The Offset Podcast - The Offset Podcast EP 046: Listener Submission - Greenscreens
Episode Date: December 17, 2025Happy Holidays!It’s our last episode of Season 2 and for the year! We’ll return in mid-January 2026 for Season 3 with some great new topics, new types of content and much more.If you�...��ve enjoyed The Offset Podcast help us improve it. If you have 5-10 minutes we’d love it if you could take our audience survey by visiting this link: https://wkf.ms/4acQMvbThank you to our amazing audience we couldn’t do the show without your support. Wishing you and yours a great holiday season!————Continuing our recent efforts to address some viewer/listener submitted topics, in this episode of The Offset Podcast we’re exploring a question we got from audience member Jared about how to integrate greenscreen work into a color/finishing pipeline.It’s a great topic, and one as you’ll see and hear Joey gets particularly amped up about!Since we’re a podcast and not a tutorial channel, this episode is more about big picture strategies and techniques then the nitty gritty of particular sliders and knobs, but we still cover a ton. Some of the specific topics we explore include:Success in greenscreen work starts on setKeying on the Resolve Color page vs Fusion pageFusion’s Delta KeyerGetting more comfortable with mattes and transparencyEdge refinement, matte finessing and the difference between the twoCombining keys and using garbage mattesControlling spill with spill suppressionLight wrap, shadow creation and getting better compositesHanding off to a VFX professionalIf you like The Offset Podcast, we’d love it if you could do us a big favor. It’d help a lot if you could like and rate the show on Apple Podcasts, Spotify, YouTube, or wherever you listen/watch the show. Also if you liked this show consider support the podcast by 'buying us a cup of coffee' - https://buymeacoffee.com/theoffsetpodcast
Transcript
Discussion (0)
Hey, everybody. Welcome back to another episode of the Offset Podcast.
And today we're addressing a viewer or a listener submitted a question about green screen work.
Stay tuned.
This podcast is sponsored by Flanders Scientific, leaders in color accurate display solutions for professional video.
Whether you're a colorist, an editor, a DIT, or a broadcast engineer, Flanders Scientific has a professional display solution to meet your needs.
Learn more at flanderscientific.com.
Hey everybody, welcome to another episode of The Offset Podcast.
I'm one of your host, Robbie Carmen.
And with me, as always, is Joey Deanna.
How are you doing, Joey?
Good.
I was going to say Joey Deanna master ace mechanic because you're wearing your beul shirt.
And I know you've been hard at work at some motorcycle surgery over the past couple days.
But I thought it was a perfect opportunity to mention your mechanic work because today's topic, today's episode,
reminds me of being a mechanic a little bit, of green screen work.
There's a lot of moving pieces, small parts, technical stuff that can kind of, you know,
give you success or give you challenges depending on how you do it.
So, yeah, we wanted to kind of address a question that we got from one of our viewers,
our listeners, Jared, about integrating green screen work into their color pipeline,
their finishing workflow.
You know, and this is a really interesting topic.
I think the one thing that we want to just stress
at the very top here, this is not a tutorial.
This is a podcast, right?
So this is not something we're going to be like,
okay, now move this slider, now move that slider.
Now, you know, whatever.
We're going to more address this kind of
from a conceptual workflow, big picture level.
There's obviously, right, Joey,
a plethora of tutorials out there on the old interwebs
that you can get to learn how to do this better.
on the color page, infusion, in other places.
But this is one of those things where I think once you kind of get some of the big picture
conceptual parts of green screen work, then it really doesn't matter where you're doing the
work, right?
Whether you're in fusion, nuke, after effects, resolve, whatever, you can kind of go into
it armed with some of the baseline knowledge, right?
So I just wanted to be clear about that off the top, that this is less about step by step
in this particular episode and more about conceptual stuff.
Now, Joey, before we begin, I just want to say we've made it to the end of season two, man.
This is, it's the last episode of the year for 2025.
Our seasons run from January 15th to December 15th every two weeks.
So we got another, I don't know, whatever, how the mouth works out, 24, 25 episodes out there.
This is our second year of the show.
And I think you'll agree, big thanks to all of our viewers and listeners.
They've helped us grow the show this year.
We're excited about 2026 and moving forward with new episode ideas, new features.
We're thinking about how to grow the podcast, get it into more people's hands.
And you guys have been a large part of making that happen.
The podcast has grown significantly in 2025, and we're hoping to do more of that in 2026.
Now, to that end, Joey and I realized, we don't know a lot about you guys, our audience, about what you do, you know, who you
are, that kind of stuff. So one of the things that we would love, if you could, you know, spare
five or ten minutes to do is that we've created a user or a viewer survey to help us address,
you know, kind of where we're at with the podcast, things that we could be doing better,
things that we're thinking about doing, get your opinions on, but also just to get to know
you a little bit as well, who our viewership is. So if you wouldn't mind, just jumping over
to the link and you can scan the QR code here that you see on screen to that survey. And we'll
we'll put this on social media and a few other places.
Be super helpful for us.
Again, it's probably about five or ten minutes of your time.
I'll just multiple choice.
There's no right or wrong.
Just go through it as best you can.
And we'd really appreciate the time spent to do that.
It will help us guide our season three,
which will be back January 15th of 2026.
We take about a month off for the holiday season here
just to kind of refresh and recharge your batteries,
but we'll be back in mid-January.
And we're also pleased to tell you that Flanders Scientific will be back as a sponsor for our third season, which super big thanks to those guys.
They've really helped us out over the past couple of years.
And of course, you can head over to flanderscientific.com.
Check out some of their new offerings, including the XMPC, which we'll have to do a whole episode about that particular lineup in that monitor because it's pretty exciting stuff they have coming down the pipeline.
All right, my man.
Well, let's dive into the subject de jour.
Why don't you kind of give us a brief of the question that Jared posed and then we can we can dive in?
Yeah, so Jared said, essentially he said, I'd love to hear your thoughts on dealing with green screen compositing.
Is this something you often deal with?
How have you refined your processes and any tips on approaching your keying?
So yeah, like Robbie said, we want to talk about approaches and strategies.
We don't want to tell you what buttons to click.
but overall the question, I'm sure Robbie will tell you.
So do we deal with green screen often?
Yeah, we do.
