Cullen, I shot and am editing my first music video and your channel and knowledge of colour management that you've imparted to me has helped me immensely to achieve a STYLISTIC look for this music video, not JUST a colour-managed correction alone. The band is beyond happy to see their stylistic preferences so easily applied to every shot as I now know how to wrangle all that footage whichever way they want. Thanks to you.
I think my top-2 color management mistakes are: 1. Expecting your tools to behave the same way as they do when grading a rec709 image, and then deciding that color management sucks because that approach doesn't work well on large dynamic range log images. 2. Using tone mapping in the CST that serves as your IDT. This makes it so that Linear Gain (and really every other tool) can't give photometric results.
@@he.smile_ Don't just trust me though! Run the experiment yourself. Shoot two clips at different exposures and then match them with linear gain. If you had tone mapping upstream of the gain adjustment, you'll find it might not be possible! (depending on the tone mapping settings and the contents of the scene)
@@henri-julien Sure thing. IDT is from the ACES terminology, meaning "Input Device Transform". In short, your color management pipeline is typically the following nodes. 1. IDT - this is a color space transform starting with your video file's color space and ending up in your working color space (say, ACEScct/AP1, or DaVinci Intermediate/DWG). 2. Your grade 3. Your last node would be the ODT (Output Device Transform). This is a color space transform from your working color space to your display/delivery color space. As the display's gamut is typically smaller than your working color space and because you have no operations after this step, you would typically want your ODT to have tone mapping and some kind of gamut mapping strategy.
@@thatcherfreeman Ah! Amazing, this clarifies a lot. I am mostly working with SLOG and FLOG footage, so in those cases are you suggesting that I not use any tone mapping in my IDT and then use linear gain to get photometric exposure changes? Would you also suggest not using any tone mapping for something like Iphone footage or drone footage that is not shot in a log gamut? Or in this case is the tone mapping in the IDT preferential? Thank you so much for your help!
Hey Cullen, learned a lot from @AlexTech and @Darren Mostyn before coming across you. Well found you before but could understand a word of what you were saying!. Now I'm doing OK but on colour management I do a different thing and it seems to work out fine. CM (slog3) into DWG timeline and output but then add a CST to rec or cineon/kodak on the clip or timeline. Reason is - the scopes and vector then show all the colour available and it seems to prevent the need for compression dctl's. I have a few other odd ways of grading too - happy to share if you want to discuss. as in no serial nodes just 3 stacks of parrallels. (first for lggo - hSv-curve and balance) yes in a parallel not a serial... think about it each one gets the pure information. (only fail is noise image multiply the bottom end)
Thanks for all the Great videos you make all the time. ❤ 🙏🙏 My question is, I've been trying to recreate a look from a film but finding it difficult. Could you make a tutorial on a specific kind of grade? The Clean Warm look which kind of looks like a Film Print Emulation was used. ??? 🙏🙏🙏
Hello Cullen, Thanks for this new video. Again very interesting. I had a question: have you made a video on how to color grade different faces, different skins together on one single clip? Thank you, Cullen.
So helpful, as always. Thank you sir! I'll give it a shot and keep it n mind, never giving up. BTW, where chan I find the mistake no. 5? I can see only 4 mistakes here. no. 5 : not listening to Cullen
Thanks for the effort you put in those videos! i've been experimenting the last weeks with CM, and recently i tried Aces, i must say i'm really hooks to it, it produces better result than in DWG. And your film lut looks really great. Are your voyager luts Aces compatible?
Cullen I always hear you saying that the Davinci resolve's basic film looks aren't made for color managed however, when you're color managing with CST plugin inside the pipeline wouldn't if I just convert my DWG/ DWI-> Cineon log and then apply the LUT isn't that acceptable? Or is this a form of half semi-color managing? Is there more color science behind this that I'm not aware of? Thanks, as per usual the video was informative!
You absolutely can do that, it's just that when you're using Resolve's film looks you're applying that look in Rec709 color space with a Cineon gamma curve. It's just not as flexible IMO. "Semi-color managing" is just when you get your inputs and outputs wrong when applying a transform lut. With Cullen's free film luts you can drop them in your working color space (DWG or ACES) and modify them to your liking, change their output gain, etc. Basically more flexibility without messing with your color management. It's better to seperate your color management (which should be a purely technical transform) and your overall look. Resolve's film looks don't allow you to do that. But if you like them, you can absolutely use them while also using DWG as a working space. Just put your output to Rec709 Cineon before applying the lut (This doesn't work if you're color managing in project settings). Hope this helps.
Any videos that talk about data levels? Talking about full/video levels, how to know if it’s actually necessary to change. For example, I fly FPV drones and use GoPros. If anybody knows anything about GoPros, they have more information than one might think. Their “flat” color profile with “native” white balance is the rawest form and what I use most of the time, but after I stabilize in reelsteady or gyroflow and render lossless to bring into davinci, if I check “full” levels instead of auto, it changes the image as if it brings the gamma up. But I tested a non-stabilized raw video to compare and “auto” is actually representing what “full” levels look like, so if I change it to “video” levels, it changes the image and crushes blacks. To sum it up, stabilized video rendered as “GoPro Cineform YUV 10 bit” will change when I check full levels, this tells me davinci is detecting it as video levels automatically, but the raw MP4 on auto is detecting it as “full” levels already. I’m just wondering if when I change the stabilized versions to “full”, is it just the same as bringing the gamma way up or is it actually showing me more information?
Hey! This a good question, and one that comes with different answers. I showed some examples of using LUTs that fit nicely into the color management system we use on the channel. The LUTs used in the example are both expecting and returning Davinci Wide Gamut/Davinci Intermediate. In this scenario we would not need to use a CST. However, if you're trying to use a LUT that is setup to receive and return a different color space, you would need to use a CST to apply the LUT correctly. Hope that helps!
