Camera with integrated lighting

I’m confused? Where is the lighting that’s integrated with the camera? This is just post processing? Is the title right? It’s confusing whatever
 
Looks like a software solution, which reads the correct lighting for people in the foreground from the background.
 
Then this is more of integrated grading and not integrated lighting. Lighting is not just boosting values to match foreground and background. Lighting is actual light so you don't have to fart with values. This is a consumer feature designed to compensate for laziness and a lack of talent or skill and nothing more. Because users just can't seem to grasp the concept of why their video calls look worse than any VHS I ever shot and devices need to make up for that.
 
I'm lightly horrified at ramifications of this (as if we as DP's haven't already lost enough control when we are shut out of color, now our footage can get re-lit as well). But, at least I'm entertained by the Peter Lorre-esque VO on the clip... "hold on to your papers"!
 
I saw this exact same presentation like 6 months ago but without the Peter Lorre VO

But I can't believe that people care so much about their zoom call backgrounds that replacing them is the first marketed application for work like this
 
At some point no real lighting will be done on set. They'll just flood everything with flat even light, capture raw and do all lighting in post. And if light field capture becomes viable, you'll be able to change the angle and lens in post too.
 
Batutta said:
At some point no real lighting will be done on set. They'll just flood everything with flat even light, capture raw and do all lighting in post. And if light field capture becomes viable, you'll be able to change the angle and lens in post too.
Click to expand...

maybe but I'm guessing there will always be a qualitative advantage to at least approximating your look on set, and so high-end production will only use these post tools for tweaking and embellishing or accomplishing things that can't be done physically
 
ahalpert said:
maybe but I'm guessing there will always be a qualitative advantage to at least approximating your look on set, and so high-end production will only use these post tools for tweaking and embellishing or accomplishing things that can't be done physically
Click to expand...

This isn't going to be desired by creatives, it will be desired by the suits and the money people who want maximum ability to tweak anything they want after the fact. The way Marvel makes films these days is practically halfway there. If they can go all the way with this approach, they will do it. Directors will just be on set for blocking and shaping performance. Everything else will be handled by discrete units.
 
Batutta said:
This isn't going to be desired by creatives, it will be desired by the suits and the money people who want maximum ability to tweak anything they want after the fact. The way Marvel makes films these days is practically halfway there. If they can go all the way with this approach, they will do it. Directors will just be on set for blocking and shaping performance. Everything else will be handled by discrete units.
Click to expand...

Yes but only if the technology/techniques are good enough.

Marvel films are really beautiful. (Although not beautiful enough to make ME watch simplistic children's films that I feel so little emotional/spiritual/philosophical connection to.) The productions put a tremendous premium on the lighting, composition, and effects. Marvel is one of the last vestiges of Hollywiod perfectionism.

They could cut more corners NOW if they valued executive control more than production value.

And there are things they "can" do now that don't work very well - like Henry Caville's mustache removal for Man of Steel (or whatever it was) reshoots. That SFX snafu was widely mocked, and by laypeople, not film people.

My guess is that they'll still shoot the scenes properly lit or mostly lit, and use post-production to enhance or tweak. And then if a suit decides they should have done it differently, they'll attempt to relight it in post and see if they can do a good enough job that they avoid a reshoot.

But things like Birdman, where they used tons of power windows and 2D relighting in post effects to compensate for the lack of lighting they could do while shooting long takes with 360-degree coverage - now they'll be able to do a better job of it faster.

But lower-budget productions (than Marvel) will probably go to town with the technology because it'll end up looking better than what they could have done anyway. Or, for example, documentary interviews where some cash-strapped producer really wants to shoot the interviews themselves anyway and spare the expense of hiring a DP. And the lighting can be done in post with finishing funds and matched to a cinematic style that they settle on during editing.
 
Marvel films change more after the shoot than you might imagine. For example, they usually have practical Superhero costumes on set that they completely paste over with CGI, with designs often changing. Even suits like Spider-Man's or Black Panther's which you'd think wouldn't be difficult to do practically, they use CGI. If you watch behind the scenes footage, they BARELY use sets, just the bare minimum for the actors to interact with and the rest is CGI. I see the trend increasing.
 
It takes too much human labor to use, make, build physical items/matter to look like what is now expected in these movies. And as they become increasingly complicated, eventually anything in this category will have to become almost 100% computer-based unless it's supposed to look unrefined.

It's the same for the lighting; the creativity of it is unappreciated by most and you can mix up lighting probably in dozens of different ways for most scenes by closing your eyes and throwing random light around and most people will not even notice.

So when it becomes possible to do it with your fingers in software in which the differences soon will become unrecognized, it's the next choice in the continued transition to machines completing the tasks so more content, more money can be made in a tighter window.
 