It's become a little bit of a specialty.
And I will say it's become more of a Joey specialty.
Rob just I just go along for the ride on a lot of this.
But, you know, the fact of the matter is that green screen,
I mean, we don't have to, everybody knows what green screen, blue screen is,
that kind of stuff is.
but, you know, for as budgets have gotten tighter, location, travel, etc., has all gotten tighter, green screen,
and now to a certain extent, also volumetric stages, by the way, have become a really value-driven way to do things like interviews.
Of course, you know, in, you know, big narrative films, green screen has a solid place for VFX work, right?
Most of the things that we're seeing in our work as colorists, more have to do with stand-ups, interview setups,
you know, that kind of sign, a little less on the on the VFX kind of work.
But, you know, occasionally we'll get, you know, something like,
hey, we've got to do a screen replacement on this TV that, you know,
had a green frame on it or a phone or something like that.
And you have to kind of combine some of these techniques.
But I agree, man.
I think that there's a lot of real big nuance in the controls here.
I think what we want to do is kind of talk about the big picture.
And I'll start out and then we can, I'll hand it over to you to start diving in.
one of the things that I think about green screen all the time is that your success with this work
is less about what you do in post-production.
I mean, it's obviously important.
But it really, like more than any other kind of production, your success with this is going to start in camera on set with lighting and all that kind of stuff.
And there's, again, there's a lot of, there's a lot of nuance to this.
But I mean, I think some of the things that I would just mention are, hey, good camera choices.
Highly compressed cameras, you know, low chromosub sampling, heavy compression, not great for green screen work, right?
You want to shoot, you know, as high quality as you can.
Same thing goes with lensing.
Lighting, right?
Having good, even lighting on the background, lighting for the foreground and kind of building that separation out with some depth of field.
You don't want to have somebody right up on the screen.
And then, you know, also things like wardrobe, right?
Like if you're shooting on green screen, having somebody wear a green sweater, problematic, right?
So there's a lot of production things, and some of that you can't solve.
So I know for a lot of us, you just get a pile of stuff and you're expected to work your magic.
But I think, you know, when, if you're consulting with clients about this kind of stuff, hey, giving them a kind of a cheat sheet or a tip to be successful when they're shooting is important.
And I would also just say that a lot of DPs and camera people, like, sure, they can shoot green screen.
But there are special operators, specialists who do this kind of work all the time.
If you're doing a lot of green screen work, it's worth the money to hire somebody who's like,
yeah, this is my expertise.
I do green screen all the time because you'll just, in the long run, you'll have a much better,
much better go at it.
But I think, Joey, the first question I have for you is, okay, so we get green screen work.
Where is that, like, is this something that we should be doing as colorists?
Is this something that we're like, nope, this is, you've got to go.
for a VFX person, like in your mind,
is it worth a colorist adding this
to their repertoire and why? And then
I guess sort of a second related question is
is there a sort of a
line in the sand of where you're like, yep, this is something
that's a colorist in a color pipeline, color workflow,
I can do versus no, no, no, no, I got to send this off
to a specialist who does this day in and day out.
Let's start with those two parts of it.
Yeah, I mean,
I think it's definitely something that
colorists, if they're interested in this kind of work, should embrace and should learn and should
practice. And I actually really enjoy doing green screen comps. I think it's a, it's a fun challenge,
both creatively to figure out how to make it all kind of sell the shot, but also technically to
really dial in how to get the perfect detail and to really impress the client, make it really
good looking. I love doing green screen work. I think it's just fun. So if you enjoy it, it's a
great way for you as a colorist to kind of add some hours and add some billables to your
project. You know, nobody's going to ask you to do it for free. But like Robbie said,
there are some situations where you're going to be like, no, this is too complicated. This
needs to go to a really talented VFX artist. And kind of the first kind of decision, like
the demarcation line that you're going to come to when a client asks you, hey, can you do these
screen screen composites for me is that one.
So the first thing you look at is the kind of quality of the footage, the complexity of
the comp, how much, you know, am I going to need to make the background?
Do I need to invent the background?
Do we need a designer to make a background?
Is it a photographic background?
Look at all of those equations and kind of think what I usually do with my clients is like,
hey, send me one or two test shots.
I will give you a quick example of what we can do here.
If we want to try to do something drastically more complex, then we should start talking about going to a VFX artist.
You don't want to surprise the client halfway through being like, oh, I said I can do the green screen, but actually I can't.
Yeah.
And when I look at that footage, I think there's a couple telltale signs that will really kind of guide you on this.
This is not a complete list, but a couple things to look for.
Number one, I'm always looking at the quality of lighting on the screen itself.
Right. If you have a dark spot over here and a bright spot over here and a lot of wrinkles and that kind of stuff, like that's a cue that this is going to be a little more challenging, right?
Same thing goes with like if you have a lot of noise in the image, right? Because remember, we're sampling pixels when we do green screen work.
If there's a lot of noise or artifacting or whatever that's going to get in the way of a clean selection of those pixels, you could potentially have some problems, right?
The other telltale sign for me about success in green screen work is, especially with people, of course, is features, right?
If somebody has, you know, big, big hair, you know, light colored hair or whatever that accepts a lot of spill.
Like you see this all the time with people with blonde hair, right?
Where they're, you know, getting a lot of that color cast from either the practical lights being used or the background.
Like, you can evaluate these kind of things really easily without actually having.
to do the work. But I do agree with you, Joey, like from a process point of view, even if you've
done those evaluations and you're like, yeah, I think I can make this work. Like, a test is always
a good, good thing. And I would also say, like, when you ask for that test, like, ask for the,
if your client can, pull out the worst case scenario. And what I mean about the worst case,
the worst case scenario is people gesticulating. The camera moves and half the green screen goes off,
you know, off of camera, now you're looking at a wall, like whatever, those kind of situations
where you might have to involve, you know, some more sophisticated techniques.
You always start with the worst case scenario because then that will make you look like a
hero with the good stuff.
All right.
So let's just assume for a second that you've said yes.
We're going to, you've evaluated some footage.
And you're like, yep, I think we can do this.
