WEBVTT - How Camera Stabilization Works 0:00:04.160 --> 0:00:07.520 Get in text with technology with tech Stuff from stuff 0:00:07.520 --> 0:00:13.760 works dot com and welcome to tech Stuff. I am 0:00:13.840 --> 0:00:17.360 your host, Jonathan Strickland. I'm an executive producer at how 0:00:17.440 --> 0:00:22.200 Stuff Works. I like you and I like technology, and 0:00:22.320 --> 0:00:24.200 I don't really know that much about you. I just 0:00:24.239 --> 0:00:26.160 get a good feeling, but I know a pretty good 0:00:26.160 --> 0:00:30.040 amount about technology. Today, I want to talk about, you know, 0:00:30.760 --> 0:00:33.640 tech and as I'm sure most of you guys know 0:00:33.680 --> 0:00:35.720 by now, some of you know because you're doing it 0:00:35.800 --> 0:00:38.640 at this moment. I live stream tech Stuff recordings on 0:00:38.680 --> 0:00:42.440 Wednesdays and Fridays over at twitch dot tv slash tech Stuff. 0:00:42.840 --> 0:00:44.720 There's a chat room over there, and I like to 0:00:44.760 --> 0:00:47.640 spend some time chatting with viewers and listeners before the 0:00:47.680 --> 0:00:51.000 show and during breaks. And during one of those breaks 0:00:51.000 --> 0:00:54.080 in a recent show, one listener put forth a request 0:00:54.120 --> 0:00:56.160 for a future episode of tech Stuff and the request 0:00:56.240 --> 0:01:00.320 was for camera stabilization or image stabilization. So that's what 0:01:00.320 --> 0:01:04.920 I'm gonna talk about the future is now. So I'm 0:01:04.920 --> 0:01:07.759 gonna talk about image stabilization in general. And there are 0:01:07.760 --> 0:01:11.880 two really big, really general ways to go about this. 0:01:12.720 --> 0:01:16.120 One is to use mechanical elements to help compensate for 0:01:16.160 --> 0:01:19.600 the little or sometimes not so little, jittery movements we 0:01:19.680 --> 0:01:22.440 make as we try to capture video or still images 0:01:22.520 --> 0:01:27.120 using handheld cameras. And the other is the software approach, 0:01:27.240 --> 0:01:31.240 in which algorithms and programs attempt to reduce the jitters 0:01:31.240 --> 0:01:34.560 sometimes reverse the jitters using some sort of trickery on 0:01:34.600 --> 0:01:36.759 the back end. We're gonna look at both of those 0:01:36.800 --> 0:01:40.280 different approaches in this episode. Now, before I do that, 0:01:40.600 --> 0:01:42.280 I think it might be a good idea to talk 0:01:42.280 --> 0:01:45.240 about how digital video cameras work in the first place, 0:01:45.319 --> 0:01:48.640 so that you understand what is going on before we 0:01:48.720 --> 0:01:52.680 even get into stabilization. That way, we can better understand 0:01:53.280 --> 0:01:56.560 the actual mechanics of stabilization when we get there. So 0:01:56.920 --> 0:02:01.200 let's crack open a digital camcorder. Let's say, you guys 0:02:01.240 --> 0:02:04.840 remember cam quarters, right, I imagine there's still a thing. 0:02:05.000 --> 0:02:06.400 I don't know. I haven't been in the market for 0:02:06.440 --> 0:02:09.280 one for a really long time, But these days there 0:02:09.320 --> 0:02:11.280 are lots of other devices out there that can capture 0:02:11.320 --> 0:02:15.040 high resolution video, including several smartphones. So cam quarters a 0:02:15.160 --> 0:02:17.600 term I don't really hear very much anymore. Anyway, that 0:02:17.639 --> 0:02:20.480 doesn't matter, as the cam quarters inwards are pretty much 0:02:20.480 --> 0:02:22.320 the same as you would find in any digital camera 0:02:22.360 --> 0:02:25.040 device these days. Only now we've gotten really good at 0:02:25.080 --> 0:02:27.640 maturizing those components, so you don't have to lug a 0:02:27.720 --> 0:02:30.360 big thing on your shoulder, nor do you need something 0:02:30.360 --> 0:02:33.440 that is part camera part VCR. And if you don't 0:02:33.440 --> 0:02:35.360 know what a VCR is, you'll have to listen to 0:02:35.400 --> 0:02:37.120 an older version of tech stuff. I'm not going to 0:02:37.200 --> 0:02:42.080 go into that here. First, cameras record visual images. I 0:02:42.160 --> 0:02:45.720 know that breaks it down to an incredibly