WEBVTT 00:00:00.000 --> 00:00:03.919 There's a misconception that movies shot 00:00:02.480 --> 00:00:06.640 with digital cameras [music] and those 00:00:03.919 --> 00:00:08.515 shot with film cameras work in contrast 00:00:06.639 --> 00:00:10.000 with one another. That one process 00:00:08.515 --> 00:00:12.720 [music] negates the other. Like you're a 00:00:10.000 --> 00:00:14.519 film person or a digital person. But a 00:00:12.720 --> 00:00:16.960 movie shot digitally [music] can be 00:00:14.519 --> 00:00:18.920 printed and projected on film. And a 00:00:16.960 --> 00:00:20.304 movie shot on film will very likely at 00:00:18.920 --> 00:00:21.080 some point go through a scanning process 00:00:20.304 --> 00:00:22.600 [music] 00:00:21.079 --> 00:00:25.359 for at least digital effects and 00:00:22.600 --> 00:00:27.640 editing. And with IMAX, both ends of 00:00:25.359 --> 00:00:30.719 that equation, the scanning and the 00:00:27.640 --> 00:00:33.399 recording of large format 70 mm film 00:00:30.719 --> 00:00:34.079 require purpose-built equipment and an 00:00:33.399 --> 00:00:36.679 experienced [music] 00:00:34.079 --> 00:00:38.719 technician to operate those machines. I 00:00:36.679 --> 00:00:39.960 visited the 70 mm film scanning and 00:00:38.719 --> 00:00:42.159 recording labs [music] at IMAX 00:00:39.960 --> 00:00:44.600 headquarters to watch these machines in 00:00:42.159 --> 00:00:46.504 action, including one of the most 00:00:44.600 --> 00:00:48.399 beautiful bespoke machines I've ever 00:00:46.505 --> 00:00:52.600 [music] seen, this film scanner that has 00:00:48.399 --> 00:00:55.280 scanned almost every single IMAX film. 00:00:52.600 --> 00:00:57.880 Hey, Fred. Hello. Such a pleasure. 00:00:55.280 --> 00:01:00.399 Likewise, thank you. Um okay, so this is 00:00:57.880 --> 00:01:02.840 the old scanner, and that's the new 00:01:00.399 --> 00:01:04.359 scanner. I have so many questions about 00:01:02.840 --> 00:01:06.040 this, but I want to just start with like 00:01:04.359 --> 00:01:07.519 how long have you been working for IMAX? 00:01:06.040 --> 00:01:08.520 I've been working for IMAX I'm in my 00:01:07.519 --> 00:01:10.159 15th year now. 00:01:08.519 --> 00:01:11.319 >> 15th year? Have you been scanning for 00:01:10.159 --> 00:01:13.479 them the entire time? 00:01:11.319 --> 00:01:15.799 >> Scanning and film recording. Amazing. 00:01:13.480 --> 00:01:17.400 >> So, if you have a piece of IMAX film and 00:01:15.799 --> 00:01:17.679 it needs to be a digital image, that's 00:01:17.400 --> 00:01:19.600 me. 00:01:17.680 --> 00:01:21.400 >> That's you. And if it's a digital image 00:01:19.599 --> 00:01:22.159 and you need a piece of IMAX film, 00:01:21.400 --> 00:01:24.320 that's also me. 00:01:22.159 --> 00:01:26.200 >> Amazing. Okay, so I I we're going to 00:01:24.319 --> 00:01:27.879 talk about the process 00:01:26.200 --> 00:01:31.640 uh that film goes through in your room 00:01:27.879 --> 00:01:33.679 here. But I right off the bat, I walk in 00:01:31.640 --> 00:01:36.599 and I see a machine with a marble 00:01:33.680 --> 00:01:38.600 countertop and these space-saving holes 00:01:36.599 --> 00:01:41.319 and chrome polish, and then I touch it 00:01:38.599 --> 00:01:43.199 and I'm like, wait a minute. Cuz cuz 00:01:41.319 --> 00:01:44.959 Greg said outside that this was one of 00:01:43.200 --> 00:01:45.439 This is serial number one of this 00:01:44.959 --> 00:01:47.839 machine. 00:01:45.439 --> 00:01:49.200 >> This is number one. This is the only 65 00:01:47.840 --> 00:01:51.040 mm 00:01:49.200 --> 00:01:52.560 scanner that that this company built. 00:01:51.040 --> 00:01:55.120 Well, dude, whoever made this 00:01:52.560 --> 00:01:56.920 >> 35 mm scanners afterwards, and they made 00:01:55.120 --> 00:01:58.560 like, I don't know, pile of them. But 00:01:56.920 --> 00:02:00.239 this is the only one for 65. 00:01:58.560 --> 00:02:03.480 >> Okay, but whoever made this is an 00:02:00.239 --> 00:02:05.199 important artist. Yes. Okay, you agree. 00:02:03.480 --> 00:02:06.320 This is a masterpiece. 00:02:05.200 --> 00:02:08.240 >> Yes. Yes, sir. 00:02:06.319 --> 00:02:09.840 >> Oh my gosh. And then when you were 00:02:08.240 --> 00:02:11.680 calibrating, I watched this whole thing 00:02:09.840 --> 00:02:14.680 move or there's all these servos inside 00:02:11.680 --> 00:02:16.200 here. Yep, motors and belts and you 00:02:14.680 --> 00:02:18.360 can't lift the hood because it's 00:02:16.199 --> 00:02:19.919 granite. And it's made granite so that 00:02:18.360 --> 00:02:21.560 it's super stable so that it's 00:02:19.919 --> 00:02:24.839 measurements are the same. OH MY GOD, 00:02:21.560 --> 00:02:26.159 IT'S NOT even just aesthetic. It Yes. 00:02:24.840 --> 00:02:29.120 It is 00:02:26.159 --> 00:02:31.280 temperature. Uh because a few degrees in 00:02:29.120 --> 00:02:33.159 temperature can change you know, metal 00:02:31.280 --> 00:02:34.759 will change a little bit and that'll 00:02:33.159 --> 00:02:37.079 throw this whole thing out of alignment. 