I think that the fixed node tree has the benefits that you mentioned ( i.e. ripple changes etc. ) but most important to me is the time it reduces when using set ups for complicated OFX or other things like power windows that are easy to do, but turning on the node to be up and ready to tweek when doing a lot of shots shaves off time - it's never faster to set up than to just turn on the node and start working. Also the fixed node tree has another add on, that it keeps you organized. Imagine how hard it would be in a kitchen if you had to go to a different random drawer to look for a fork or knife - in our house we have a set drawer, with a molded slots for getting knives, forks and spoons. Sometimes a good memory is from being organized in the first place. You can orphan nodes as well - so when swapping the CST for ACEs - you can then click the orphaned node and swap the nodes - which keeps the numbering for ripple changes. Love your videos - great work and great information. youTube is a bit of crap shoot in that regard - I appreciate the ones that are on point. Thank you
I'm curious about "ripple changes" -- is that different than "groups"? And how do trims work, I've tried to do that where I push in on a clip within a group, so that all the clips in that group have the same framing, but it only seems to work on 1 clip. Maybe I'm missing something. Great video -- hope there's more!
The sizing attributes would only apply to a whole group if you used the node sizing on a node in that group. The general sizing in the color page is clip by clip.
Super helpful video! Struggling to match the contrast of my resolve viewer (on Mac) to that of the exported clips. The shadows appear slightly lifted in exports. Experienced this issue when grading solely in 709/2.4 color space but integrating wg/intermediate into my process seems to have amplified the issue slightly. Any advice? I still have "Use Mac display..." ticked in my preferences.
This is a widely known issue with Mac’s display, I suspect solution could be a small gamma change via a LUT to emulate final render. I haven’t had the time to get into it yet, but have you tried Cullen’s fix? - Jai
Suggestion : wouldn’t it be better to put your scene node right after the secondaries? This way it would be easier to qualify and make HSL adjustments if let’s say you have a warm push for the scene! Great video 👍🏾
This is a good point, mostly the scene correction isnt big so can be before doing keys, but if its quite extreme such as a day for night look then doing it in the other order is better. Actually the real rule of thumb is how was it shot especially regarding white balance, and how does it need to end up? Heavily tinited at any stage or not? You might need to clean it up at some point in the node graph to pull keys and then put the tint back afterwards, or you could have a seperate "well balanced" branch that is just used for pulling keys. My node structure is pretty fluid job to job depending on requirements.
Hey! Jai this side, guy behind the channel. Paul is hosting couple of episodes here at Colorist Foundry. Paul is a senior colorist. He's originally from South Africa and currently based out of Taiwan. I think this video might do a better job at his introduction than me :) ruclips.net/video/Bp2x91rpQnw/видео.html
Regarding workflow(s): shotmatching arguably is very important since it remained even in you simplest node tree. If one lacks a control surface, printer lights on the num-block seem a perfect and fast alternative. Adjusting offset with clicks of printerlights requires a pure log-space if I am not mistaken to emulate a true exposure adjustment. Many log-spaces have toes, so one expects shadows to not behave like an exposure adjust. Which 'pure' log spaces are there to perform shot-matching fast then? Is it "just" ACEScc? Does it make sense to CST into ACEScc for one node and then go to something that is more amenable to other adjustments? Would love to hear from you on these topics (printer lights, pure-log vs. log-with-toe, CSTs for just one node). Cheers!
As far as I know, the closest thing today we have to changing exposure photometrically is gain applied in linear. If you like using printer lights shortcuts, which currently only ever change offset, And you have Resolve Studio, you can use Thatcher Freeman's Rebind LGGO DCTL to make printer lights actually affect gain in linear :)
@@rockindude3001 Yep, you're totally correct - Offset in ACEScc is the same as gain in linear, except for minor differences in handling of negative values.
Exactlt what Kaur said. Funny thing is how those are fairly recent developments only possible once color space transforms came in. For many years prior to aces people used offset on camera log and then in the telecine days we used lift gamma gain on the “log-ish” signal that the scanned negative gave us. The only thing i like about the hdr tool is that it has a global offset in stops. And as long as your timeline space is set correctly it behaves the same as gain in linear ( it probably is just that under the hood).
these videos are really helpful, appreciate the time and effort that has been put into these. grateful they're free as well!!
Glad you like them!