simplistic and 0:02:45.800 --> 0:02:48.400 almost silly statement, but it is something to keep in 0:02:48.440 --> 0:02:52.040 mind that we are talking about visual information. Now there's 0:02:52.040 --> 0:02:56.440 also audio information. Cameras mostly incorporate microphones these days. If 0:02:56.440 --> 0:02:58.400 you have a video camera that doesn't have a microphone 0:02:58.720 --> 0:03:01.320 and you aren't using external microphone, you might be an 0:03:01.360 --> 0:03:05.160 avant garde performer, or you may just have a really 0:03:05.240 --> 0:03:09.919 crappy camera. Breaking this down even further, cameras capture light. 0:03:10.040 --> 0:03:13.880 Because everything we see is from light that's bouncing off 0:03:13.919 --> 0:03:18.440 of surfaces or filtering through materials. Cameras attempt to capture 0:03:18.480 --> 0:03:22.120 that combination of lighting effects in an effort to replicate 0:03:22.160 --> 0:03:25.720 it in some way, typically to replicate in a way 0:03:25.760 --> 0:03:28.880 that is as faithful as possible to the to what 0:03:28.919 --> 0:03:31.560 we would see with our naked eyes, although obviously you 0:03:31.600 --> 0:03:35.000 can set cameras so that you're capturing stuff and altering 0:03:35.000 --> 0:03:37.600 it so it doesn't look the way it would naturally. 0:03:37.760 --> 0:03:40.320 Thus you have all the different types of filters out there. 0:03:40.760 --> 0:03:44.240 You Instagram fanatics out there know all about that, taking 0:03:44.520 --> 0:03:50.480 photos with various filters, But generally speaking, cameras are capturing light. 0:03:51.560 --> 0:03:54.480 Old cameras did this by focusing that light on photo 0:03:54.600 --> 0:03:57.880 sensitive film. So you would take light, you would direct 0:03:57.920 --> 0:04:01.360 it to this film, and that would alter the film chemically. 0:04:01.680 --> 0:04:04.080 You would then develop the film, putting it through other 0:04:04.200 --> 0:04:08.040 chemicals that would give you a negative, a reverse of 0:04:08.440 --> 0:04:10.880 the image that you want, at least in color, and 0:04:10.920 --> 0:04:13.920 then you would use that negative to produce prints, and 0:04:14.000 --> 0:04:17.240 that would replicate what your eyes saw based upon what 0:04:17.360 --> 0:04:20.680 the camera was capable of capturing. Assuming you had a 0:04:20.720 --> 0:04:23.720 good camera, it would look pretty much the way you 0:04:23.760 --> 0:04:27.440 saw it in person. Now, cameras typically use one or 0:04:27.480 --> 0:04:29.880 more lenses to focus light in this way. In fact, 0:04:29.920 --> 0:04:32.760 when we say a camera lens, you really mean a 0:04:32.800 --> 0:04:36.920 series of lenses that are encased in some sort of 0:04:37.080 --> 0:04:40.080 form factor. It's very easy to think of a camera 0:04:40.160 --> 0:04:43.560 lens as being a single thing, because you might have 0:04:44.240 --> 0:04:47.000 a camera body and then a selection of different lenses 0:04:47.000 --> 0:04:49.400 where you can attach or detach them. But each of 0:04:49.440 --> 0:04:54.280 those lenses contains several glass lenses inside of it. It's 0:04:54.279 --> 0:04:58.000 not just a lens inside a casing. A digital cameras 0:04:58.000 --> 0:04:59.800 by the way, you do this too. Uh. They do 0:05:00.360 --> 0:05:04.800 use film to capture images. They use image sensors to 0:05:04.920 --> 0:05:07.919 capture them, but they still use lenses to focus the 0:05:08.000 --> 0:05:12.479 light onto the appropriate part of the camera. Uh. This 0:05:12.600 --> 0:05:16.880 being the sensor as opposed to photosensitive film. The image 0:05:16.920 --> 0:05:20.400 sensors are solid state devices, and the two most common 0:05:20.440 --> 0:05:23.240 types of image sensors are cc D also known as 0:05:23.360 --> 0:05:28.760 charge couple devices and c MOSS or CMOS complementary metal 0:05:28.760 --> 0:05:32.680 oxide semiconductors. But what the heck does that even mean? Well, 0:05:32.839 --> 0:05:36.240 CCD sensors are the older of the two, and they 0:05:36.400 --> 0:05:38.599 lead the way for many many years in terms of 0:05:38.640 --> 0:05:41.719 image quality. CCD