Robin Phillips

modern keying tools can handle a lot, but the reality is that this is probably going to require rotoscoping. granted, even that is not a big deal anymore with some of the advanced roto tools out there. even after effects has a rotobrush that can get most of a human figure (or parts if you do the mask in layers) set against even complex backgrounds. Action VFX has a bunch of basic tutorials that are really good at getting across the basics, check this one out to get an idea of how some of these roto tools work As for the shadow cast, honestly the green screen isnt the concern so much is do you want to have that shadow projected onto the TV plate. Thats a little more advanced, but is a pretty common thing to do. the usual cheat is to take a copy of the mask, make it dark and transparent, then warp it till it looks right on the bg plate. You might also be able to salvage the shadow that spills onto the green screen too, but I wouldnt get my hopes up on that. Regardless these are things even junior artists should be able to accomplish these days. One thing to note, make sure your scans are optically pin registered. most scanners can do this now, but you dont want to use like an earlier spirit 4k that doesnt have the feature, otherwise your compositor will be fighting the scanner gate weave. The Scanity (fotokem has one), the regular Scanstation (lots of vendors have this one) are examples of scanners that have optical pin registration. The arriscan has a true physical pin registration and that feature was always the gold standard for film VFX, but a true physical pin registered scan costs a lot because the scanner cant run at full speed. Back in the day it was always also nice to get an overscan so the compositor would have the perforations to stabilize if need be, but that shouldnt be vital so long as you're using a late era film camera with good registration (others here can speak to movement registrations better than me) and an optically pin registered scanner. Though if you're using like a 1st generation BL, you might want to get an overscan.

the light streaks are made by the physical pull down of the film during exposure, so even a rotary shutter digital camera like the Alexa Studio can't do this. replicating this is going to be a post effect. I havent looked for one, but its entirely possible if not likely someone has made this as a plugin. Your other option is to just shoot film for the relevant shots, say with a 435, and find a vendor with the timing box that lets you do this offset, but unless you shoot short ends its gonna be pricy.

have you seen the footage of the actual trinity test? its basically a flash and then a mushroom cloud similar to a properly executed gas bomb. you can see one of the color shots of it here Here is also some black and white footage which held its dynamic range better Much of what we got use to seeing publicly were either vastly larger bombs, or footage turned on stuff that we wanted to observe during detonation. Those test shots of the trees being hit by the shockwave or the houses' paint burning just before the structure blows are far more terrifying than a straight on look at a gen 1 US nuclear bomb. I believe it was the H-bombs that really introduced the moisture cloud shockwaves that are associated with the tests in the pacific. Weirdly I think that also just depends on the conditions in the air around the detonation zone. You can see that sort of burst shockwave in the 2020 Beruit explosion, which was, and I emphasize we're talking relative here since we're talking about nuclear weapons, a somewhat small detonation but the nature of it and the moisture in the air resulted in one of those dense cloud shockwaves. So I think that if this had been approached digitally, the look dev would have been the actual trinity test. As you can see above, it kinda looks like a gas bomb with a center that went a little faster than its outer component. So if accuracy was part of the goal, I dont think it would have looked any more impressive.

is there a list of locations showing 70mm 5 perf? seems pretty easy to find the list of 15/70 locations but searching for the non imax 70mm version just brings up search results for... the imax 15/70 version

Its Greig Fraser, and hes put in the work with his collaborators to really dial in digital and make the most out of a given sensor and make something unique out of it. I dont think that can really be said for a lot of the work thats out there. And yet they still saw value in printing to film and re-scanning it for the final image.

Film cameras are more or less just a box that holds the film. Your "look" comes from which stock you use, how you expose it, what lenses you use, how you color it, what gauge you're using (65mm, 35mm, 16mm, 8mm) etc. This assumes you're shooting with what we'd consider a typical crystal sync camera for production use. Now if you have film registration problems that can introduce excessive gate weave. There are hand crank cameras whose "look" will be dependent on consistant turning of the crank. You can get a very slightly different vibe from the cameras if the shutter is more side mounted than bottom mounted. But thats really about it. As for the digital cameras having their own look, the truth behind that depends on if you're using proper color management and how much time and money you want to spend manipulating the image to look a certain way. As David pointed out with Yedlin's demo, its certainly possible to match different systems to either film or whatever look you're after. But What Yedlin's demo doesnt demonstrate is the work required under the hood to really nail that. If it was as simple as he suggests, everyone would be doing it and Arri would have pre-packaged their cameras to perform in that manner as a turn key option. One reason to shoot film if you want the film look is to just avoid having to deal with all that potentially costly post processing.

there are angenieux s16 zoom lenses out there, some very nice late model ones. The canons will be easier to find, but IMO those lenses have a little more chromatic aberration and a slight magenta vibe for my tastes, and its really tricky to correct out in the DI if you dont like it.