I think that the fixed node tree has the benefits that you mentioned ( i.e. ripple changes etc. ) but most important to me is the time it reduces when using set ups for complicated OFX or other things like power windows that are easy to do, but turning on the node to be up and ready to tweek when doing a lot of shots shaves off time - it's never faster to set up than to just turn on the node and start working. Also the fixed node tree has another add on, that it keeps you organized. Imagine how hard it would be in a kitchen if you had to go to a different random drawer to look for a fork or knife - in our house we have a set drawer, with a molded slots for getting knives, forks and spoons. Sometimes a good memory is from being organized in the first place.
You can orphan nodes as well - so when swapping the CST for ACEs - you can then click the orphaned node and swap the nodes - which keeps the numbering for ripple changes.
Love your videos - great work and great information. youTube is a bit of crap shoot in that regard - I appreciate the ones that are on point. Thank you
In my kitchen, its not so much about saving time but rather not burning stuff. 🤣
Thank you for your work ❤
I'm curious about "ripple changes" -- is that different than "groups"? And how do trims work, I've tried to do that where I push in on a clip within a group, so that all the clips in that group have the same framing, but it only seems to work on 1 clip. Maybe I'm missing something. Great video -- hope there's more!
The sizing attributes would only apply to a whole group if you used the node sizing on a node in that group. The general sizing in the color page is clip by clip.
Super helpful video! Struggling to match the contrast of my resolve viewer (on Mac) to that of the exported clips. The shadows appear slightly lifted in exports. Experienced this issue when grading solely in 709/2.4 color space but integrating wg/intermediate into my process seems to have amplified the issue slightly. Any advice? I still have "Use Mac display..." ticked in my preferences.
This is a widely known issue with Mac’s display,
I suspect solution could be a small gamma change via a LUT to emulate final render. I haven’t had the time to get into it yet, but have you tried Cullen’s fix? - Jai
This is educative
Suggestion : wouldn’t it be better to put your scene node right after the secondaries? This way it would be easier to qualify and make HSL adjustments if let’s say you have a warm push for the scene!
Great video 👍🏾
This is a good point, mostly the scene correction isnt big so can be before doing keys, but if its quite extreme such as a day for night look then doing it in the other order is better. Actually the real rule of thumb is how was it shot especially regarding white balance, and how does it need to end up? Heavily tinited at any stage or not? You might need to clean it up at some point in the node graph to pull keys and then put the tint back afterwards, or you could have a seperate "well balanced" branch that is just used for pulling keys. My node structure is pretty fluid job to job depending on requirements.
Just found you three hours after you uploaded this video. Best first video I have ever watched. Thanks! Subbed! - Where are you based? NZ?
Hey! Jai this side, guy behind the channel. Paul is hosting couple of episodes here at Colorist Foundry. Paul is a senior colorist. He's originally from South Africa and currently based out of Taiwan. I think this video might do a better job at his introduction than me :) ruclips.net/video/Bp2x91rpQnw/видео.html
this looks great , where i can see the other parts (1, 2)
ruclips.net/video/7j-fJxrLRUk/видео.html & ruclips.net/video/Bp2x91rpQnw/видео.html&
Regarding workflow(s): shotmatching arguably is very important since it remained even in you simplest node tree. If one lacks a control surface, printer lights on the num-block seem a perfect and fast alternative. Adjusting offset with clicks of printerlights requires a pure log-space if I am not mistaken to emulate a true exposure adjustment. Many log-spaces have toes, so one expects shadows to not behave like an exposure adjust. Which 'pure' log spaces are there to perform shot-matching fast then? Is it "just" ACEScc? Does it make sense to CST into ACEScc for one node and then go to something that is more amenable to other adjustments? Would love to hear from you on these topics (printer lights, pure-log vs. log-with-toe, CSTs for just one node). Cheers!
As far as I know, the closest thing today we have to changing exposure photometrically is gain applied in linear. If you like using printer lights shortcuts, which currently only ever change offset, And you have Resolve Studio, you can use Thatcher Freeman's Rebind LGGO DCTL to make printer lights actually affect gain in linear :)
@@KaurH but that is what I mean: mathematically offset-in-pure-log is gain-in-linear. Then the convert to ACEScc trick should work without a DCTL.
@@rockindude3001 Yep, you're totally correct - Offset in ACEScc is the same as gain in linear, except for minor differences in handling of negative values.
Exactlt what Kaur said. Funny thing is how those are fairly recent developments only possible once color space transforms came in. For many years prior to aces people used offset on camera log and then in the telecine days we used lift gamma gain on the “log-ish” signal that the scanned negative gave us. The only thing i like about the hdr tool is that it has a global offset in stops. And as long as your timeline space is set correctly it behaves the same as gain in linear ( it probably is just that under the hood).
Do you have that power grade available somewhere to download?
✨👋💪😮👍✨
Is that not a South African accent I hear?
It is!