was considered to be the top of 0:05:41.720 --> 0:05:45.640 the image quality. Was also more expensive and less energy efficient, 0:05:45.839 --> 0:05:50.080 but it produced the best images. A CCD sensor was 0:05:50.160 --> 0:05:53.160 able to produce much sharper, higher quality images, but s 0:05:53.320 --> 0:05:57.599 MOSS sensors are more efficient the more power efficient they're Also, 0:05:57.720 --> 0:06:01.080 there's also some more built in functionality with MOSS sensors 0:06:01.120 --> 0:06:04.040 than there are with C c d s. Plus, eventually 0:06:04.080 --> 0:06:07.120 the C MOSS quality pretty much caught up with and 0:06:07.279 --> 0:06:11.360 in some ways surpass C c ds. So while there 0:06:11.400 --> 0:06:16.799 are differences, and they're still different kinds of of cameras 0:06:16.800 --> 0:06:19.120 out there that have different kinds of sensors, and different 0:06:19.120 --> 0:06:23.400 photographers will favor one versus the other, the differences in 0:06:23.600 --> 0:06:28.200 quality between the two have largely diminished. It's not as 0:06:28.360 --> 0:06:32.760 dramatic as it once was. Uh, some budget cameras have 0:06:32.880 --> 0:06:34.599 C c d s, A lot of the better cameras 0:06:34.640 --> 0:06:38.200 have C mosses, but that's not that's not necessarily across 0:06:38.240 --> 0:06:41.920 the board. Like photo reactive film, those image sensors are 0:06:42.000 --> 0:06:46.240 photo sensitive. So when light in the form of photons 0:06:46.640 --> 0:06:50.000 hits the photo sites, photo sites are the little photo 0:06:50.040 --> 0:06:54.040 sensitive elements on these image sensors. Each photo site essentially 0:06:54.080 --> 0:06:57.680 is a pixel, so you get one photo site per pixel. 0:06:58.360 --> 0:07:00.760 When when a photon hits a photos ight, this produces 0:07:00.800 --> 0:07:04.520 an electrical charge. Now a bright light will end up 0:07:04.560 --> 0:07:07.800 generating a stronger charge than a dem light will. The 0:07:07.839 --> 0:07:12.240 photo sites must register the relative brightness of the light 0:07:12.320 --> 0:07:15.080 that's hitting them, as well as the relative brightness of 0:07:15.120 --> 0:07:18.320 the color of light that is red, green, or blue light, 0:07:18.680 --> 0:07:21.440 in order to reproduce color images. Otherwise you're just gonna 0:07:21.440 --> 0:07:24.560 get black and white images. Some sensors do this by 0:07:24.560 --> 0:07:28.640 having a color filter above each pixel photo site, uh, 0:07:28.680 --> 0:07:31.680 and they will have a different color filter over the 0:07:31.720 --> 0:07:35.080 different photo sites. This is called a Bear filter array 0:07:35.360 --> 0:07:38.680 is named after Bryce Beyer, who invented the array. It's 0:07:38.680 --> 0:07:43.239 a filter consisting of a mosaic or array of red, green, 0:07:43.360 --> 0:07:46.400 and blue filters above the photo sites on the sensor chip. 0:07:47.040 --> 0:07:49.280 So it looks kind of like a checkerboard, except you've 0:07:49.280 --> 0:07:53.800 got three colors, not two, and uh. One fourth of 0:07:53.880 --> 0:07:58.280 those filters are red, one fourth of those filters are blue, 0:07:58.680 --> 0:08:01.400 and the other half are green. So you have twice 0:08:01.400 --> 0:08:04.040 as many green as you have either blue or red. 0:08:04.400 --> 0:08:06.360 So why is that Why would you have twice as 0:08:06.360 --> 0:08:09.720 many green filters? Well, that's to mimic the color sensitivity 0:08:09.800 --> 0:08:13.240 of the human eye, because our color sensitivity is not 0:08:13.320 --> 0:08:17.280 even across the board from a biological standpoint. Some filters, 0:08:17.280 --> 0:08:20.280 some Bayer filters, will include different shades of green, so 0:08:20.360 --> 0:08:22.880 you won't just have one single shade of green filter 0:08:23.000 --> 0:08:24.720 you might have to you might have one that's like 0:08:24.760 --> 0:08:26.920 a dark forest green and one that's kind of more 0:08:26.920 --> 0:08:30.280 of a lighter green or something along those lines. Uh. 0:08:30.480 --> 0:08:34.480 Some filters will use clear filters in place of some 0:08:34.559 --> 0:08:38.479 of the green, and that's in order to improve light sensitivity, 0:08:38.480 --> 0:08:42.360 but in general, that's how most of those filters work. 0:08:42.840 --> 0:08:45.840 When the camera records an image, initially, it's a picture 0:08:45.880 --> 0:08:48.400 that's just red, green, and blue in the raw file. 0:08:48.559 --> 0:08:50.720 So if you were to look at a raw photo 0:08:50.760 --> 0:08:54.000 image before any processing has happened, it would look really weird. 