00:02:34.759 --> 00:02:40.479 So they built it on granite, polished 00:02:37.080 --> 00:02:42.520 granite. Because you're scanning pixels 00:02:40.479 --> 00:02:45.399 and you can't have any variance 00:02:42.520 --> 00:02:45.840 whatsoever. We are scanning a frame this 00:02:45.400 --> 00:02:47.240 big. 00:02:45.840 --> 00:02:48.920 >> Yeah. 00:02:47.240 --> 00:02:51.439 And how many pixels is that total? 00:02:48.919 --> 00:02:52.519 >> And we are scanning it at a over 8,000 00:02:51.439 --> 00:02:53.639 pixels across. 00:02:52.520 --> 00:02:55.719 >> Crazy. 00:02:53.639 --> 00:02:57.919 The way this machine was set up is it 00:02:55.719 --> 00:02:58.879 only does a vertical line of pixels at a 00:02:57.919 --> 00:03:02.439 time. Okay. 00:02:58.879 --> 00:03:05.479 >> So it's slicing the image 8,424 00:03:02.439 --> 00:03:06.639 times per image. That must 00:03:05.479 --> 00:03:08.119 That must take a while. 00:03:06.639 --> 00:03:10.359 >> put 12 and 1/2 seconds. 00:03:08.120 --> 00:03:13.480 >> 12 and 1/2 seconds per frame? 00:03:10.360 --> 00:03:15.440 >> Per frame. So every IMAX that's ever 00:03:13.479 --> 00:03:17.959 been scanned went through this scanner 00:03:15.439 --> 00:03:20.599 until Until VERY RECENTLY. YES. 00:03:17.960 --> 00:03:22.640 >> WOW. THAT IS A 00:03:20.599 --> 00:03:25.879 That you kind of have to work backwards 00:03:22.639 --> 00:03:27.039 from deadlines and deliveries just to 00:03:25.879 --> 00:03:28.039 ACCOMMODATE THIS. 00:03:27.039 --> 00:03:29.079 >> YES. 00:03:28.039 --> 00:03:32.039 WOW. 00:03:29.080 --> 00:03:34.280 >> YEAH. Is this machine was semi-retired 00:03:32.039 --> 00:03:35.496 at this point? Is it emeritus? I'm 00:03:34.280 --> 00:03:35.960 sorry, it is very 00:03:35.497 --> 00:03:38.800 >> [laughter] 00:03:35.960 --> 00:03:40.800 >> very much semi-retired. I do test it, 00:03:38.800 --> 00:03:42.360 you know, weekly to make sure that it's 00:03:40.800 --> 00:03:42.920 still running because this is our 00:03:42.360 --> 00:03:45.800 backup. 00:03:42.919 --> 00:03:48.000 >> Right. If this goes out, this This is 00:03:45.800 --> 00:03:49.320 it. Now, the nice thing is we haven't 00:03:48.000 --> 00:03:53.360 had to use it. Right. 00:03:49.319 --> 00:03:55.319 >> This has only seen me testing it for 00:03:53.360 --> 00:03:58.080 a year now. Or a year? All right. 00:03:55.319 --> 00:03:58.919 >> So it's not actually been in production. 00:03:58.080 --> 00:04:01.719 It's such a 00:03:58.919 --> 00:04:04.319 >> is capable of being being that 25 years 00:04:01.719 --> 00:04:06.199 after it was made. For a machine. I 00:04:04.319 --> 00:04:07.959 mean, there's a way in which I can tell 00:04:06.199 --> 00:04:10.919 that the people that put this together 00:04:07.960 --> 00:04:11.320 love machines and also love products. 00:04:10.919 --> 00:04:13.319 Yes. 00:04:11.319 --> 00:04:16.399 >> Like they weren't just making a one-off, 00:04:13.319 --> 00:04:18.879 they were making it a beautiful product. 00:04:16.399 --> 00:04:21.919 >> Yes. This is inspired by I feel like 00:04:18.879 --> 00:04:24.120 German manufacturing from mid-century, 00:04:21.920 --> 00:04:27.080 from Scandinavian. I mean, it's like 00:04:24.120 --> 00:04:28.552 there's so much. Wow. Yeah. Okay, now 00:04:27.079 --> 00:04:29.439 let's talk ABOUT THIS PROCESS. 00:04:28.552 --> 00:04:31.480 >> [laughter] 00:04:29.439 --> 00:04:33.680 >> LET ME PUT A PIECE OF FILM ON. OKAY, NO 00:04:31.480 --> 00:04:36.360 MIND. SO, 00:04:33.680 --> 00:04:39.000 take a piece of film and we thread it 00:04:36.360 --> 00:04:40.720 up, starting this side. 00:04:39.000 --> 00:04:41.680 It starts Make sure it doesn't want to 00:04:40.720 --> 00:04:44.120 walk away on us. 00:04:41.680 --> 00:04:44.120 >> Fair. 00:04:46.319 --> 00:04:50.639 Make sure we take a few up, a few rolls 00:04:48.600 --> 00:04:52.280 laps up on the on the roll. 00:04:50.639 --> 00:04:54.000 And then 00:04:52.279 --> 00:04:56.239 I like to pull a little bit extra and 00:04:54.000 --> 00:04:58.920 thread it through here. So threading up 00:04:56.240 --> 00:05:01.840 the gate first so that it sets correctly 00:04:58.920 --> 00:05:04.160 and then with the sprockets, 00:05:01.839 --> 00:05:05.639 make sure it locks in place. 00:05:04.160 --> 00:05:08.040 So I just want to highlight here for a 00:05:05.639 --> 00:05:11.240 second like, you know, film is all about 00:05:08.040 --> 00:05:11.760 the fast shuttling and with precision. 00:05:11.240 --> 00:05:13.800 Yes. 00:05:11.759 --> 00:05:16.360 >> That's got to be an exponentially harder 00:05:13.800 --> 00:05:19.199 problem at this size and scale, 00:05:16.360 --> 00:05:20.800 especially for tiny variances of 00:05:19.199 --> 00:05:22.879 sprockets, which I would imagine you 00:05:20.800 --> 00:05:23.680 guys have reduced as close to zero as is 00:05:22.879 --> 00:05:24.959 humanly possible. 00:05:23.680 --> 00:05:28.120 >> Exactly. And that's why you may notice 00:05:24.959 --> 00:05:30.000 the film path is very precisely defined. 