Batutta said:
Marvel films change more after the shoot than you might imagine. For example, they usually have practical Superhero costumes on set that they completely paste over with CGI, with designs often changing. Even suits like Spider-Man's or Black Panther's which you'd think wouldn't be difficult to do practically, they use CGI. If you watch behind the scenes footage, they BARELY use sets, just the bare minimum for the actors to interact with and the rest is CGI. I see the trend increasing.
Click to expand...

Yeah I've seen the BTS. They wouldn't be doing all CGI if it weren't good enough to cut the mustard. Ok that's your point - that post lighting WILL get good enough. But for the sake of argument, I think they'll still at least rough in the lighting on set and embellish later because I think the result will be better. and if they can do so many alterations afterwards anyway, then they won't have bought into the final look anyway. So, similar to wearing practical costumes that they embellish
 
NorBro said:
, it's the next choice in the continued transition to machines completing the tasks so more content, more money can be made in a tighter window.
Click to expand...

I think this idea of machines completing tasks is a bit of a red herring. Lights are machines. Humans are using lots of machinery to complete practical lighting designs.

"AI" (which isn't really intelligence anyway, just more sophisticated computer programs making more of the decisions at a slightly higher level in the decision hierarchy) isn't going to be about making decisions for humans, it's going to be about saving time and laying off technicians.

Photoshop now has these AI "neural filters". The first final release filters available are for skin smoothing; and there's one for matching the color and texture styles of paintings or other source material. There are a couple others available. The filters don't make decisions for you, there are still many parameters that the human controls. The filters just automate some of the decision making of what details should remain (eg stubble) and what should be blurred (eg pimples)

Or theres the "Infinite Color" panel which can spit out color remixes of your photo at the push of a button. And you can click repeatedly and cycle through an infinite set of random choices until you find one you actually like. The computer does the heavy lifting, but you make the choices
 
It's a red herring to those who seek comfort in not thinking about the future and change.

It is certain that routine, everyday tasks will be taken over by machines without human input while others will always require human input.

Just depends on the process. But what you're describing in software is elementary and had its groundwork beginnings in the 90s. People have much bigger plans for automation.

Part of it is deep learning and recognizing the task at hand and applying that "knowledge". Surely there will be errors but there is a long way to go.
 
To further elaborate and use an example of something you're fond of: AF

Canon's short description and insight into their operations:

"The system is based on an algorithm that teaches itself by scanning millions of images. The system essentially learns how to recognise the heads of people, even if they're skiers wearing goggles, racing drivers wearing helmets, or gymnasts upside down or even facing away from the camera. Deep learning is a huge leap forward from the intelligent autofocus systems of previous Canon cameras."

https://www.canon-europe.com/pro/stories/intelligent-autofocus-explained/

___

The camera learned what a human head is and is making decisions on what to focus on.

This is one out of I imagine hundreds of thousands of ways machines can learn to complete tasks and make decisions that doesn't just require them to work around conveyor belts in factories.

Expecting a machine to have a human brain is illogical, but they will eventually be a part of society and will be trusted to do things.

It is a no-brainer.
 
NorBro said:
It's a red herring to those who seek comfort in not thinking about the future and change.
Click to expand...

I fully expect that most of my editing work will eventually be handled by machines. BUT the producer will still be there, giving top-level creative commands. Just to a computer and not to me. Even if it's more frustrating and requires more rounds of input from the producer to get the desired result, they'll save a ton of money by not hiring me. And some parts of the process will be far faster for the machine to handle.

But there is some editing I do that will never be replaced by machines. It will be a producer hiring me and then me using more advanced software that speeds up implementing my creative decisions.

It's about knowing what computers/software are inherently good at and what they're not. I studied a few computer languages, and the fact of the matter is that software is a passive machine, no matter how complex. Like build a better mouse trap. and AI is just software. "AI" is a marketing term. (Of course, one could argue that the human brain is a similarly passive machine.)

Face-recognition: excellent example. The only reason it's so useful is because so much of the time faces are what we want to focus on. But even then, the machines have a tendency to focus on the wrong faces or not to know when we want to shift to someone or something else. The top-level creative commands are still required by humans.

Much of the work being done by machines is about cost-savings. You can still get a superior result from the human touch. The question will be when and where people will pay for the difference
 
It's always really important to separate the personal creativity from direct decision-making in automation in these kind of conversations because the technology is not yet neurologically linked.

Machines cannot yet predict what we're thinking so any kind of example of this is nonviable for them and will defeat the purpose of their existence.

There has to be a system in place for this to work.

You can teach a machine how to edit and how to make 7 different videos for each day of the week and when to upload each one.

But if a human is constantly changing its mind and one day decides to add a shot of his or her cat or dog in the video or what it had for dinner last night than this doesn't work. You know?

So yeah, he or she will then always have to be there because he or she is the only one who knows what it wants.

___

P.S. With AF, it's only randomly focusing now because people don't have chips inside of them yet. I'm not saying we will, but we will have to if we want the machines to learn more about our brain activity.

But this has of course already started with eyeball movement.
 
Last edited:
You must log in or register to reply here.
Back
Top