0:08:54.080 --> 0:08:58.240 It wouldn't look color realistic at all. It could look 0:08:58.280 --> 0:09:01.840 pretty awful. It's what photogra is often called false color. 0:09:02.360 --> 0:09:05.800 But then you have this image processing unit that's part 0:09:05.880 --> 0:09:09.360 of digital cameras, and it can take this raw file 0:09:09.440 --> 0:09:12.959 and convert it into more natural imaging using what is 0:09:12.960 --> 0:09:17.559 called a demosaic stage, So it takes all that jumblee 0:09:17.640 --> 0:09:20.840 information and makes some meaning out of it. One cool 0:09:20.920 --> 0:09:23.520 thing that happens with this software is how it determines 0:09:23.559 --> 0:09:26.880 the color of an individual pixels. So remember images are 0:09:26.920 --> 0:09:30.400 made up of millions of these pixels. A pixels essentially 0:09:30.400 --> 0:09:32.680 a point of light, or if you prefer a single 0:09:32.760 --> 0:09:36.680 point of color in an image, and the more pixels 0:09:36.720 --> 0:09:39.360 you have, the smaller they Let's say that you have 0:09:39.400 --> 0:09:42.600 the same dimensions of a photo. Let's start with that. 0:09:43.120 --> 0:09:45.800 So let's say you're using an eight by ten photo. 0:09:45.960 --> 0:09:49.319 This is just for the purposes of explaining pixels and resolution. 0:09:50.480 --> 0:09:53.199 If you increase the resolution of an eight by ten photos, 0:09:53.200 --> 0:09:55.960 so you're not changing the size of the image, You're 0:09:56.000 --> 0:09:59.160 just changing the resolution. If you increase the resolution, that 0:09:59.200 --> 0:10:02.800 means you're packing or pixels into that same physical space. 0:10:02.880 --> 0:10:06.280 That means the pixels themselves have to be smaller. If 0:10:06.360 --> 0:10:10.480 you are reducing the resolution, you're decreasing the number of pixels. 0:10:10.480 --> 0:10:13.320 That means the size of each individual pixel gets larger. 0:10:14.120 --> 0:10:16.839 So if you've ever seen an image where someone has 0:10:16.880 --> 0:10:21.720 taken a digital photo that was for a very small 0:10:22.720 --> 0:10:25.240 set of dimensions and then they've blown it up, it 0:10:25.240 --> 0:10:29.920 looks very blocky or pixelated, that's because it was a 0:10:29.920 --> 0:10:33.040 certain resolution and a certain size, and now you've stretched it, 0:10:33.320 --> 0:10:36.199 so all of those pixels have individually been stretched as well. 0:10:36.880 --> 0:10:39.080 If you want to increase resolution, then you have to 0:10:39.120 --> 0:10:42.400 reduce pixel size to add more pixels to that image. 0:10:42.840 --> 0:10:45.720 This only works to make an image sharper up to 0:10:45.800 --> 0:10:49.800 a point. Higher resolution does not automatically mean a better 0:10:49.880 --> 0:10:53.000 quality of picture. There are other elements that are also 0:10:53.120 --> 0:10:57.920 very important, things like color representation, so in contrast ratio, 0:10:57.960 --> 0:11:00.840 So it's not just resolute shan that you need to 0:11:00.840 --> 0:11:03.559 think about. This is why we've done episodes of tech 0:11:03.600 --> 0:11:06.960 stuff in the past where we've talked about the megapixel myth, 0:11:07.440 --> 0:11:09.200 the idea that if you go out and buy a 0:11:09.240 --> 0:11:13.000 camera that has more megapixels that automatically takes better photos 0:11:13.040 --> 0:11:16.840 than a camera that has a lower megapixel count. That's 0:11:16.920 --> 0:11:20.200 just not necessarily the case. It might be true, but 0:11:20.320 --> 0:11:22.800 it's not because of the megapixels. It's because of other 0:11:22.880 --> 0:11:26.320 elements as well. The only time