00:05:28.120 --> 00:05:32.800 I'm not making loops. I'm not making 00:05:30.000 --> 00:05:34.920 approximately the right size. I'm going 00:05:32.800 --> 00:05:36.759 it fits right here, it fits right here, 00:05:34.920 --> 00:05:38.600 and everything lines up exactly. 00:05:36.759 --> 00:05:39.920 >> So this machine knows what to expect 00:05:38.600 --> 00:05:41.439 when this piece of film is through that 00:05:39.920 --> 00:05:43.240 sprocket hole and when it's going to get 00:05:41.439 --> 00:05:44.959 over here and it's never wrong. 00:05:43.240 --> 00:05:46.560 >> And it's never wrong. 00:05:44.959 --> 00:05:46.919 So, 00:05:46.560 --> 00:05:48.199 then 00:05:46.920 --> 00:05:48.720 >> Oh, and now it goes through all these 00:05:48.199 --> 00:05:51.680 things. 00:05:48.720 --> 00:05:54.440 >> all all these things. So, 00:05:51.680 --> 00:05:56.560 there is, I believe, a an edge code 00:05:54.439 --> 00:05:59.560 because film comes with uh from Kodak 00:05:56.560 --> 00:06:00.879 comes with a an edge code printed right 00:05:59.560 --> 00:06:02.519 into the film. Oh, okay. 00:06:00.879 --> 00:06:04.639 >> It's a latent image until obviously they 00:06:02.519 --> 00:06:07.439 process the film. Unfortunately, this 00:06:04.639 --> 00:06:09.079 one does not work, never has. So, it's 00:06:07.439 --> 00:06:10.680 there. It was there with the idea of 00:06:09.079 --> 00:06:11.079 making it work, but it never did. 00:06:10.680 --> 00:06:13.240 >> Copy. 00:06:11.079 --> 00:06:15.680 >> So we go around there. These two 00:06:13.240 --> 00:06:17.480 basically are controlling the tension. 00:06:15.680 --> 00:06:17.920 How much tension do do we take up? 00:06:17.480 --> 00:06:20.160 >> I see. 00:06:17.920 --> 00:06:21.319 >> And And watch when I spin it up, see how 00:06:20.160 --> 00:06:23.240 that Mhm. 00:06:21.319 --> 00:06:25.279 >> And it'll find a balance. And as the 00:06:23.240 --> 00:06:26.639 film is transported, it'll make sure 00:06:25.279 --> 00:06:28.799 that this thing's pulling at the right 00:06:26.639 --> 00:06:29.519 amount, even if the roll's really big or 00:06:28.800 --> 00:06:31.240 really small. 00:06:29.519 --> 00:06:33.599 >> So that it's never winding it too tight 00:06:31.240 --> 00:06:35.639 or too loose, it's always at a specific 00:06:33.600 --> 00:06:39.080 >> Oh, exactly. Wow, of course it has to 00:06:35.639 --> 00:06:40.479 be. Because too loose and your film goes 00:06:39.079 --> 00:06:43.319 on the ground. 00:06:40.480 --> 00:06:44.600 Too tight and you risk snapping the film 00:06:43.319 --> 00:06:46.920 or 00:06:44.600 --> 00:06:48.600 pulling the film past the the sprockets 00:06:46.920 --> 00:06:51.240 and creating problems that way. 00:06:48.600 --> 00:06:53.640 >> I cuz if film is surprisingly brittle. 00:06:51.240 --> 00:06:56.319 Yes. I remember as a projectionist, I 00:06:53.639 --> 00:06:58.360 remember having one smacking around and 00:06:56.319 --> 00:07:00.514 creating a lot of celluloid confetti. 00:06:58.360 --> 00:07:00.960 Yes. 00:07:00.514 --> 00:07:02.800 >> [laughter] 00:07:00.959 --> 00:07:04.639 >> When we get into film recording, I'll 00:07:02.800 --> 00:07:05.240 tell you about chewed-up film. 00:07:04.639 --> 00:07:06.959 >> Okay. 00:07:05.240 --> 00:07:09.439 >> Cuz those machines like to chew film 00:07:06.959 --> 00:07:10.599 occasionally. So do you only do tension 00:07:09.439 --> 00:07:12.199 on the take-up side? 00:07:10.600 --> 00:07:14.360 >> let me do this side as well. It's the 00:07:12.199 --> 00:07:15.279 same thing just in reverse. 00:07:14.360 --> 00:07:18.560 Or this. 00:07:15.279 --> 00:07:20.679 Go around this way. This way and then 00:07:18.560 --> 00:07:22.160 pull tight. 00:07:20.680 --> 00:07:23.920 And that lets it find its level. 00:07:22.160 --> 00:07:25.200 >> gate closes when I DO THAT. 00:07:23.920 --> 00:07:26.800 >> OH, WOW. 00:07:25.199 --> 00:07:28.599 >> It automatically knows, oh, I got 00:07:26.800 --> 00:07:30.120 tension on both sides. That means I can 00:07:28.600 --> 00:07:32.680 try to close the gate. That's some 00:07:30.120 --> 00:07:35.399 really nice user experience right there. 00:07:32.680 --> 00:07:38.480 >> When the gate tries to close, if it 00:07:35.399 --> 00:07:38.919 can't pin that one Yep. and that one 00:07:38.480 --> 00:07:40.759 >> Yep. 00:07:38.920 --> 00:07:42.879 >> Those pins actually go through the film 00:07:40.759 --> 00:07:45.599 from the other side to hold it in 00:07:42.879 --> 00:07:47.040 position. Yep. If they go if if 00:07:45.600 --> 00:07:48.840 everything's not lined up correctly, 00:07:47.040 --> 00:07:50.240 they will try to pin through the film. 00:07:48.839 --> 00:07:50.479 Very gently, they'll do it and they'll 00:07:50.240 --> 00:07:51.519 stop. 00:07:50.480 --> 00:07:53.000 >> won't poke through it. 00:07:51.519 --> 00:07:54.680 >> And they'll go, nope, there's something 00:07:53.000 --> 00:07:56.480 there. Nope, you know what? Forget it, 00:07:54.680 --> 00:07:58.199 I'm not doing this. So it's not going to 00:07:56.480 --> 00:07:58.600 punch a hole in your film. That's 00:07:58.199 --> 00:08:00.519 beautiful. 