WEBVTT 1 00:00:01.966 --> 00:00:05.733 Hey guys and welcome to this video! We're going to talk a little bit about 2 00:00:05.733 --> 00:00:09.733 the Lean Production process, start getting you thinking about your work flow! Right, 3 00:00:09.733 --> 00:00:13.733 Mikey!? Yes! Now, I've struggled with the concepts of Lean Production, because 4 00:00:13.733 --> 00:00:17.733 that's not generally how we're taught to do things. We do 5 00:00:17.733 --> 00:00:21.733 everything, almost, in our lives in batches, and when you switch over to Lean 6 00:00:21.733 --> 00:00:25.733 Production it actually makes things a lot easier, but it takes a while to get your 7 00:00:25.733 --> 00:00:29.733 head around and Ben's going to explain that a bit more in detail now. Yeah! So what's 8 00:00:29.733 --> 00:00:33.733 lean about? Really it's about three things. Lean is about delivering value 9 00:00:33.733 --> 00:00:37.733 to the customer, i.e. something that the customer will pay for, the 10 00:00:37.733 --> 00:00:41.733 end customer. So say you are making bowling pins and a bowling ball, in this case, 11 00:00:41.733 --> 00:00:45.733 and you're handing it over to a game developer, 12 00:00:45.733 --> 00:00:49.733 eventually there's going to be a game made, and somebody on the app store is going to be flicking through and deciding 13 00:00:49.733 --> 00:00:53.733 whether they're going to buy that game. And if what you're about to do isn't going to add to that end 14 00:00:53.733 --> 00:00:57.733 experience, it's not value. So that's what value is, it's about delivering something 15 00:00:57.733 --> 00:01:01.733 you'll pay for. Yep! And in this case, that's exactly how we're making 16 00:01:01.733 --> 00:01:05.733 the course. We output something to you guys as quickly as possible, and then let you guys give 17 00:01:05.733 --> 00:01:09.733 us the feedback, as quickly as possible. And you want to employ that, basically, 18 00:01:09.733 --> 00:01:13.733 in your modeling, so you want to make sure you get to the end result as quickly as possible. Find 19 00:01:13.733 --> 00:01:17.733 all the problems you're going to have and then jump back again and reiterate and make 20 00:01:17.733 --> 00:01:21.733 them better. It's all about a feedback cycle, making sure that you get that feedback 21 00:01:21.733 --> 00:01:25.733 quickly and improve your model or improve the course or whatever 22 00:01:25.733 --> 00:01:29.733 you're working on, as quickly as possible. Yeah, so feedback is key. 23 00:01:29.733 --> 00:01:33.733 Mikey says deliver the end result. Now we need to be careful about that, let's deliver a first version of the 24 00:01:33.733 --> 00:01:37.733 end result. Yes! That's the whole point here! And like sometimes, maybe, 25 00:01:37.733 --> 00:01:41.733 when an images loads on the web and it's just really fuzzy and then it gets sharper and sharper. 26 00:01:41.733 --> 00:01:45.733 That's kind of like Lean, because it's giving you an impression of the entire thing. Very 27 00:01:45.733 --> 00:01:49.733 early on you can go, am I or am I not interested in waiting for the rest of this image? 28 00:01:49.733 --> 00:01:53.733 And that's a bit like your bowling pins or your bowling ball. What you want to do is mock something super 29 00:01:53.733 --> 00:01:57.733 simple up, get it through to your game developer, for example, and then, get their feedback. 30 00:01:57.733 --> 00:02:01.733 They might go, oh well the pins have got too many polygons for my colliders ore the ball doesn't look right 31 00:02:01.733 --> 00:02:05.733 or the axis is wrong or the handedness is wrong or something. So get a really 32 00:02:05.733 --> 00:02:09.733 basic version and then keep iterating. So in this video we want to help you to understand 33 00:02:09.733 --> 00:02:13.733 that Lean is not about Pre-Production, Production, and Post Production. 