you really have to 0:11:26.320 --> 0:11:29.880 worry about really high megapixels counts as if you wanted 0:11:29.920 --> 0:11:31.960 to take a photo and then blow it up to 0:11:32.000 --> 0:11:34.760 a really large size and you didn't want to have 0:11:34.800 --> 0:11:37.120 too much distortion when you were doing that. You want 0:11:37.120 --> 0:11:39.839 a really high megapixel count so that you can do 0:11:39.880 --> 0:11:43.920 that without losing too much on the resolution side. Well, 0:11:43.960 --> 0:11:47.040 the way this image processing software tends to work is 0:11:47.080 --> 0:11:49.920 it looks at each individual pixel and then it looks 0:11:49.960 --> 0:11:53.199 at all that pixels neighbors and it starts to say, well, 0:11:53.240 --> 0:11:56.640 based upon all the neighbors of this pixel, what color 0:11:56.720 --> 0:12:01.880 should this specific pixel be. And it starts to extrapolate 0:12:01.960 --> 0:12:05.040 information based on this and it makes some decisions about 0:12:05.080 --> 0:12:08.520 what color each pixel should be based upon the nature 0:12:08.640 --> 0:12:12.040 of its neighbors. And as it turns out, this software 0:12:12.120 --> 0:12:16.959 is pretty good. It can reproduce colors fairly faithfully. And 0:12:17.000 --> 0:12:19.040 that seems pretty interesting to me when you think that. 0:12:19.080 --> 0:12:20.839 You know, when you how do you start with that? 0:12:21.000 --> 0:12:25.080 Right when you first start, you don't have any necessarily 0:12:25.160 --> 0:12:28.000 any natural colors. It's all the red, green, blue. But 0:12:28.480 --> 0:12:32.840 by using this kind of process of deduction, the image 0:12:32.840 --> 0:12:37.720 processing software can create a natural looking image. And it's 0:12:37.720 --> 0:12:40.480 all about again looking at all the neighbors of this 0:12:40.559 --> 0:12:42.920 one pixel to determine what color it should be. This 0:12:43.000 --> 0:12:45.600 happens super fast. By the way, it sounds like something 0:12:45.600 --> 0:12:47.720 that would take a really long time, but the software 0:12:47.760 --> 0:12:50.440 is incredibly fast. Now. Bear filters are found in many, 0:12:50.480 --> 0:12:54.440 but not all cameras. Some use other types of filter systems. Uh. 0:12:54.480 --> 0:12:57.960 The fob on sensor has red, green, and blue filters 0:12:58.040 --> 0:13:01.000 over every single photo site. So instead of having either 0:13:01.120 --> 0:13:04.839 a red, or green, or blue filter over each photo site. 0:13:05.240 --> 0:13:08.640 Every single photo site has all three, and some photographers 0:13:08.640 --> 0:13:14.479 say that this produces more natural looking images. Others say 0:13:14.520 --> 0:13:19.800 they are completely bonkers crazy wah wah, who uh. I 0:13:19.840 --> 0:13:23.440 guess it really depends on your perspective and what you 0:13:23.520 --> 0:13:27.080 have experience with um, because as far as I can tell, 0:13:27.280 --> 0:13:30.840 it's not that different. But then my visual acuity is 0:13:30.880 --> 0:13:34.520 nowhere near the level of say a professional photographers. The 0:13:34.600 --> 0:13:37.360 important thing to remember here is that the image sensor 0:13:37.520 --> 0:13:41.840 generates an electrical charge based off these photons, whether it's 0:13:42.000 --> 0:13:44.440 measuring the red, green, or blue, or just the amount 0:13:44.440 --> 0:13:47.720 of light in general, and the camera has to measure 0:13:47.760 --> 0:13:52.160 that that electrical charge and convert that value into a 0:13:52.240 --> 0:13:55.080 digital value, or convert that electrical charge into a digital 0:13:55.160 --> 0:13:59.559 value with an analog to digital converter, and then you're 0:13:59.600 --> 0:14:02.720 left with digital data for all the processing. So the 0:14:02.840 --> 0:14:05.440 software does all the work on the back end, so 0:14:05.480 --> 0:14:08.079 you've got the hard work on the front end where 0:14:08.120 --> 0:14:10.760 the image has to her image sensor has to take 0:14:10.800 --> 0:14:16.679 all these this light information converted into electrical charges, measure 0:14:16.720 --> 0:14:19.560 that