I think the first question that I have is that, and part of this is honestly is driven by my,
let's just call it as a resolve user let's just call it fusion phobia right um you know fusion's been
around for gosh i'm probably going on 10 years now and it's still one of those tool sets where i'm like
i'm a little scared of it right i personally i feel like a lot of the time for green screen work
i can accomplish pretty pretty good results right on the color page but like in your opinion like
you got a pile of footage like where are you going to
start this process, right? So if you evaluate it, you think you can do it, it's got, you know,
you have a good run out of a good quality footage. Are you jumping to the color page first?
You know, in the color page, for example, you know, it's different tool sets, right? You got the,
you know, 3D key here on the color page. You got the delta key here over and the fusion.
There's obviously a lot more matte type tools going on over in fusion. So like,
where's that demarcation line for you?
So, yeah, like we were kind of talking about earlier, the biggest driver of this is your
sources and what you need to do.
A lot of people are really surprised when I tell them how much green screen work we do
directly on the color page.
For the more simple things, it's my preferred way to do it because you can deal with
color issues really easily.
The performance is incredibly, incredibly fast, right?
Most green screens on the color page are real time.
So if you have a thousand shots that you need to render green screens on,
you might not want to lean towards Fusion just because of render times.
However, when you get into something more complicated than what you want to build on the color page,
and we'll talk a little bit more about how you can build on the color page,
but you'd be surprised how much you can build on the color page
in terms of using the background, adding perspective, doing some cool kind of things to sell the shot.
But if you need to do a ton of masking, or if you have a really problematic green screen, or if you need to bring in a lot of different elements, like the background is something you're going to make yourself, or if you need to do something with three dimensions where maybe you're kind of faking a camera move, those are all situations where fusion gives you the most possible flexibility.
I was going to say I think that where it really breaks down for me in that regard and to be honest, where I hand off to you more of the time is when we start getting a lot of that like kind of more sophisticated tracking needs, right?
Like especially when we'll talk a little bit about this later about with garbage mats and stuff like that, perspectives changing lighting.
Like, you know, like the camera wraps around and does a curve around somebody.
you know, those kind of things.
You know, that's, you know, those kind of situations.
I think the other question I have about this is, or just conceptually about the choice,
is, you know, to me, the color page, as a colorist, the color page makes a lot of sense to me.
I'm already used to, you know, in a node-based workflow, I have my own color pipeline
built or if I'm using project wide.
Like, the second you jump over to fusion, those pipelines, while manageable, change a little
little bit, right? Like, it's, you know, it's a, it's a different, in the order of operations,
it comes in a different place in the resolve pipeline. So therefore, like, if you're not there
a lot of the time, it does require a little bit of a brain shift. Even if the tools might be
more sophisticated, better over infusion, like, that's why I usually try first on the color
page, because it's sort of like, this is what I already know, this is what I'm already comfortable
with. I'm not using the fusion page day in and day out. So it's like, let me see if I can
you know, get started with what I know.
Two more things about kind of the interaction
with fusion and color and with keying.
One other, and this is a big one,
situation that I almost always jump to fusion for
is if the green screen is problematic
or the subject is problematic enough
that I need the Delta keyer.
A lot of people don't know what the Delta keyer is.
So I just want to talk about that a little bit.
The two kind of big differences between the primary keying tools
and fusion and color.
Now, the 3D keyer from Keeer,
is available in fusion.
So if you want to use the 3D keyer infusion,
you're welcome to.
But the 3D keyer is an amazing green screen keyer.
It does beautiful work with edges and spill suppression.
And out of the box, the 3D keyer does a better job
on green screens than most keyers I've used.
Where I want to go to the Delta keyer
is a couple of specific kind of scenarios.
And I want to back up a little bit here and kind of say,
what is the Delta keyer?
The delta kear is exactly what it sounds like.
It's a kear that works based on delta or differences.
So this confuses a lot of people because they usually just drop the delta keyer on there.
They heard the delta key is awesome.
They drop it on there.
They sample their green and it looks like garbage.
And they're like, what did I do wrong?
The delta keir is not just sampling the color of the screen.
The delta keyer is expecting what's called a clean plate input.
This is your green screen without anybody.
in front of it, right? And then it takes those two streams and makes a key from it. Now, you might say,
I don't have that. I don't have a clean plate. We didn't shoot one. If you're shooting locked down,
it's very useful to shoot that clean plate natively. Yeah, it's like the equivalent of that into the
Delta key as a clean plate. That works amazingly. However, there is another effect in fusion called the
clean plate effect that will take all basically all the green from your green screen and grow it inside.
to like envelop the subject.
It makes a really weird looking image.
But you feed that into your Delta Kier as the clean plate.
And the result of this is it's kind of taking the green just itself,
but with all of its imperfections.
So if the green screen is uneven,
the Delta Kier does a great job dealing with it.
Or, and this is the big one, someone with big frizzy hair
that lets a lot of green through little holes in it,
the Delta Kier does a thousand
times better job with that exact specific scenario than any other kear I've ever used.
So if you grab someone with super frizzy hair, compare the 3D keyer and a good setup
on the Delta keyer, you're going to go with Delta keyer every time.
So that's like, that's the demarcation line for me.
If it's an uneven green screen or a frizzy hair that needs a lot of detail or frizzy clothing
that needs a lot of detail, Delta keyer all day long.
Other than that, I do tend to gravitate towards the cost.
page because it's more familiar and it's more real-time and faster.
You know, the thing that you just mentioned about the clean plate is interesting.
I didn't realize, first of all, I didn't realize that that's exactly how that works.
Plus, I didn't realize the clean plate effect to build it from if you don't have that.
But it reminds me of like, you know, an audio, a production audio guy capturing room tone, right?
It's like, why wouldn't you capture that, you know, clean just so you have it, you know,
just representative of, you know, the scene, I think is super helpful.
So keep that in mind when you're talking to clients as well.
Be like, hey, you're going out and doing a production.
Make sure you record, you know, a minute or so of just nothing but the green screen.
That would be helpful, possibly helpful down the road.
Yeah, absolutely.
And it can be a single frame, right?
If as long as the camera's locked down and doesn't change perspective, like if someone walks away from the green screen,
there's a clean plate.
You know, you're ready to go.
I think the other thing worth mentioning at this stage, no matter whether you're in the color page or in the fusion page, is that you do have to get a lot more comfortable with the concepts of matwork and transparency, right?