00:07:58.600 --> 00:08:03.400 >> 25 years ago, 00:08:00.519 --> 00:08:05.120 somebody made this brilliant decision 00:08:03.399 --> 00:08:06.560 that has saved so many rolls of film. 00:08:05.120 --> 00:08:08.639 >> If any of the people that worked on this 00:08:06.560 --> 00:08:10.319 is watching this video, reach out to us 00:08:08.639 --> 00:08:12.000 in the comments cuz I want to talk to 00:08:10.319 --> 00:08:14.519 you about this thing. Do you even This 00:08:12.000 --> 00:08:15.240 this beautiful. These these circles and 00:08:14.519 --> 00:08:17.639 lights here. 00:08:15.240 --> 00:08:19.680 >> It's almost a face. It Oh, there there 00:08:17.639 --> 00:08:20.639 is. It's almost a little like a droid. 00:08:19.680 --> 00:08:22.879 Yeah. 00:08:20.639 --> 00:08:24.399 Okay, so now the film is loaded up. So 00:08:22.879 --> 00:08:26.639 once it's loaded up, I try to make sure 00:08:24.399 --> 00:08:28.039 I I will advance a frame, a couple 00:08:26.639 --> 00:08:31.439 frames just to make sure that it is 00:08:28.040 --> 00:08:32.038 actually going to transport correctly. 00:08:31.439 --> 00:08:33.080 Like that. 00:08:32.038 --> 00:08:35.158 >> So that was one frame. 00:08:33.080 --> 00:08:37.719 >> Those This is one frame. 00:08:35.158 --> 00:08:40.599 Gotcha. That's moving 15 perfs. 00:08:37.719 --> 00:08:42.680 >> It's moving 15 perfs at a go. Okay. And 00:08:40.599 --> 00:08:44.158 as you can see, very smoothly. Yep. Yep, 00:08:42.679 --> 00:08:45.439 through custom etched and machine 00:08:44.158 --> 00:08:47.319 buttons that didn't get bought in a 00:08:45.440 --> 00:08:47.936 catalog. Right? 00:08:47.320 --> 00:08:48.640 Yes. 00:08:47.936 --> 00:08:50.320 >> [laughter] 00:08:48.639 --> 00:08:52.080 >> Then I'll wind down to the beginning of 00:08:50.320 --> 00:08:53.960 the roll because this is leader for 00:08:52.080 --> 00:08:57.000 handling purposes. That way I don't have 00:08:53.960 --> 00:08:58.400 to handle people's film. 00:08:57.000 --> 00:08:59.480 And if we look through here, 00:08:58.399 --> 00:09:01.759 >> Oh, yeah, see the color test. 00:08:59.480 --> 00:09:05.639 >> see the We can see the picture and I can 00:09:01.759 --> 00:09:05.639 actually make adjustments to 00:09:06.440 --> 00:09:12.280 Oh, I can go one perf at a time so that 00:09:09.360 --> 00:09:13.120 I'm in frame. Gotcha. So that's the one 00:09:12.279 --> 00:09:16.199 perf jog. 00:09:13.120 --> 00:09:18.080 >> Yes. Amazing. Which is stop plus 00:09:16.200 --> 00:09:19.440 forward. I can also, of course, do stop 00:09:18.080 --> 00:09:21.480 plus backward to go the opposite 00:09:19.440 --> 00:09:23.320 direction if I need to. 00:09:21.480 --> 00:09:25.320 And we are almost ready to scan a frame, 00:09:23.320 --> 00:09:26.879 but before I do, make sure I take the 00:09:25.320 --> 00:09:27.680 calibration strip. 00:09:26.879 --> 00:09:29.360 Oh. 00:09:27.679 --> 00:09:32.039 >> scan like that, but it will not look 00:09:29.360 --> 00:09:33.680 very good. So What does the calibration 00:09:32.039 --> 00:09:35.279 strip do? It's a color calibration? 00:09:33.679 --> 00:09:37.839 >> It's It's a color calibration and it 00:09:35.279 --> 00:09:39.639 basically emulates the unexposed film 00:09:37.840 --> 00:09:41.600 that is processed. 00:09:39.639 --> 00:09:43.840 We call it the film base. Gotcha. 00:09:41.600 --> 00:09:46.040 >> So it is 00:09:43.840 --> 00:09:48.519 just a piece of clear Right, that is 00:09:46.039 --> 00:09:50.240 sort of the neutral of film. Yes. And 00:09:48.519 --> 00:09:52.125 yes, it's a little scuffed up. It's been 00:09:50.240 --> 00:09:52.600 used many years. 00:09:52.125 --> 00:09:55.639 >> [laughter] 00:09:52.600 --> 00:09:57.560 >> The fact that it's here and not here 00:09:55.639 --> 00:09:59.039 throws it so far out of focus, you don't 00:09:57.559 --> 00:10:00.959 the blemishes or what have you. 00:09:59.039 --> 00:10:02.799 >> the blemishes don't matter. Speaking of 00:10:00.960 --> 00:10:04.040 which, one of the things that we do in 00:10:02.799 --> 00:10:06.120 this room is we have to control the 00:10:04.039 --> 00:10:07.919 temperature and humidity very tightly. I 00:10:06.120 --> 00:10:10.399 would imagine the humidity could affect 00:10:07.919 --> 00:10:12.679 image quality, right? Humidity will 00:10:10.399 --> 00:10:14.720 cause the film to warp. 00:10:12.679 --> 00:10:16.719 The IMAX frame is so big that it 00:10:14.720 --> 00:10:19.160 actually can warp as it goes into the 00:10:16.720 --> 00:10:20.600 gate and still feed through the machine. 00:10:19.159 --> 00:10:22.639 Is it because the celluloid is slightly 00:10:20.600 --> 00:10:23.159 hygroscopic, it absorbs a little water? 00:10:22.639 --> 00:10:23.879 EXACTLY. 00:10:23.159 --> 00:10:25.360 >> WOW. 00:10:23.879 --> 00:10:26.919 >> EXACTLY. 