and convert that into digital information, and then the 0:14:19.600 --> 0:14:22.960 software takes that digital information and interprets it to create 0:14:23.000 --> 0:14:25.880 the image that you're looking at. There are other things 0:14:25.880 --> 0:14:28.520 that take into consideration here, such as the shutter speed 0:14:28.600 --> 0:14:31.000 of the camera. Shutter is a device that cuts off 0:14:31.120 --> 0:14:34.600 light exposure to the sensor in the camera. The shutter 0:14:34.600 --> 0:14:40.040 speed determines the exposure time of a camera's UH the 0:14:40.080 --> 0:14:43.640 image that you're taking, So a short exposure is good 0:14:43.680 --> 0:14:45.560 if you're trying to take an image of something that 0:14:45.760 --> 0:14:48.360 is moving very quickly, but you need a lot of 0:14:48.440 --> 0:14:51.720 light to do that. You don't you're not leaving the 0:14:51.720 --> 0:14:54.720 shutter open very long for light to get to the 0:14:54.760 --> 0:14:57.680 image sensor. The same was true of film cameras. You 0:14:57.680 --> 0:14:59.640 need a very fast shutter speed, but you need a 0:14:59.680 --> 0:15:01.960 lot of light in order for the light to be 0:15:02.000 --> 0:15:06.680 able to register against the the photo sensitive material, whether 0:15:06.720 --> 0:15:10.240 it was film or an image sensor. If you're using 0:15:10.240 --> 0:15:13.200 a very slow shutter speed where the shutter is open 0:15:13.240 --> 0:15:15.600 for a longer amount of time, and this is all 0:15:15.640 --> 0:15:18.080 relative by the way, we're talking about fractions of a second. 0:15:18.920 --> 0:15:20.920 If you're leaving the shutter speed open for a longer 0:15:20.920 --> 0:15:25.160 amount of time, that's great for low light UH activities 0:15:25.160 --> 0:15:26.640 like if you want to take an image of something 0:15:26.680 --> 0:15:30.120 that's in really low light. Using a longer shutter speed 0:15:30.160 --> 0:15:34.960 where it's open longer is a good idea, But any 0:15:35.080 --> 0:15:38.280 motion is going to insert a lot of blur into 0:15:38.320 --> 0:15:41.760 that image, So you want the you wanted to be 0:15:41.880 --> 0:15:45.920 very still. If you're going to use a slower shutter speed. 0:15:46.160 --> 0:15:48.800 If you wanted to do something like high speed camera 0:15:48.880 --> 0:15:51.320 work where you're going to show something back in super 0:15:51.400 --> 0:15:55.400 slow motion, that shutter is moving incredibly fast. You're talking 0:15:55.400 --> 0:16:00.960 about taking thousands of images every single second. In order 0:16:01.000 --> 0:16:03.120 to do that, you have to have it very very 0:16:03.160 --> 0:16:07.400 well lit. So if you've ever been on a set 0:16:07.480 --> 0:16:12.240 where they're shooting super high speed UH footage in order 0:16:12.280 --> 0:16:15.360 to show back at at slow speed, then you know 0:16:15.360 --> 0:16:18.120 what I'm talking about. The lights in those kind of 0:16:18.160 --> 0:16:21.280 situations tend to be out of control. They're really actually 0:16:21.360 --> 0:16:24.920 very much in control. They're just extremely well lit. So 0:16:25.280 --> 0:16:27.400 the whole reason I wanted to talk about this bit 0:16:27.640 --> 0:16:29.920 is that if you move a camera around while you're 0:16:29.960 --> 0:16:33.400 taking images, whether it's snapshot or video, you can end 0:16:33.440 --> 0:16:35.840 up with a jittery mess and it's really difficult to 0:16:35.880 --> 0:16:40.480 hold completely still. If you've ever tried to just hold still, uh, 0:16:40.520 --> 0:16:42.480 you know even if you've got the hands of a surgeon, 0:16:42.520 --> 0:16:44.480 you're gonna notice a little bit of a jitter in there. 