And how those can be piped around various parts of the interface.
Because like if you're new at this, you hear me hear me out for a second.
You've probably done something similar to, okay, I'm going to go onto my edit page timeline.
I'm going to stack two clips on top of each other, right?
I'm going to put the background on video track one, the foreground on video track two.
I'm going to key video track two on the color page.
And then I'm going to add an alpha output on the color page, pipe that.
So now I've built a composite, right?
And like, there's nothing wrong with that.
But it shows, it's, you know, I'm guilty of that all the time.
But it is one of those things where it's probably the lowest, you know, the lowest.
the lowest sophistication method of of doing that work.
I think one thing that as you get better and more comfortable with this,
starting to understand that any node really in the color page or fusion
can potentially pipe out a matte image to work with and that kind of stuff.
And so like you just need to get a little bit more conceptually comfortable with,
you know,
because with how a mat works,
how edge refinement works,
how transparency works.
because then you can start seeing the puzzle pieces of how those various things to go together
within the constructs of a node tree either on the color page or on the fusion page
and I'll be honest with you that has been the for me getting better at green screen
that's been the hardest challenge going oh wait a second you used a color generator that
was masked off to do this mat or whatever like those kind of things conceptually just
require a little bit more thinking than you're probably used to yeah and yeah and
you know, that's another kind of one of those fundamental questions you're going to ask yourself when you approach a green screen project, whether it's do I use color or fusion or some other program, doesn't matter.
The next kind of thing you want to think about is how do you actually arrange this in the software?
Because up to this point, we're just talking about doing the key itself, not a whole composite.
So do you do what Robbie just described and put the background on V1 in your edit page, then make a keyable clip with fusion or color?
on top to make your comp.
That's one way to do it.
Like everything else that we've talked about or we'll talk about,
there's pros and cons to each one of these approaches,
and you need to tailor it to your project.
Did your client give you an edit of backgrounds?
If they did, this is a real easy way to get started
because you could just lay all the backgrounds in
and just go shot to shot and work on the keying.
However, the downside here is
if you build your composite outside of Fusion
or color, you lose some ability to work in kind of the three planes of the green screen,
which is I can color and texture affect the background.
I can color and texture affect the foreground,
or I can color and texture affect all of them together.
So all of those things, when you start building your comp,
you're going to realize, oh, I need to match the color between the foreground and the background.
I need to maybe darken the background a bit to bring the foreground out.
then maybe I'll throw a vignette and some film grain on the whole comp together to really sell the shot, right?
Those things, those details get harder when you've separated it on the edit page.
So the next logical question is, well, how can I put the background in my comp if I don't put it on the edit page?
That's a great question because the way I would, like I always think about node trees, whether in the fusion page or the color page,
as single shot specific, right?
It's not like, you wouldn't think of a node tree as,
hey, here's a node tree simultaneously dealing with two or more shots at the same time,
but that's why I mentioned the Matt stuff and that kind of stuff earlier
because you actually can do that.
Yep, and this is where a lot of people kind of, again,
they get surprised when I show them some of the green screen setups
we've built on the color page because they don't realize you open the media tab,
on the color page, drag in your background.
Now you can comp your background using layer mixer nodes right in the color page,
which means you basically can have a stream on the top of your green screen,
a stream on the bottom of your backgrounds.
You can arrange color effects, blurs, whatever on the background,
combine them with a layer mixer.
And then on the output side, after they're combined, you can color them together.
But you can also still in the same node tree, color them separately.
Now, could you go back on the edit page and jump back
and forth between clip to clip. I'm going to darken the background a little bit. I'm going to
brighten the foreground a little bit. Maybe I'll throw an adjustment layer on top of the whole thing.
You can compound it. You can do that. But that's not, that's no fun. It's like it doesn't, it's not
conducive to creativity because you're jumping around. Now, the last thing I want to mention on this is
that's all on the color page. If we want to do our comp all in the color page,
background foreground, foreground, using an external mat. And it's a great way to work. But let's say this is one of
those situations where, hey, I'm on Fusion page.
I needed to do some really complicated masking.
I still want that capability.
I still want to be able to grade both the foreground and the background and them together
completely independently on the color page using the color tools I'm familiar with.
On the Fusion page, you can always, and this is another thing that people just don't always realize when they're just getting started,
if you make a second Media Out node on the Fusion page, feed it the mat from whatever you're
using the key, just the mat, not your whole comp.
And then you go back into the color page, right click on your node tree, add source.
Now you have a mat that you can feed into the key input of any node.
So you can do the comp work in camera space on the fusion page with the background
altogether, then jump to the color page and still use your mat that you built
in fusion with the fusion tools to do all your color work.
So if you do a green screen composite infusion,
Don't think that you are tied into color matching and color grading that comp only on the fusion page.
You can bring the mats over to color.
Yeah, you're connecting.
You're connecting the two pages together.
Yeah.
Yeah.
It's interesting that that's that's more of the workflow I meant about kind of combining the two pages together.
I guess my question related to that was, what do you see the advantages of doing that?
matwork and fusion page
and then piping it over to the color page
is it just more sophisticated
compositing tools are available
over on the fusion page and
obviously the color page has better color tools
is that simple? It's that simple and also
the color page
has the advantage of using your control panel.
Yes, there's a lot of color tools in the fusion
page but they're all kind of mouse
and pointer and stuff.
I'm going to go to the color page
where I'm most familiar coloring, do
my color nodes and
it's just a it's an easier workload if you're doing your comp and fusion let's just recap because
we we've talked about a lot so far and it can get a little confusing so I just want to just real
quick recap the kind of this the strategies number one we could start you know as simple as possible
on the edit page layer things up key something on the on the page you know either on the edit page
or the color page add an alpha you know alpha output just combine those two but it's a lot of jumping around
right you're going back and forth between the two different clips
if you compound those, then you have another layer you're dealing with of complexity, right?
So we can do this all on a color page node graph by adding in external mats, which as you said,
can be your background can be an external mat.
You can use media and combine multiple clips in the same node tree all on the color page to do that.
You could do that exact same thing solely on the fusion page, or you can combine the two.