these days its usually just shooting at the slow mo rate the whole time, and doing the speed change in the computer. on film cameras like the 435 you could speed ramp while maintaining exposure during the take with the right controller, which can be really cool, but its tricky to nail and expensive if blowing through tons of 35mm

the thing is, if you're renting you at least need to have someone else you know with the same camera nearby who rents in case yours fails. Fortunately I know folks who could potentially sub in for my cameras, but the reality is I really want to find a 2nd SR3 Adv both for parts but also as a backup camera if/when I start renting again (which sucks now, market is an order of magnitude price different from when I got my cameras). You dont want to be the only rental option for a film camera, have it fail, and then have an entire production pissed at you, especially if they get litigious. Even if your contracts are solid, someone can make your life hell if they really want. TBH I think anyone getting a private 35mm film camera these days needs to be doing it for personal projects, and maybe you'll send it out with a friend once in a while. Hell the only reason I have a 435 was because the price was right and I was able to acquire everything I needed to run motion control and stop motion, which protects me for shooting VFX stuff on film on for hire or personal 16mm projects. But I'd never have bought it if the camera hadnt been part of a pandemic liquidation sale. If you really want a personal 35mm film camera, honestly I'd be looking at a 2C or 3 with crystal base, and if you really wanna be in the rental game maybe look into lenses that you know will rent and be durable enough for rental. That being said, Alan Gordon has a 4 perf 535B for sale for under 14grand which might be a steal for a sync sound 35mm camera these days. I just would not count on one of these cameras being a money maker.

Anyone know if this is going to affect tungsten bulbs? https://thehill.com/policy/energy-environment/3932145-energy-dept-to-ban-sale-of-some-light-bulbs-heres-why/

AM camera was selling glue on replacements a while ago.

This is such a weird conversation to me, that a $250,000 camera when new is still too expensive at $5k. You are aware that the ability to own any of this gear is a super recent phenomenon right? Is it "financial gate keeping" that an Alexa 35 costs 80 grand? or is that the market price based on costs and demand? No one is entitled to any of this stuff. For nearly 100 years the only access to this kind of equipment was through rentals on shows with the budgets to afford them, and you learned by working on those shows in lower tier gigs or through classes and demos. It has never been the case that filmmakers could afford to fully own 100% of the tools necessary to create a hollywood scale feature film (save for a hand full who became billionaires or owned studios), so it make no sense to expect that to be a possibility. Thats not financial gatekeeping either, its just economics. But since we do live in the future, TBH you can pick up an FX30 for 2 grand and it'll outperform a camera like the F900 all day, and that was a main stay of mid 2000s tv production. Is it an F35? no. Is it 35mm film? no. is it more than enough to learn to light and shoot with for chump change compared to the highest end of the market? absolutely. Same with older DSLRs. Hell, you can make a pretty decent looking movie on an iphone if you want. Building your skillset is what matters most, not access to a specific sensor or filmstock.

I think you're view of the market is too narrow. if there is some specific application necessary for a digibetacam workflow (local tv news that hasnt upgraded maybe? doesnt have to be US domestic), or for spare parts or the like. Plus sometimes if you need period footage you'll want to shoot on a period camera. GLOW did this, retrofitting some SD NTSC broadcast cameras for certain shots. Thats not a sunk cost fallacy scenario. In the cast of an F35, it can be useful in a sports scenario, capturing blank gun fire as an element cleanly (to avoid rolling shutter issues), or just because someone likes how it looks. Thats no crime. Hell, NASA has a ton of research footage of the recent SLS launch that was shot on 16 and 35 high speed along with their digital because the film held the difference between the highlights and blacks better than the high speed digital. In almost all senses the photosonics cameras are obsolete, yet they allowed for a clearer picture of what was happening with the engines than the newer stuff could. Ultimately, though, Im not quite sure why you care. Like, who cares if someone wants to use a certain tool that you dont want to use? What matters is what they do with it, and if it achieves the result they want. After all, what we're talking about is art.

the answer is, enough to keep the market price of the camera still at around $5k.

I've found 4k is almost always worth it, especially if theres any chance of VFX work. An arriscan for 1.3x the scanstaion sounds like a pretty good deal, though obviously if you're sending a ton of film that will add up. IMO if you can do a true RGB scan thats usually optimal. That being said, if you do not have a machine that can handle the mass of DPX files and that lab wont transcode for you, the scanstation may be the best option. The best scanstation 6.5k footage I've seen for 16 has always been when it runs in HDR mode. If you go with the scan station, ask for prores 4444XQ over traditional 4444. Its a subtle difference, but noticeable.