00:10:25.360 --> 00:10:29.000 And we've had 00:10:26.919 --> 00:10:30.559 we've experienced it where it's like we 00:10:29.000 --> 00:10:33.000 scan something then when we scanned it a 00:10:30.559 --> 00:10:35.719 couple days later it's like, wait, why 00:10:33.000 --> 00:10:36.519 is it a different focus? Wow, that's 00:10:35.720 --> 00:10:38.960 amazing. 00:10:36.519 --> 00:10:40.600 >> Let's We really jumped on it, made sure 00:10:38.960 --> 00:10:42.800 that everything was 00:10:40.600 --> 00:10:44.159 properly set up and So then temperature 00:10:42.799 --> 00:10:46.479 and humidity control is part of the 00:10:44.159 --> 00:10:48.279 calibration process of this room. Yes, 00:10:46.480 --> 00:10:49.639 and that's why you hear air running 00:10:48.279 --> 00:10:51.039 continuously 00:10:49.639 --> 00:10:53.840 >> in this room. So you got a lot of 00:10:51.039 --> 00:10:55.661 hoodies and fleeces. You make sure 00:10:53.840 --> 00:10:57.080 you're buttoned down warm and you 00:10:55.662 --> 00:10:58.759 [laughter] take care of the all of that 00:10:57.080 --> 00:11:01.040 you make sure that that's good. Um and 00:10:58.759 --> 00:11:02.720 this is the business end of the scanner 00:11:01.039 --> 00:11:04.519 coming through this way. Uh so the light 00:11:02.720 --> 00:11:06.680 column is here, the light is underneath. 00:11:04.519 --> 00:11:08.799 Gotcha. Goes up and I think they use a 00:11:06.679 --> 00:11:12.039 prism or a mirror and deflect it into 00:11:08.799 --> 00:11:14.120 the camera which is here. Gotcha. Um And 00:11:12.039 --> 00:11:16.360 it is a linear array so it's not taking 00:11:14.120 --> 00:11:17.839 the entire picture at the same time. 00:11:16.360 --> 00:11:20.552 It's taking 00:11:17.839 --> 00:11:22.160 one set of pixels vertically up to 6,144 00:11:20.552 --> 00:11:24.120 [laughter] 00:11:22.159 --> 00:11:27.679 pixels vertically. So the scanner is 00:11:24.120 --> 00:11:29.919 just basically one pixel wide and film 00:11:27.679 --> 00:11:33.120 frame tall. Uh ish 00:11:29.919 --> 00:11:34.639 >> Yes. ish Exactly, yeah. 00:11:33.120 --> 00:11:35.759 12 and 1/2 seconds. 00:11:34.639 --> 00:11:37.838 >> 12 and 1/2 seconds. 00:11:35.759 --> 00:11:40.759 >> I I'll be frank, knowing what it's 00:11:37.839 --> 00:11:42.360 trying to do that sounds fast. 00:11:40.759 --> 00:11:43.720 But it also sounds slow. 00:11:42.360 --> 00:11:45.200 >> Incredibly slow. [laughter] When we look 00:11:43.720 --> 00:11:47.759 at the other machine you'll be like, 00:11:45.200 --> 00:11:49.759 that is very very slow. Um can we just 00:11:47.759 --> 00:11:52.000 able to actually just witness a scanning 00:11:49.759 --> 00:11:53.480 sequence? Um yes, let give me a moment 00:11:52.000 --> 00:11:56.720 to set that up. 00:11:53.480 --> 00:11:58.639 Oh, one other thing you you hear the 00:11:56.720 --> 00:12:00.639 hear the hiss? Yeah. That's actually we 00:11:58.639 --> 00:12:03.360 actually have air knives on it. 00:12:00.639 --> 00:12:05.759 Partly to blow off the dust, partly to 00:12:03.360 --> 00:12:07.800 keep the film, you know, moving in it 00:12:05.759 --> 00:12:09.120 smoothly. Right, right. Where are the 00:12:07.799 --> 00:12:11.719 air knives? Are they here? 00:12:09.120 --> 00:12:11.919 >> They are actually in this. Right, right. 00:12:11.720 --> 00:12:13.960 Okay. 00:12:11.919 --> 00:12:15.519 >> So with if there was no film in the gate 00:12:13.960 --> 00:12:16.519 I could show you you could stick your 00:12:15.519 --> 00:12:18.600 put your hand in the gate and see 00:12:16.519 --> 00:12:21.360 actually feel them. Gotcha. Uh but the 00:12:18.600 --> 00:12:22.560 whole goal is when it's here no dust has 00:12:21.360 --> 00:12:24.480 been able to get anywhere near it. 00:12:22.559 --> 00:12:26.799 Exactly. Uh I really appreciate you 00:12:24.480 --> 00:12:28.879 walking me through this. This is it's 00:12:26.799 --> 00:12:30.759 it's amazing. Let's talk about this 00:12:28.879 --> 00:12:31.838 brand new one. Yes. 00:12:30.759 --> 00:12:33.679 So what's the biggest difference between 00:12:31.839 --> 00:12:36.440 the new machine and the old machine? Is 00:12:33.679 --> 00:12:38.479 it just speed? Part of it speed, part of 00:12:36.440 --> 00:12:41.320 it is the process that it takes the the 00:12:38.480 --> 00:12:43.560 image. Mhm. Um the old machine took it a 00:12:41.320 --> 00:12:45.879 single vertical line of pixels at a 00:12:43.559 --> 00:12:48.599 time. This actually will scan an entire 00:12:45.879 --> 00:12:51.080 frame except of course a gigantic frame 00:12:48.600 --> 00:12:53.720 like an IMAX frame because an IMAX frame 00:12:51.080 --> 00:12:56.639 is so big. So it actually scans the IMAX 00:12:53.720 --> 00:12:59.200 frame in thirds. Oh, so it does 1/3, 00:12:56.639 --> 00:13:01.120 middle third, other third. Yes. And it 00:12:59.200 --> 00:13:02.240 does its own stitching of them? Um no, 00:13:01.120 --> 00:13:03.360 we actually do that on a separate 00:13:02.240 --> 00:13:04.639 machine. Oh, amazing. 00:13:03.360 --> 00:13:06.440 >> We could do it on the separate on the 00:13:04.639 --> 00:13:06.960 same machine but it is a separate 00:13:06.440 --> 00:13:07.960 process. 00:13:06.960 --> 00:13:09.519 >> Gotcha. 