You could do your mat and key work on the fusion page and then pipe that over to the color page as a new input to then continue doing your color work with the combined composite.
So you really have some choices of where you're most comfortable and what, you know, workflow practices are going to best match the project itself.
Yeah.
And to go back to the edit page example, there are times where you do want to keep the background and foreground completely separate.
For example, if you have a project where you need to render individual clips back to the client for them to reassemble and edit,
the only way to do that is to keep your background on a completely separate clip that you're only using for kind of pre-visualization.
So like I said, kind of early on, there's no, there's no one rule about any of this.
It's all very project.
I get it.
I get it.
So if we have those big picture things out of the way, let's talk about some of the specific, I think, challenges that people run into with.
keen, right? Obviously, the first one is edge refinement. A keen is all about edge refinement,
how smooth that edge looks like, how, you know, how much detail, you know, you don't want to
give somebody a massive haircut, right, just because you can't get a, you can't get a good edge to
it. So 3D keyer, obviously we have some, you know, all the keyers on the color page, you have
some edge refinement, you know, blurring, you know, shrink or grow of the mat,
de-noising of the mat, that kind of stuff. I'm guessing, just, you know, all the keyers on the color page, you
I'm guessing just like everything else, there's some more sophisticated options for that kind of thing over in the fusion page, right?
Yes, there is.
But some of those options also exist on the color page as well.
What I want to say conceptually to anybody thinking about edge refinement on your keys, edge refinement is for the edge.
And what I mean by that is, whether you're on fusion or color or even in After Effects, the goal of using any edge refinement tool is to get a,
good edge. It is not to fix problems in your key. So, for example, somebody's wearing a green
pin on their shirt. Okay. You can get away with cranking the fill white or the clean white up
until that's a solid and you've now ruined your edge, but you've got a done key. Great. No,
no. Use your edge tools for the edge. And if there are other problems bring in up,
power window to track that little pin that's on their their shirt, for example, right?
I think that's a use the tool for what it's what it's for instead of trying to overdo
the edge refinement. So another thing you can do is combine multiple garbage mats or combine
multiple keys. Again, this is completely page or software independent. You look at your green screen,
you're refining the edge. I can get the hair exactly right, but it's messing up their shirt.
their arms or I can get their clothing exactly right but it's giving them a
haircut well you don't need to have it all on one node do a mat that splits their
their head that has one key on their head and then do an outside node or an
invert or another mat whatever whatever your software calls it and do a
different key for the body with different edge refinement rules people get
really caught up in making the entire image work with one effect even if like
Okay, I've seen people try to like crank up these edge of assignment tools to clean up stuff, you know, garbage on the outside of the edges.
Like noise, if there's noise on the edge of the frame where there's some lens vignetting.
Who cares?
Power window the whole thing.
So you're only dealing with the edge with your edge tools.
I think it's a, it's a gem of knowledge what you said.
It's so simple.
I've never thought about saying it that way, but use the edge tools for the edge, not for the key.
Because I think that I'm guilty of that from time to time, right?
you know, clean black, clean white, whatever, just trying to get like, and that, that's, that, that, that, that seesaw back and forth, like, it can work sometimes, right? Like, yeah, sure, you can, you maybe get lucky and clean up the key and refine the edges at the same time, and it's acceptable. But that back and forth is the real challenge for there. If you're not, if you're not, if you don't realize what those edge tools are, they're not a solve for the key necessarily. They're a solve for refining the edge, which I think is great. Now, let me,
let me ask you this because you mentioned garbage mats
and I think that's a
an edge type issue
you mentioned a couple different scenarios on it
but if I separate it out where like the edge
tools that we mentioned are for the edge
things like garbage mats are a way of dealing
with problems of the key is that
is that a good way of saying that like whether that's like
the screen off the edge whether it's like
you know you want to do a holdout for something
on maybe on somebody's body
that is getting keyed by accident
It could even be for it I can get this perfect except for the side of their head
I can just put a garbage mat around that that kind of thing
So garbage mats are for like sort of filling holes or masking off problems of the key
Not really for a cleanup of the edges necessarily
Exactly so you know one example that we just talked about is
There's some lens vignetting around the edge of the screen so your key is perfect in the middle
it's perfect for everywhere that there's a keyable edge, right?
Why would you refine that key anymore?
If it's perfect everywhere you need to see it,
why would you refine that more to make it work with the vignetted green screen
when your subject never walks in front of the vignetted part of the video?
Now, if they do, it's another problem.
But in most cases, the green screen extends beyond where you need to key.
Yeah, and this is harkening back to the beginning of our conversation about this
where I think the evaluation and understanding
what like actually being familiar with the shots
because what I find in those situations, right,
is that like I get the key perfect,
I add a garbage mat,
but then I didn't really watch down the clip the entire way.
Yeah, and their hand moves out of it.
Halfway through, I see somebody's arm
just get cut off in the middle of the shot, right?
Because they then hit the garbage mat
on the side of the frame and all of a sudden their wrist,
like, you know, got cut off by something or whatever.
So it's also important when you're doing that kind of those,
those bulk fixes with like garbage mats or multiple keys that you have to go through the entire
clip too, not just a couple frames on it to make sure it's right.
Yeah.
So now that we've kind of defined edge versus bulk, let me ask you a couple specific edge questions
because I think if we now know that we're not trying to fix the key necessarily with the edge,
the edge refinement tools, but adjust those edge, I think the first place that I always start
is like what level of detail do we need to have on,
because in a perfect world,
it's 100% detail, right?
You know, it's perfectly crisp.
You're not losing a strand of hair,
but in the real world, like, we can cheat that a little bit.
Where is the, where's the balance?
Is it just a matter of, yeah, that looks good?
Or are you looking at something specific
when you're starting to adjust edges of,
Yeah, I don't want to lose this kind of detail or this type of thing.
It really is, for the most part, in my opinion, kind of what works and what looks good.
However, the one thing I will say, every time you're doing that edge refinement, you're tweaking, maybe
you're doing some de-noise, you're doing a little bit of shrink, you're doing a little bit of grow,
whatever it is.
Every time you think this looks good, I'm really liking this direction, you need to go to that
highlight mode and you need to not only go to the highlight mode, go to the highlight mode, go to the
highlight black and white mode if you're on the color page.
