriously.
Regarding Troy, I would point people to his web site which is a very informative repository of information “colored” with his personality which I find quite enjoyable: https://hg2dc.com/why/