0:16:44.960 --> 0:16:47.800 Most of us have some of that whenever we operate 0:16:47.800 --> 0:16:51.040 a camera. Some of it's more obvious than others, and 0:16:51.840 --> 0:16:54.600 a little bit people can tend to look past, but 0:16:54.680 --> 0:16:56.600 more than a little bit, it gets very distracting. So 0:16:56.640 --> 0:16:58.960 if you ever wanted to do something like move around 0:16:59.080 --> 0:17:03.200 while you're shoeing video, you might have so much jetd 0:17:03.360 --> 0:17:05.400 that's distracting, and you want to figure out a way 0:17:05.440 --> 0:17:07.760 to reduce that, and we found numerous ways to cut 0:17:07.800 --> 0:17:10.879 that down. Some ways just involved locking down the camera, 0:17:11.160 --> 0:17:14.000 so you might put it in a tripod, and then 0:17:14.240 --> 0:17:18.960 the camera's motion is strictly limited and you can operate 0:17:19.000 --> 0:17:20.959 it without it having too much jetter. But then you 0:17:21.040 --> 0:17:24.760 have a pretty static shot. You could use cranes and dollies, 0:17:24.840 --> 0:17:29.240 which tend to have very smooth movement along certain directions, 0:17:29.240 --> 0:17:31.480 but you are limited in the ways you can move 0:17:31.680 --> 0:17:35.640 the camera. In both of those, you can't go anywhere 0:17:35.640 --> 0:17:37.560 an actor can go. For example, if you wanted to 0:17:37.560 --> 0:17:41.399 shoot a film, you could put the camera on a 0:17:41.560 --> 0:17:45.960 dolly that isn't on a track. So there's some track systems. 0:17:46.119 --> 0:17:48.560 Those are great. If you have uneven ground or rough 0:17:48.680 --> 0:17:50.760 terrain and you put tracks down, you can have a 0:17:50.840 --> 0:17:53.560 nice smooth camera operation, but again you're limited to just 0:17:53.680 --> 0:17:56.160 moving on the tracks, or you could do a wheeled 0:17:56.320 --> 0:17:59.360 dolly that can move around uh an area, but then 0:17:59.400 --> 0:18:02.280 you need some that's an area that's pretty smooth and flat. 0:18:02.920 --> 0:18:05.080 That really still limits what you can do with a camera. 0:18:05.640 --> 0:18:09.679 So one of the ways that we have come up 0:18:09.760 --> 0:18:13.320 with two improve camera operation and to be able to 0:18:13.359 --> 0:18:15.880 go in the same places where it say performers can 0:18:15.920 --> 0:18:20.879 go is with steadicams. Now. The steadicam was invented in 0:18:20.920 --> 0:18:23.880 the mid nineteen seventies as a way for camera operators 0:18:23.920 --> 0:18:26.680 to move around while running a camera and get smooth footage. 0:18:26.960 --> 0:18:30.400 Whether it's with a documentary or with a film where 0:18:30.400 --> 0:18:33.560 you're following actual characters. The steadicam can remove the jitter 0:18:33.760 --> 0:18:36.800 created by walking or running. And when we move around 0:18:36.840 --> 0:18:39.440 in our own bodies, this is happening all the time. 0:18:39.480 --> 0:18:42.040 There's always this jitter as we're walking or running, but 0:18:42.080 --> 0:18:44.280 we don't really notice it because our brains are really 0:18:44.280 --> 0:18:46.199 good at smoothing all of that out to create a 0:18:46.200 --> 0:18:49.640 more steady experience in our consciousness. So it's only when 0:18:49.640 --> 0:18:53.480 you're really concentrating on it that you become aware of it. 0:18:53.480 --> 0:18:55.199 It's sort of similar to how the world does not 0:18:55.320 --> 0:18:58.560 spontaneously go away and then come back every time you 0:18:58.600 --> 0:19:01.160 blink your eyes. It's only when you really think about 0:19:01.200 --> 0:19:03.720 blinking your eyes that you start to notice it. And 0:19:03.760 --> 0:19:08.120 now you're noticing it, aren't you sorry about that? My bad? 0:19:08.640 --> 0:19:10.840 Don't worry. You'll you'll stop thinking about blinking your eyes 0:19:10.840 --> 0:19:12.879 in just a minute and then everything will be fine. 0:19:13.440 --> 0:19:16.080 So how does a steady cam stabilize the image so 0:19:16.080 --> 0:19:18.640 that you can get the really cool effects you see 0:19:18.680 --> 0:19:21.920 in movies like the Amazing copa shot in Good Fellas. 0:19:22.280 --> 0:19:24.520 So if you're not familiar with the sequence I'm talking about, 0:19:24.920 --> 0:19:27.840 it's it's in the movie. It's well in the film. 0:19:27.920 --> 0:19:30.880 There's a sequence that's a little bit more than two 0:19:30.880 --> 0:19:34.320 and a half minutes long, and it's a an uncut shot. 0:19:34.640 --> 0:19:38.040 It tracks two characters as they cross a New York street. 