And look at that mat at a hundred percent high contrast because you might be introducing
a haircut somewhere where you didn't expect it.
You might be adding a hole or removing a hole where you didn't expect it, especially,
okay, women with dangley earrings.
That earring can sometimes kind of move close to their face to where it would be solid
and then a hole.
And it can, if you add too much like shrink, Matt shrink, you can make it so it.
like, depending on where they swing their head,
their ear might look a little weird and flickery, right?
Those issues, once you kind of get it visually where you want,
don't just trust that.
Look at your mat in high contrast and verify, trust but verify, right?
It will give you an indication.
Now, again, no hard rules here.
If the final image looks great, you know,
it doesn't really matter exactly what the high con mat looks like,
but the high con mat is a great way to spot little things
that you can get better
that you might not see
when you're looking at the whole company.
That's a good one.
And I've noticed that I tend to,
and I think part of it is because I'm guilty of
trying to use edge refinement tools
sometimes to fix larger, bigger key issues.
The one that I notice all the time
is like the haloing effect
that happens with some of the cleanup stuff
and the denoise, right,
where you'll get a little bleed through on those edges.
So like you're improving the key,
but at the detriment of the clean edges, right?
And so like, that's the best thing you've probably ever said on the podcast
about thinking about those two differences right there.
Let me also ask you this when it comes to edge refinement.
I think the biggest one that can, you know,
depending on the shooting situation can be really straightforward
or it can be really difficult is the spill that happens
in how to deal with spill.
suppression and that kind of edge issue.
So just so you everybody's familiar, what I mean is that you go and shoot somebody on a green
screen and let's just say they're fair skin or they got blonde hair or whatever, or maybe
they're too close to the actual green screen or the lighting's too strong.
You can get some spill or some reflectance from the background onto the subject in the
foreground.
And it kind of just looks nasty, right?
You've got to get the screen thing.
Now, obviously, most keyers are going to have something called spill suppression.
What is spill suppression?
What's the goal of it?
When should, do you always use it?
What, like, what's the, like, what is the best practices for that?
Spill suppression is exactly that.
It's looking for the screen color.
So you usually have to tell it what color it's looking for.
And it's kind of doing almost how the color compressor works,
where it's pushing towards,
it's kind of pushing the hue and saturation towards another color.
But it's doing it specifically to the opposite color.
of your screen.
To eliminate it, yeah.
This works better than trying to manually do that with curves or with a primary grade almost every time.
Don't be afraid of spill suppression.
Built into the 3D keyer, it's built into the Delta keyer if you're using it there.
However, sometimes it's too much.
You can always dial it back.
And sometimes it's not enough.
There are OFX built into resolve that is the same spill support.
but external.
So if you take that, put it in another node after your key, feed it your mat.
So you don't want to spill suppress the background in most cases, right?
So feed it your mat.
You can add more spill suppression if you need more.
Usually that's not the case.
Usually if you add too much, they end up looking a little burnt.
But it's funny, in the Delta keyer, the options for spill support, like the preset options
for spill suppression are like, I forget the exact wording, but it's like rare, medium,
rare burnt
but in general
spill suppression is the first place to look
for fixing that and then
you can also do additional color
correction to kind of bring
the background and the foreground together like if you're
in if you're comping them into a really warm
environment and they were lit cool
you're going to want to warm them up in color
right but you're going to want to do that
to them without green on their face
so spill suppression first
then color correction to match the background in the foreground.
Yeah, yeah, okay, that makes sense.
And I've had, I've had really good luck with the 3D here spill suppression.
Like, it can work.
It's very good.
It can work some pretty good magic.
But like I, you're right.
There's other, there are other tools available for that.
Now, one of the, so let's say you got all of this stuff down.
Like, you're getting pretty good results.
You refined your edge.
And again, there's levels of sophistication for, for this as we've talked about.
But the one thing I think everybody who's ever heard or done some green screen work has heard from their clients is something, I'm paraphrasing here, but it's something like this.
It looks like the person is just like sitting on top of the background, right?
It's just like you got a perfect key, but now it still looks like you have two disparate images, right?
You have the background and then you got the person keyed, but they don't really look like they're part of the same entity, the same, the same same same.
thing, right? And I think that that's a common complaint where it's just like, yeah, that looks like
it was green screened, right? And it's just like, how do we get that? So I know that there are a number
of strategies about this. You know, one of the, sort of as a side note, this is one of the reasons that I
think volumetric stages have taken off so well, this kind of thing, because you can really integrate
somebody so much better into a background in a volumetric set up with a screen behind them.
So why do you think that is?
well it's because of the natural spill that happens with and like the way that the light wraps around i'm using a pun here right the light wraps around somebody uh and then interacts with that and goes both ways right the foreground lighting hitting the background and the background lighting it's more of the background kind of wrapping around the person as you if you're looking at somebody right if you're looking at somebody standing on a street corner right you're getting the photons from that person obviously but you're also seeing the photons as they come from the background
and kind of wrap around that person
and it makes it feel like,
oh, this is one plane,
this is one integrated thing.
So I mentioned,
I was joking,
but light wrap,
what is a light wrap?
Why is it kind of a key component,
pardon the pun again,
of really selling an integrated look?
Yeah, and that's kind of,
on the overall level, right,
the idea here is to sell the shot,
to make it not look like it's somebody cut out and paste it on top.
First layer of that is color matching,
which any color should be familiar with,
make them feel like they're in the same color and lighting.
Great, we can do that.
Second part of that is texture.
If there's one is really noisy and one's really clean,
maybe de-noise the one that's noisy,
or if you want some film grain,
do some film grain across the whole comp.
But like you said,
the reason why volumetric stages look so good
is because the actual background helps to light the subject.
And we can emulate that in our green screen with something called a light wrap.
Now, all a light wrap is is essentially taking the background and bleeding it into the foreground around the edge a little bit in a soft and pleasing way.
So if there's a bright blue background, you're going to emulate a little bit of blue coming in to,
your subject. This is shockingly easy to do when you start thinking about things in kind of a node-based
environment like fusion or color. Because you've got the ingredients, right? You've got the background.