00:13:07.960 --> 00:13:11.000 That allows us to do a little bit of 00:13:09.519 --> 00:13:12.360 extra work if we need to do some 00:13:11.000 --> 00:13:15.039 processing or something like that in 00:13:12.360 --> 00:13:17.800 between. Okay. And this one you may 00:13:15.039 --> 00:13:19.879 notice threads up the opposite direction 00:13:17.799 --> 00:13:20.519 from the the other machine. Goes right 00:13:19.879 --> 00:13:22.279 to left. 00:13:20.519 --> 00:13:24.720 >> It goes right to left. 00:13:22.279 --> 00:13:26.000 Apparently if we went left to right this 00:13:24.720 --> 00:13:27.680 whole apparatus would be have to be 00:13:26.000 --> 00:13:31.000 upside down and this whole big thing 00:13:27.679 --> 00:13:32.479 would have to stick up in the air. So 00:13:31.000 --> 00:13:34.039 >> Yes. That's so funny. 00:13:32.480 --> 00:13:35.560 >> Because I was I was freaked out that 00:13:34.039 --> 00:13:36.679 it's horrible when you have to scan on 00:13:35.559 --> 00:13:38.359 both machines at the same time. 00:13:36.679 --> 00:13:38.719 >> Oh cuz you're going back and 00:13:38.360 --> 00:13:40.279 uh yeah. 00:13:38.720 --> 00:13:42.160 >> and forth and you're like, which one is 00:13:40.279 --> 00:13:44.480 wound the right way? Sometimes you just 00:13:42.159 --> 00:13:46.159 have to stop and just yeah, I can 00:13:44.480 --> 00:13:48.360 imagine. This is where being on the 00:13:46.159 --> 00:13:50.120 computer helps because on the computer 00:13:48.360 --> 00:13:51.680 forward is always forward, backward is 00:13:50.120 --> 00:13:52.000 always backward. [laughter] 00:13:51.679 --> 00:13:54.159 Okay. 00:13:52.000 --> 00:13:56.200 >> So the setup is actually similar here as 00:13:54.159 --> 00:13:58.600 you do a rough setup and then you do the 00:13:56.200 --> 00:14:01.920 spool winding. Exactly, yes. Then fill 00:13:58.600 --> 00:14:01.920 in the details as it were. 00:14:03.279 --> 00:14:06.679 That's 00:14:05.360 --> 00:14:09.039 There we go. And this one also has 00:14:06.679 --> 00:14:11.039 tension arms. Gotcha. 00:14:09.039 --> 00:14:12.679 Some things don't change. You always 00:14:11.039 --> 00:14:15.838 need tension arms to make sure that the 00:14:12.679 --> 00:14:15.838 motors on the 00:14:16.000 --> 00:14:20.399 on the rolls of film aren't taking up 00:14:18.799 --> 00:14:21.719 too tightly or too loosely. And your 00:14:20.399 --> 00:14:23.958 goal is that each tension arm is 00:14:21.720 --> 00:14:26.279 vertical. Roughly. It they au- 00:14:23.958 --> 00:14:28.399 automatically will you know, they'll 00:14:26.279 --> 00:14:29.799 stop if at a certain point. I go mostly 00:14:28.399 --> 00:14:31.559 go approximately the right place, 00:14:29.799 --> 00:14:34.559 sometimes a little past and then they 00:14:31.559 --> 00:14:36.439 stop cuz that's where they need to be. 00:14:34.559 --> 00:14:38.519 It's called reader. We can read the key 00:14:36.440 --> 00:14:39.560 codes that are that Kodak puts on the 00:14:38.519 --> 00:14:41.439 film 00:14:39.559 --> 00:14:43.719 and that way the editorial people can 00:14:41.440 --> 00:14:45.000 tell us scan these exact set of frames 00:14:43.720 --> 00:14:46.839 and we can say 00:14:45.000 --> 00:14:47.639 okay, wherever it is on the roll find 00:14:46.839 --> 00:14:49.320 it, scan this. 00:14:47.639 --> 00:14:49.879 >> Well that makes that process much 00:14:49.320 --> 00:14:52.640 faster. 00:14:49.879 --> 00:14:53.039 >> makes it much faster and very accurate. 00:14:52.639 --> 00:14:54.319 Okay. 00:14:53.039 --> 00:14:56.120 >> So instead of having to take those 00:14:54.320 --> 00:14:57.240 numbers and go, okay, what's the number 00:14:56.120 --> 00:14:59.000 at the beginning of the roll, do these 00:14:57.240 --> 00:14:59.240 calculations, how far down do I have to 00:14:59.000 --> 00:15:01.240 go? 00:14:59.240 --> 00:15:02.959 >> And then go ahead and pick. Exactly, 00:15:01.240 --> 00:15:05.279 just find it. 00:15:02.958 --> 00:15:07.439 So that's kind of what we have for the 00:15:05.279 --> 00:15:09.519 setup on that. And again notice it is a 00:15:07.440 --> 00:15:11.800 vertical operation rather than a 00:15:09.519 --> 00:15:13.319 horizontal one. Uh hence the glass doors 00:15:11.799 --> 00:15:15.958 and the dust mitigation. But it's 00:15:13.320 --> 00:15:20.320 effectively the same machine just more 00:15:15.958 --> 00:15:21.719 than an order of magnitude faster. Yes. 00:15:20.320 --> 00:15:24.520 Dude, I 00:15:21.720 --> 00:15:26.079 already just in the last 10 minutes here 00:15:24.519 --> 00:15:28.559 my mind has been blown like three or 00:15:26.078 --> 00:15:31.399 four times and it's all about the scale. 