0:19:38.560 --> 0:19:41.879 They walk down some stairs into the kitchen of a 0:19:41.920 --> 0:19:46.080 busy restaurant. They walk through the kitchen, passing by dozens 0:19:46.080 --> 0:19:48.800 of extras as they're moving around in the kitchen as 0:19:48.840 --> 0:19:53.520 cooks or or waiters, they emerge onto a dining floor 0:19:54.359 --> 0:19:56.439 and they're seated right up front at a stage. And 0:19:56.480 --> 0:19:59.480 this is all in a single uncut shot. And it 0:19:59.560 --> 0:20:02.119 was using a film camera, not a digital camera. This 0:20:02.240 --> 0:20:05.640 was in the early nineties, and there were other challenges there. 0:20:05.640 --> 0:20:08.240 For example, the way the film camera worked. It was 0:20:08.280 --> 0:20:11.920 actually pulling film from one side of the camera and 0:20:12.000 --> 0:20:15.040 it would go in through the camera where you would 0:20:15.040 --> 0:20:18.480 expose the film to light coming through the camera. The 0:20:18.560 --> 0:20:21.800 exposed film would then go into a canister on the 0:20:21.800 --> 0:20:24.080 other side of the camera. Now, what that meant was 0:20:24.119 --> 0:20:26.840 that as you shot film, the weight of the camera 0:20:26.920 --> 0:20:32.400 started to shift because the side worthy the the exposed 0:20:32.440 --> 0:20:34.879 film was coming out that started getting heavier and heavier, 0:20:35.240 --> 0:20:37.399 so you had to compensate for that. It's actually pretty 0:20:37.400 --> 0:20:40.359 remarkable if you watched that sequence that it doesn't just 0:20:40.400 --> 0:20:43.480 slowly start tilting towards the right because the camera was 0:20:43.520 --> 0:20:47.200 getting progressively more heavy on the right side. By the way, 0:20:47.240 --> 0:20:52.000 if you're curious, there's some great behind the scenes documentaries 0:20:52.040 --> 0:20:54.840 and interviews about the copa shot. It's one of those 0:20:54.880 --> 0:20:57.840 things that's talked about in film school. It took them 0:20:57.920 --> 0:21:01.160 eight takes and that was it, and they actually finished 0:21:01.160 --> 0:21:03.640 it in half a day of shooting, which if you've 0:21:03.640 --> 0:21:06.399 ever been on a film shoot for a two and 0:21:06.440 --> 0:21:10.400 a half minute sequence uncut, to get completed in eight 0:21:10.440 --> 0:21:13.680 takes is pretty phenomenal. Also, it tells you the difference 0:21:13.720 --> 0:21:19.080 between someone, say like Martin Scorsese and Stanley Kubrick, because 0:21:19.119 --> 0:21:21.960 if Kubrick were shooting that, he would have still been 0:21:21.960 --> 0:21:27.439 doing it years later. Uh. Anyway, the steadicam was a 0:21:27.480 --> 0:21:31.000 big part of why this shot was even possible because 0:21:31.040 --> 0:21:34.640 it meant following these actors through these different environments, including 0:21:34.720 --> 0:21:39.240 downstairs and through a crowded kitchen, and it was something 0:21:39.280 --> 0:21:42.000 that you just could not do on a track or 0:21:42.080 --> 0:21:45.600 on a wheeled dolly. Uh. And it was all because 0:21:45.600 --> 0:21:47.679 of the steadicam. Now, let's say he came itself was 0:21:47.720 --> 0:21:50.439 invented again in the seventies, not in the nineties. It 0:21:50.520 --> 0:21:52.520 was invented by a guy named Garrett Brown who was 0:21:52.560 --> 0:21:55.240 a commercial director and a producer. He was not an engineer, 0:21:55.480 --> 0:21:57.520 and he was just trying to come up with a 0:21:57.600 --> 0:22:01.320 way to remove all this jetter that was coming around 0:22:01.400 --> 0:22:04.359 whenever you wanted to do a handheld shot and you 0:22:04.400 --> 0:22:07.480 wanted to follow an actor along a place where you 0:22:07.520 --> 0:22:11.880 couldn't have these other traditional camera setups, and he got 0:22:11.880 --> 0:22:15.720 a rough idea of it and called it the Brown's Stabilizer. 0:22:15.760 --> 0:22:19.359 In nine three, fortunately, a more refined version would become 0:22:19.400 --> 0:22:23.280 the first steadicam, and it consisted of three components. You 0:22:23.320 --> 0:22:28.639 had an articulated iso elastic arm that would attach to 0:22:29.200 --> 0:22:33.640