Now, let's go way back. This is another one of those reasons why it's really nice to have your background
in your comp, not separate on the edit page because it's really hard to do a light wrap if it's
separate on the edit page. In fact, there's really no good way to do it. But if you have to have,
your background inside your comp infusion or in color so you've got your background you've got your
foreground and now you have a mat right you have at some point in this tree that you've built built a
really good mat that is giving us the exact edge of our subject and our foreground so we can use that
to take some of the background combine it with some of the foreground just around the edges now the
first thing people think is like yeah but i don't want to have it hit
the whole foreground, that's going to look ridiculous.
It is.
How do we get just those edges?
Simple math.
Take your high contrast mat, break it out to another node.
Again, this is Fusion, Nuke, After Effects, any software.
This has nothing to do just with resolve.
But you take your mat.
Okay.
Then you take on another stream, your mat inverted.
You blur one of them.
And then you combine them with,
a difference mode. So essentially you're subtracting the blurred one from the solid one or vice versa.
The idea being you're taking the blur and pushing it inside where your foreground is and giving you
just a visual mat of the edge. Then take your background, add it as another layer as usually an
ad blending mode of really low opacity and use your mat to limit that layer. So essentially all
we're doing is we're taking the mat, we're converting it to just an edge, and we're giving it a little
bit of blur, softness on the inside, and then we're using that to comp the background back on
with the teeniest bit of opacity. And there's a couple things that you can kind of play off each other, right?
You can play off the opacity, first and foremost, it's the overall strength, or the strength
of that blur between those two inverted mat layers to kind of show how much, or try to modulate
how much that light wrap protrudes inside your foreground.
So it's one of those things, and it's not appropriate for all comps, right?
Because if the background is massively further behind someone in kind of theoretical space,
it's going to have way, way, way less light impact on your subject.
And that's one of those areas where LED stages are actually harder to do,
because let's say the background is supposed to be massively far away,
that wouldn't have as much spill on the subject, right?
And they have to light accordingly.
So kind of think about what would sell the shot.
Would this environment they're in,
if it's super bright and futuristic and stuff,
yeah, that's probably going to spill a little bit onto them more,
give it a little bit more light wrap.
If it's a mountain super far away in a dark scene,
probably less light wrap or no light wrap.
One of the things I think about the pre-vis of this part of things too
is that it doesn't, and I've seen this problematic a lot,
specifically with integration with light wrap,
is that people don't really go into green screens
with an idea of what, oftentimes,
what they're going to composite, like they're like,
oh, we'll just key it, and then we'll add in this thing, you know,
then later.
I do think it's important just to say again,
that if you know in advance,
generally what you're going to key behind something,
that can give you a real leg up
in how you actually light the green screen,
you know, as well, right?
If you're putting somebody, you know, on a beach in the Caribbean, right?
Like, you might go, hey, listen, we're going to light this person a little warmer,
a little more yellow, red, whatever, like, like, you know, you know what I'm saying?
Like, you have some creative choices because that's, that's the hard, like,
you're right, the light wrap helps integrate the two layers.
But that's not to say that you, you know, you mentioned, you mentioned this about color
matching.
Like, I actually find that's, like, that's a big part of it too, right?
Like if somebody just looks, they're on a beach, but they look pale.
Like, what's going on there?
That's the most overall dramatic thing.
Yeah, that looks weird, right?
So these things are like light wrap and whatever helps sell it,
but they're no substitute for kind of trying to get the two elements kind of in line to begin with, right?
Yeah.
And the last thing I'll say in terms of kind of selling the shot,
think about adding a little bit of dimensionality,
as in maybe your background isn't exactly the right perspective.
give it a little corner pin to make it kind of work better.
Or give it a little bit of fake depth of field by tweaking a lens blur.
Or even do a little tiny camera move where the subject,
like if there's a little kind of push in maybe,
you can add a little push in where the subject is pushing in a little bit more than the background.
So it looks like as a little bit of perspective.
Not right for every project, but, you know, it's just something to think about.
Think about the fact that you can manipulate this background more than just,
I'm plopping this person on this done background.
Yeah, and movement certainly helps as well.
The other thing I related to this,
I was going to say that goes along with light wrap.
And we've done a lot of this because we've done a lot of stuff
that's like stand up to camera, like presentation kind of stuff.
You know, if you think about somebody standing, let's just say,
I don't know, maybe something's super simple,
like a white psych environment, right?
You know, white sites can be hard to get right.
So a lot of people shoot those green screen,
but then they put a, you know, a simple white background
or a colored background behind.
Naturally when somebody's, especially if they're full figure,
like they're all, you know, head to toe in the shot,
like just standing there on that stage,
you're, that person themselves is going to cast a little bit of a shadow
onto the background.
It happens naturally, right?
If you're standing, again, standing somewhere
and like somebody's lighting you, there's going to be natural shadow.
So I think along with light wrap,
one of the things that I've seen you do really successfully
and I've really liked how it looked.
is adding a little bit of fake shadows into the scene as well.
Like the foreground element itself casting a shadow into the fake background.
Similar approach to light to light wrap,
but explain that a little bit how that works.
Yeah, exactly.
You're just,
when you're looking kind of ways to sell the shot,
think about ways you can reuse the mat you've already made.
If it is like a stand-up shot,
take that mat, corner pin it to be downwards a little bit,
blur the absolute crap out of it.
so it's nice and soft, and then uses to feed either a really low opacity black solid or even just a color effect to dim that area of the background a little bit.
And you can make a fake shadow that goes in the direction of the light of your comp because that's another area where things can look really weird.
Right. If the light is coming from the left side of the screen on your background and coming from the right side of the screen on your subject, that's never going to look right.
you could play a little bit with like the really cool relighting tool.
Again, giving it your mat you've already made so it only affects the subject.
Or you could play with it on the background.
But you might not be able to get what you want just with that.
But you could do power windows, gradients on the background to kind of emulate the lighting direction of the foreground if you want to.
Or depending on the person, maybe flip the background or flip the person.
Just think about different, you know, look at the scene and just think, how can I sell this shot?
And then think, now that I've got the background and the foreground separately, and I have this great mat that I've already spent all this time making, can I do it with a manipulation of the background?
Can I do it with manipulation of the foreground?
Or can I do it using something I build with this mat that I've already gotten my node tree ready to go?