00:15:28.559 --> 00:15:33.039 The scale of this frame of film is is is 00:15:31.399 --> 00:15:35.039 diabolical in terms of all your 00:15:33.039 --> 00:15:37.480 processes with it. Yeah. This is 00:15:35.039 --> 00:15:39.919 scanning. You guys also do a lot of 00:15:37.480 --> 00:15:41.320 recording. What's the other part of this 00:15:39.919 --> 00:15:43.958 process? Can we walk through that? 00:15:41.320 --> 00:15:44.640 >> you don't have film and you need film. 00:15:43.958 --> 00:15:45.799 Let's do that. 00:15:44.639 --> 00:15:48.480 >> Okay, good. 00:15:45.799 --> 00:15:50.078 All right. This is film recording. Yes. 00:15:48.480 --> 00:15:52.759 IMAX has been around and has been 00:15:50.078 --> 00:15:54.919 scanning for a long time. So this is 00:15:52.759 --> 00:15:57.000 real legacy equipment back to the early 00:15:54.919 --> 00:15:58.120 days. It is. If you remember the Silicon 00:15:57.000 --> 00:16:01.759 Graphics machines 00:15:58.120 --> 00:16:02.600 >> Yes, very well. Yeah, we used those to 00:16:01.759 --> 00:16:04.600 run these machines. 00:16:02.600 --> 00:16:06.159 >> Oh, hilarious. Good god, I remember one 00:16:04.600 --> 00:16:08.360 year at the ILM Christmas party that was 00:16:06.159 --> 00:16:11.078 actually the door prize for best costume 00:16:08.360 --> 00:16:13.240 was an O2 workstation. Nice. Okay, so 00:16:11.078 --> 00:16:15.719 what is the technology that is used to 00:16:13.240 --> 00:16:16.519 get the image the digital image printed 00:16:15.720 --> 00:16:18.360 onto the film? 00:16:16.519 --> 00:16:22.199 >> So to get the digital image printed on 00:16:18.360 --> 00:16:24.240 the film in the base here we have 00:16:22.200 --> 00:16:26.320 obviously boards and stuff that run that 00:16:24.240 --> 00:16:27.600 run this. But the big thing is the the 00:16:26.320 --> 00:16:29.600 tube. The tube? 00:16:27.600 --> 00:16:30.079 >> the We literally have a cathode ray 00:16:29.600 --> 00:16:33.000 tube. 00:16:30.078 --> 00:16:35.439 >> What? Like the old tube tube TVs but way 00:16:33.000 --> 00:16:37.159 more advanced obviously. There's a CRT 00:16:35.440 --> 00:16:39.959 at the heart of the scan of this 00:16:37.159 --> 00:16:41.759 printer? There is. So obviously the CRT 00:16:39.958 --> 00:16:43.559 machine gives you the precision that 00:16:41.759 --> 00:16:46.679 you're looking for on a pixel by pixel 00:16:43.559 --> 00:16:47.119 color fidelity by color fidelity basis. 00:16:46.679 --> 00:16:48.199 Yes. 00:16:47.120 --> 00:16:50.440 >> That's astounding to me. 00:16:48.200 --> 00:16:52.440 >> Exactly, yeah. Um does it do it in a 00:16:50.440 --> 00:16:54.720 mono color and then you use filters for 00:16:52.440 --> 00:16:57.560 the color? We do indeed, yes. So the 00:16:54.720 --> 00:16:59.240 tube itself is is a particular white and 00:16:57.559 --> 00:17:01.519 then we run it through some filters. So 00:16:59.240 --> 00:17:03.360 we have a We don't use the clear filter 00:17:01.519 --> 00:17:04.838 but we could if we needed to. There's a 00:17:03.360 --> 00:17:06.799 red, green and blue filter. I'm going 00:17:04.838 --> 00:17:08.958 backwards here. That's the blue filter, 00:17:06.799 --> 00:17:11.599 the green filter. So it's like the old 00:17:08.959 --> 00:17:14.560 the old three strip process reversed in 00:17:11.599 --> 00:17:17.279 order to obtain the imagery. That is so 00:17:14.559 --> 00:17:18.119 cool that the CRT provides the precision 00:17:17.279 --> 00:17:19.720 that you need. 00:17:18.119 --> 00:17:20.958 >> Yes. 00:17:19.720 --> 00:17:23.480 So 00:17:20.959 --> 00:17:25.279 this is the tube. 00:17:23.480 --> 00:17:27.838 And as you can see it is rather big. 00:17:25.279 --> 00:17:30.440 God. The piece the piece here at the end 00:17:27.838 --> 00:17:33.759 >> Yeah. is what they call the thick glass 00:17:30.440 --> 00:17:35.559 because if dust settles on the tube 00:17:33.759 --> 00:17:38.879 you have a problem. It sits there the 00:17:35.559 --> 00:17:40.240 entire time and you have a speck 00:17:38.880 --> 00:17:42.320 on your movie. 00:17:40.240 --> 00:17:43.076 And if the camera pans, you know, the 00:17:42.319 --> 00:17:43.439 the 00:17:43.076 --> 00:17:45.279 >> [clears throat] 00:17:43.440 --> 00:17:47.960 >> the speck just sits there and you can 00:17:45.279 --> 00:17:50.000 totally see it. So they the way to do to 00:17:47.960 --> 00:17:51.880 avoid that was that they came up with is 00:17:50.000 --> 00:17:53.640 to put a big really thick piece of 00:17:51.880 --> 00:17:56.679 glass. More than half the weight sits 00:17:53.640 --> 00:17:59.880 just in this glass. Wow. And it puts the 00:17:56.679 --> 00:18:02.080 the piece of dust if there is one way 00:17:59.880 --> 00:18:03.880 outside the focal plane. 00:18:02.079 --> 00:18:05.919 Oh, that's brilliant. That's so 00:18:03.880 --> 00:18:08.320 brilliant. Yeah. 00:18:05.920 --> 00:18:09.800 So you guys are are printing on the same 00:18:08.319 --> 00:18:12.919 kind of reels like two, three minutes at 00:18:09.799 --> 00:18:14.839 a time? Yes. What is the how slow is how 00:18:12.920 --> 00:18:15.720 good I was not going to say how fast is 00:18:14.839 --> 00:18:17.559 this. How slow is this? 00:18:15.720 --> 00:18:20.880 >> So once upon a time we could get a lot 00:18:17.559 --> 00:18:22.919 better speeds but with the CRTs as they 00:18:20.880 --> 00:18:24.800 age they get dimmer and we have to 00:18:22.920 --> 00:18:26.160 expose the image longer in order to get 00:18:24.799 --> 00:18:28.519 the same exposure. 