Phil Rhodes

Already yellow!

This is a particularly illuminating case in the logic and ethics of working for a knock-down rate. Anyone given the opportunity to work on a project of the quality of Northfork, for more or less any amount of money, would be crazy to turn it down. Of course, you only tend to be offered that opportunity if you have the ability to do it justice, but the point remains: some low-paid or unpaid jobs are massively worthwhile, and some - most, frankly, the overwhelming majority - aren't.

I agree wholeheartedly. In my view haze has to be controlled down to incredibly low levels, on the order of just being barely visible, to be appropriate unless there's a plot justification for it. Then it works well but it's amazingly difficult to maintain. P

I was going to say, maybe experiment with shaped mattes in front of or behind the lens.

Got any footage we can see?

I'd like to find some really narrow ones, more like the front wheels of a wheelchair, so that they can pack down between the two shelves on a collapsible cart (and the wheelchair ones are non-marking). All the collapsible carts I've seen have used wheels with tyres at least 2" thick, which makes for an unnecessarily bulky package even when it's all collapsed down. It needs to have pneumatic tyres, of course, or foam-filled ones of equivalent performance, but I'd almost given up on finding an off the shelf solution.

The big issue there is that the computer monitor won't be displaying each of the video frames exactly once (or a consistent number of times) and even if it was, the playback won't be synchronised to the camera. This means you're likely to get film frames with two video frames exposed into them. You may also face flicker issues as many modern displays use PWM-controlled backlights. If I needed to do this I'd rig something up to use a camera with a stop motion mode and go frame by frame. It'll be slow, but with testing and enough screwing around it ought to be possible to make it look OK. That assumes the camera is usably steady in stop motion mode, which I understand some super-8 cameras aren't, particularly. You may find you need to apply a very odd colour correction on the computer to make it look as you want on film. Definitely don't expect it to work first time. P

I want a cart, and I'd resolved to build one too. Casters are tricky, especially if you want something that can be packed down reasonably small. Thickness is an issue. I'd resolved to fabricate them. P

Much as I'm likely to be publicly birched for saying this, I thought that some of HBO's excellent Chernobyl had this problem (click to embiggen). I mean, is that really what hospitals look like, even in the late Soviet era? Can we maybe see the nice person's performance? Ah. Yes. We can. Yes, yes, I know, you can justify anything, it's the feel, it's the look, but there is a limit. Given how relative this stuff is, if there's literally nothing in the frame that provides a near-white reference, you're just turning the brightness down.

Yeah, but you're not trying to shoot a night exterior to make it look like Blade Runner with a lighting budget of $3 and an FS7. These things matter. P

This is a bigger topic than people think. Traditionally there was, as far as I can recall, a tendency for features to have smaller title text than TV. I have no idea why that is. Primitive caption generators were available from - what - the 60s? It was recognised early that any feature of the image - such as a horizontal stroke in text - needed to be at least two lines high to avoid horrific flicker, and in practice more than that to avoid interlace twitter, and to provide for a bit of antialiasing. The very earliest titles which were additively mixed in from another synchronised camera aimed at artwork, but the resolution and interlacing issues pertained still. Certainly for most of the time they've coexisted, feature films have had massively more resolution to play with - not now, though. Whatever the background to it, the situation now is that Netflix shows, with their not-so-delusional grandeur, often have titles in text three pixels high because, hey, that's cool.

See, what you do is, you get what's called a low pressure sodium light - they were just about still in use back then, you can get 'em on that thing that unexpectedly replaced eBay in 2028 - and you put it just close enough that it barely lifts the blacks into a sort of ill-defined, shifting, grainy hemi-demi-semi exposure that contains just enough detail to be distracting and enough grit and fuzz to really mess with noise reduction systems. But won't that make it look like someone quite literally took a dump on the frame, Grandpa? Yes, son. It will. Yaaay!

I just wrote a piece on this. I look at it this way. Let's say the fastest film stock, under normal circumstances, was and is around ISO 500. Cameras are now routinely good to at least twice that, if not more, and do better when underexposed anyway. Modern lenses are much more intended to be used close to wide open than ever before, so credit another stop to that. Much of the lighting now uses a quarter of the power it used to, assuming LED is more affordable than HMI. So shots in 2020 are overall something like three or four stops more sensitive in terms of lighting budget to signal level than those in (say) 2000. That means your 18K HMI can now be replaced with something you can plug into the wall in most parts of the world. Sounds great, eh? One of the problems, as Stuart so accurately says, is the tendency to use this as an excuse to just turn up and shoot. Even if people choose to light, though, there's at least two other issues I can think of. First is the issue that even if you don't need the 18K, and possibly not such a huge generator to run it, you still need permission from whatever authority you're working under to put one on the street, and that's probably at least as expensive in many places as the light, the generator, and the crewing. It will save money, but it's not as simple as considering the purchase price of the light. Yes, in some circumstances, really big LEDs can be walked into place on battery power, avoiding official involvement. HMI sun guns have theoretically been able to do this for a while but until recently cameras weren't fast enough to allow them to replace the big stuff. Certainly very high power soft lights can be achieved with things like the Aladdin Fabric Light from wall socket power that would require big HMIs and diffusion to do any other way, but you still have to rig it on a street corner and that's still spendy. The other problem is that working at four stops higher sensitivity doesn't give you the same look on, say, a night exteriors as we got in the famous stuff. Blade Runner was shot on film about as sensitive as an Alexa with two stops of ND in it on lenses that, to be nice, perform best at 4 and above. Shooting night exteriors under street lighting plus movie lights isn't going to look the same as if you shoot on something literally six effective stops faster and with a much lower contrast toe. The movie lights can now be dimmer, thanks to high sensitivity. The street lighting is what it always was. This leads to muddy, noisy, tinted shadows full of unwanted spill from practicals that the old school just squelched. P

I was going to say... mathematics, people.

You can use low cost LEDs, you just need to run tests and figure out what you're going to get. I have a couple of lights that take MR16s and they're almost always loaded with very cheap, very blue domestic LEDs because I find it useful. Obviously, you're not going to use it as a key light for your sympathetic lead, but it has its purpose. P