In reality, it's usually a combination of all three.
The last thing I want to ask about, and then we'll wrap this up.
I keep saying that, but this is really the last question.
you've tried. You've really, really tried and you're not getting the results that you want.
And so you've said to yourself or to the client, hey, look, we need, I need to focus on color correcting this.
I tried as much as I kid with my skills. I need to hand this off.
What does that look like if you're handing off keying work to a dedicated visual effects person?
what are you giving them for success?
What are you asking for, you know, back from them?
Is it a composited image?
Are you asking for the different components?
Like, just speak to that briefly,
how those, that handoff works in both directions.
Again, it's all very project dependent,
but ideally you're going to have a conversation
with the VFX artist or department.
Hopefully you can all be on the same page color workflow-wise.
If you're working in ACEs, that's really good.
because I can give them an ACE's plate of the foreground
and of the background and they know they can work in ACEs.
They can preview under an ACE's transform.
They can get client approval under an ACE's transform.
They can send me back an ACE's image of the whole comp.
I would rather get the whole comp because like we talked about,
there's so many things that you can do to sell the comp
by manipulating the foreground background together.
If you're, if it's such a hard comp that you're not doing,
you're sending it out for VFX, you don't
want to limit what the VFX artists can do, right?
So we want to give them basically in an idea in my ideal world,
ACE is everything foreground background or it might be a background
that they're making. So in that case, just foreground and they'll give me back an ACE's
image, but if possible, I'd also love a high contrast mat from them.
So if I do need to do some stuff to integrate it into the rest of the grade,
I'm not power windowing something that he already did a very intricate mat for.
Not always a possibility for every workflow, but it can be really useful if that works.
Yeah, the idea of, you know, layer DXRs, cryptomats, that kind of stuff comes to mind because I think that is, you know, that's the hard part here, is create.
It's not the color.
You're likely pretty good at the color work, right?
But like if it's that keen work, especially with a lot of high.
high motion. We didn't even mention motion blur, by the way. That can be a real difficult
thing to deal with in green screen work. So you want to try as much as you can eliminate
some of that motion blur. But I agree, getting some of the mats up back. I think with the
handoff to the effects, the one thing I always think of is, okay, when you get to that point,
you've probably already done some thinking or some work about the color correction, right? So the one
thing that you might want to have a part of that handoff discussion, aces or whatever,
is some sort of viewing transform
that they can look at your
color work in context, right?
And the important thing to understand about that
is that that's not something you want to bake into the shot.
So if you're doing a handoff of the green screen work,
you don't want to bake your grade
into either the foreground or the background plates.
You want to give that to them as clean
and as raw as possible.
But if you've already done some color work,
you can hand that off as some sort of transform
a lot or whatever and go, hey, while you're working
on this. You can put this in line your pipeline to view with what I have in mind. So when
they render out, they take that off. And guess what? You just reapply that color work back on
your end and Bob's your uncle. It's now like you have the same look, but with the keywork in place.
Last thing to mention in this process, because a lot of clients don't understand this,
especially clients that don't do a lot of visual effects and a lot of keying and a lot of tracking
is frames matter. Like they might be like, oh, can we get this back with hand with like,
like 12 frame handles, that's fine for a grade normally, but if I'm sending it off to VFX,
it needs to be frame to frame and the edit and everything needs to be completely locked
because you do not want that VFX artist working on frames that are never going to be used.
That is literally just flushing money down the toilet.
So if it's...
Also, you're getting to a VFX artist, you want to be cut to cut exactly to the frame.
Yeah, yeah, exactly.
I agree.
So good stuff, man.
I think there's a lot to digest here.
And obviously, I would just say, I give a couple of resources,
and we'll put this in the show notes.
A couple of years ago, buddy mine, Jeff Foster,
wrote a great book.
It's available anywhere you can buy books
or Kindle versions or whatever,
called The Green Screen Handbook.
Again, this is by Jeff Foster.
He really goes into a lot of nitty-gritty about production tips,
and there's some other books like it,
but I definitely recommend that one.
I would also recommend just the plethora of VFX tutorials from VFX Society and other places that really dive into the nitty gritty about King considerations.
And don't look for resolve specific tutorials.
A new tutorial will help you in fusion.
You can gronk so much from guys, you know, people doing, hey, I'm going to show you how to do this in nuke today or in flame or whatever.
And you might not be working on those platforms.
but the concepts and the conceptual work is really important.
Like, oh, well, Flame has this really cool tool.
How can I emulate that kind of thing over in, you know, Fusion or Resolve or whatever?
And then also, I would just say that in general with this kind of thing,
flagging problems and difficulty early in the process is better than later in the process, right?
You mentioned tests at the top of the show
And over-deliver
Totally, right?
I have seen so many people
Just get ruined by
You know, things that they over-promise that they could do
And then they go, oh, well,
I can't, it turns out I can't do this
As good as I need to, right?
Good stuff, man.
Jared, I hope this helped you out
A little bit conceptually think about some of this stuff
And like I said, for some of the details
there are a lot of other resources that we've mentioned a few of them.
Also, just a last reminder, we are all on shows on Instagram and Facebook.
You can find announcements about new episodes on those platforms.
We're also on YouTube, of course, all the major streaming platforms, Apple Podcasts, Spotify.
You can always head over to the Offset Podcasts with this link and QR code right here.
That's where we, one of the places where we distribute, but also have show notes, transcripts, and that kind of stuff.
And then just as a reminder, it would be a really, really, really big help if you can
could spend five or ten minutes with our viewer survey again right here at this link below.
Take you about five, ten minutes to fill this out.
If you wouldn't mind doing that, that would really help us as we embark on season three,
which we're taking a month off.
This is our last episode of the year.
We'll be back mid-January of 2026 with new episodes, new podcast features, all sorts of fun
stuff.
So we really hope you to come back in the new year and join us for future episodes.
And again, thanks again for making 2025 and the second season of the Offset podcast so successful.
We really appreciate all the feedback, the kind words, et cetera, that we've received.
And here's doing it again in 2026.
So until then, happy holidays.
Stay safe and our best to everybody watching to finish out the year strong.
So for the old Offset podcast, I'm Robbie Carmen.
And I'm Joey Deanna.
Thanks for listening.