I mean, I just listed some of the reasons why it still has value. I dont think the F55 is remotely in the same class image wise. The alexa is debatable, IMO the F35 makes for a more filmic and pleasing image out of the box through hdsdi than the Alexa except in certain lowlight conditions, where some of the CCDness of it can get a little wonky. But even then, the Alexa classics also dont have the best low light performance in the universe, though they dont have the weird shit that can sometimes happen on a very under exposed CCD picture. The market is the market, and so long as people are willing to pay 5k for an F35 thats what its price will be. Same for the alexa classic or any other camera really.

its picture rivals or beats the alexa classics, even the mini in some circumstances. plus that global shutter. with ultra primes it upscales to 4k in post with no problem

IMO if you want the 7219 look, shoot 7219. I normally shoot that at 400 and find it fine, 320 should give you more small grain response and a little more perceived sharpness. I would light a little more contrasty though on 19, not just to lean into its native properties but to add even more perceived sharpness. Remember we also live in a world where you can do subtle sharpening in Resolve without making the grain pop too much. Hell, if you really want you can use neat video to degrain, sharpen, then mix back in with the source. That method works great too, but boy its a resource hog. If you need a reference, check out "the wrestler", which was mostly 19. it looks great. 500T is perfectly sharp when done right. That being said, make absolutely sure your camera is as sharp as it can get. On the Aaton XTR series its FFD is 52.00mm dead nuts. My SR3 Advanced is sharpest at 51.99mm for whatever reason (differences in pressure plate tension between the two systems?). Make sure your lenses are also collimated to the sharpest possible. I've seen professional stuff, usually ads, where its clear they used 16 and its was a little soft, and to my surprise they are often shot on the 416. My best guess on these is that they're out of spec, or the flange depth is technically within tolerance so the camera still goes out, even though its not dialed in optimally. If you do go 200T, TBH I shoot that at 160 for best results.

the InCamera guys in the UK have some videos on this. One thing to be aware of if you go to 200FPS is that any little debris you have floating about (ideally being kicked up by a fan) may fall too slowly. I'd recommend a day of testing if you can afford it

you can send raw, though if you have a preferred algorithm for demozaicing you should make sure that info is passed along. Otherwise demozaiced DPX (at least 10 bit) log or prores4444XQ log is fine. You should sort out any color management methods with VFX, including if they want a viewer lut or not. Personally Im a big fan of using ACES when theres VFX involved. You dont want to give them graded footage, as it can limit their ability to give you the best composite. Sometimes you need to dig deep into a color channel to get a track or isolate something and if you grade it you can tie VFX's hands. You may also want to do a round trip test of the pipeline as well, so you know that what you get back from VFX is in the right/intended format for color

the LWZ.1 and LWZ.2 match the Ultra 16s. same coatings, nearly as sharp. its a 15.5-45 t2.6, so its not super wide. I used it as my main lens with my U16 8mm till I hunted down my longer U16 focal lengths. The only other lens that will directly match is the master zoom 16.5-110, but thats like bolting a canon on to the front of your camera. If you need wider, your best bet will be the zeiss s16 zoom 12-120. It has the zeiss contrast, but it wont be color matched. If you go with a wider canon zoom, you'll get some magenta chromatic aberration that doesnt match the zeiss glass and isnt really undoable in the grade, unless you've got a lot of time to work on it in post.

among other things, it just makes life easier. You can take your film scans and tag them as cineon and set an output transform (say rec 709), and you're suddenly not wasting time with as much technical work as you are now able to just dive in creatively. It also means that you can bring in other cameras, say an arri alexa, set it up as log c and with the same output, and now your footage is more or less living in the same general world color wise. This color management also makes outputting to different color spaces a little easier since your material was effectively normalized on the input side. You should still manually check your color for each space instead of relying on a final output transform to magically get it right, but it does make life easier

You usually want to set your input profile for film as cineon, but its not necessarily going to be perfect due to a variety of reasons (say you are mixing 2 week old film with 6 month old film, there will be subtle differences). I've also worked with some labs who had instructed me to use arri log c instead of cineon, so you want to talk to your lab/scan house just to be sure. As for Aces vs Resolve color management, I think it just depends on your needs, what you're mixing, how heavy a vfx show it is etc. Personally I lean toward ACES when dealing with heavy vfx where Nuke is the main compositing tool due to the linear workspace, but Im not convinced thats vital. It seems to make things easier for the VFX houses and simpler to solve color management problems though should shots come back wonky.

2 perf 2.40:1 uses less surface area than 3 perf 2.40:1. The grain will be more present. Remember that 2 perf does not have a native aspect ratio of 2.40:1, its much wider. there is an old panavision explainer pdf that I have, but its too large to post here and annoyingly Im not seeing it elsewhere online. but from that doc: 3 Perf 2.40 0.945 x 0.394 in. 24.00 x 10.04 mm [same area as 4 perf 2.40] 2 Perf 2.40 0.825 x 0.345 in. 20.96 x 8.76 mm 3P 2.40 is 1.3x larger than 2P 2.40

if this is repeatably happening you need to have your camera and lenses serviced. You will need your flange depth checked as well as your fiber screen mount checked. Your lenses will need to be collimated. Maybe someone here can recommend a tech or service house in London that can help you.