00:18:26.160 --> 00:18:30.919 >> Am I correct that all the CRTs that have 00:18:28.519 --> 00:18:33.000 ever been built for this are already 00:18:30.919 --> 00:18:34.560 built and you're just grabbing new old 00:18:33.000 --> 00:18:35.480 stock where it shows up? 00:18:34.559 --> 00:18:37.678 >> And seeing if we can get them 00:18:35.480 --> 00:18:38.559 refurbished which is a job 00:18:37.679 --> 00:18:40.080 unto itself. 00:18:38.559 --> 00:18:41.678 >> Oh my, can't even imagine. These are 00:18:40.079 --> 00:18:43.439 technologies that are rapidly their 00:18:41.679 --> 00:18:45.400 support structure is disappearing all 00:18:43.440 --> 00:18:46.559 the time. It is, yeah. So they get 00:18:45.400 --> 00:18:48.800 slower as they age. 00:18:46.559 --> 00:18:50.639 >> Yeah. Currently we're running somewhere 00:18:48.799 --> 00:18:52.879 between a minute and two and a half 00:18:50.640 --> 00:18:55.159 minutes per frame. Oh my gosh. You 00:18:52.880 --> 00:18:57.240 thought 12 and 1/2 seconds was slow? 00:18:55.159 --> 00:18:59.320 A minute to a minute and a half. Um and 00:18:57.240 --> 00:19:01.240 this is so like Christopher Nolan shoots 00:18:59.319 --> 00:19:03.639 on film, he brings a couple of different 00:19:01.240 --> 00:19:05.559 shots that need a an effect done to them 00:19:03.640 --> 00:19:08.840 that is executed on these. 00:19:05.559 --> 00:19:11.799 >> Exactly. Incredible. My joke is 00:19:08.839 --> 00:19:13.959 I got tired of watching paint dry. 00:19:11.799 --> 00:19:15.559 Grass grew too quickly [laughter] for me 00:19:13.960 --> 00:19:20.079 so I went into film recording. 00:19:15.559 --> 00:19:22.839 Hilarious. How many backup CRT tubes do 00:19:20.079 --> 00:19:25.258 you have in stock currently? 00:19:22.839 --> 00:19:27.278 How about the other question? 00:19:25.258 --> 00:19:27.278 >> [laughter] 00:19:27.319 --> 00:19:31.119 >> I have like seven machines in the room 00:19:29.519 --> 00:19:32.000 and about four of them are working. Oh 00:19:31.119 --> 00:19:34.359 my gosh. 00:19:32.000 --> 00:19:35.799 >> Because of the CRTs. Um so I I would 00:19:34.359 --> 00:19:38.079 imagine there are people working on 00:19:35.799 --> 00:19:39.879 solving this problem currently. Yes. 00:19:38.079 --> 00:19:41.559 >> Okay. There are various solutions by a 00:19:39.880 --> 00:19:43.640 couple different groups trying to find 00:19:41.559 --> 00:19:45.799 basically find replacements for the CRT. 00:19:43.640 --> 00:19:47.520 Because this still exceeds the precision 00:19:45.799 --> 00:19:49.440 of anything currently being built. And 00:19:47.519 --> 00:19:51.799 reliability. You want something that's 00:19:49.440 --> 00:19:54.000 not only precise but is reliable that 00:19:51.799 --> 00:19:56.200 you can put a thousand feet of film, do 00:19:54.000 --> 00:19:57.359 it a frame at a time, and 4,000 frames 00:19:56.200 --> 00:19:58.960 later 00:19:57.359 --> 00:20:01.000 get it processed, look at the print, and 00:19:58.960 --> 00:20:03.600 go, "Yep, that looks good." Amazing. 00:20:01.000 --> 00:20:05.200 >> Every time. Um I notice it's freezing 00:20:03.599 --> 00:20:06.879 cold in this room. You're way colder 00:20:05.200 --> 00:20:07.160 even than in the other in the scanning 00:20:06.880 --> 00:20:08.760 room. 00:20:07.160 --> 00:20:10.240 >> That's a little bit accidental. It's 00:20:08.759 --> 00:20:12.319 supposed to be the same temperature and 00:20:10.240 --> 00:20:13.880 humidity as the other [laughter] room. 00:20:12.319 --> 00:20:15.960 Um because we used at one point we used 00:20:13.880 --> 00:20:18.560 to have 18 of these machines. If you can 00:20:15.960 --> 00:20:20.640 imagine 18 of them packed into this room 00:20:18.559 --> 00:20:22.879 You needed to pump a lot of cool into 00:20:20.640 --> 00:20:24.320 here to keep them all running optimally. 00:20:22.880 --> 00:20:26.960 Exactly. Wow. 00:20:24.319 --> 00:20:28.960 >> So, we've made some adjustments to the 00:20:26.960 --> 00:20:31.440 temperature to accommodate for fewer 00:20:28.960 --> 00:20:33.400 machines, but it's still cold. So, in 00:20:31.440 --> 00:20:35.279 this day and age when some filmmakers 00:20:33.400 --> 00:20:37.640 can leave critical decisions about a 00:20:35.279 --> 00:20:39.839 shot up until days before the film's 00:20:37.640 --> 00:20:41.640 release, they must have a rude awakening 00:20:39.839 --> 00:20:43.599 when they come to you guys for post. I'm 00:20:41.640 --> 00:20:45.840 sure they do. [laughter] I'm hoping that 00:20:43.599 --> 00:20:47.799 they go, "Oh, okay, let's figure this 00:20:45.839 --> 00:20:49.279 out. Let's make sure that this works." 00:20:47.799 --> 00:20:52.759 Well, I mean because the deadlines are 00:20:49.279 --> 00:20:54.799 real deadlines. Um this has been such an 00:20:52.759 --> 00:20:55.559 eye-opener. Uh thank you so much, man. 00:20:54.799 --> 00:20:58.480 This is 00:20:55.559 --> 00:20:58.480 incredible.