34 00:02:13.733 --> 00:02:17.733 We want you to imagine that you're creating these bowling assets, 35 00:02:17.733 --> 00:02:21.733 these assets for a bowling game, computer game. Yep! We 36 00:02:21.733 --> 00:02:25.733 want to help you understand how many iterations you need to do. Yes that's very important, because it's very 37 00:02:25.733 --> 00:02:29.733 easy to keep going and keep going and keep going and searching for that 38 00:02:29.733 --> 00:02:33.733 100% perfect and that is an unobtainable goal. What we really want is 39 00:02:33.733 --> 00:02:37.733 that 80/20 rule! You've probably heard of that, we want to get to that core 20% 40 00:02:37.733 --> 00:02:41.733 that is going to produce 80% of our outcome. Absolutely! So this 41 00:02:41.733 --> 00:02:45.733 production method is a way of discovering. Your first base is to discover 42 00:02:45.733 --> 00:02:49.733 20% of the work that you need to do, just 20% of all the work you could do, 43 00:02:49.733 --> 00:02:53.733 that's going to deliver 80% of the value. That might well be that you're a perfectionist and the 44 00:02:53.733 --> 00:02:57.733 80%'s not good enough and that's fine, but then, we'll work out what the 30% is that gives you 45 00:02:57.733 --> 00:03:01.733 95% of the value and so on. Diminishing returns 46 00:03:01.733 --> 00:03:05.733 can rear their ugly head very quickly, whenever you're doing anything like 3D. 47 00:03:06.333 --> 00:03:09.333 So in the slides for the course, we've given you a little bit of information 48 00:03:09.333 --> 00:03:13.333 here. Let's just define Lean again, this is just a summary of Wikipedia. It says that it eliminates 49 00:03:13.333 --> 00:03:17.333 waste created through overburden or an uneven production process. 50 00:03:17.333 --> 00:03:21.333 So let's talk about that for a second. It's all about not wasting your time 51 00:03:21.333 --> 00:03:25.333 or not wasting any input, before 52 00:03:25.333 --> 00:03:29.333 you get stuff to the customer. So I've already produced some bowling pins and a 53 00:03:29.333 --> 00:03:33.333 bowling ball for Ben and his Unity course that he's working on, and 54 00:03:33.333 --> 00:03:37.333 the first thing I did was I went away and produced a wonderful, great big 55 00:03:37.333 --> 00:03:41.333 high definition, bowling pin and we soon found we were 56 00:03:41.333 --> 00:03:45.333 running into problems. Now, I spend a lot of time getting that, in my mind, 57 00:03:45.333 --> 00:03:49.333 bowling pin perfect for Ben, but, in fact, it wasn't perfect for Ben, because I didn't 58 00:03:49.333 --> 00:03:53.333 understand the constraints he was under with, what he was doing. So then, we 59 00:03:53.333 --> 00:03:57.333 very quickly went into a series of rapid iterations to get to 60 00:03:57.333 --> 00:04:01.333 exactly what Ben wanted, and that was important because, otherwise, it's 61 00:04:01.333 --> 00:04:05.333 wasting my time generating all that stuff that's of no value to Ben 62 00:04:05.333 --> 00:04:09.333 and then ultimately you guys, if you're taking the course. 63 00:04:09.333 --> 00:04:13.333 But ultimately I'm also wasting Ben's time, since he sat there waiting for an asset. 64 00:04:13.333 --> 00:04:17.333 So it goes both ways, about eliminating both sides of the waste. 65 00:04:17.333 --> 00:04:21.333 So the whole point is to really plan your work, right, so we want to eliminate waste. 66 00:04:21.333 --> 00:04:25.333 We want to maximize the flow or value, i.e. something that the customer will pay for, 67 00:04:25.333 --> 00:04:29.333 and ultimately it's about gaining the highest possible quality, it's about finding a way 68 00:04:29.333 --> 00:04:33.333 of making that thing that is at the highest possible quality. 69 00:04:33.333 --> 00:04:37.333 Now, I must stress at this point, it is a learning journey to get 70 00:04:37.333 --> 00:04:41.333 to a Lean production, and you are going to find that it's difficult, 71 00:04:41.333 --> 00:04:45.333 from time to time, to adjust to that, because you'll want to achieve the best results 72 00:04:45.333 --> 00:04:49.333 you possibly can. However, we urge you to iterate as many 73 00:04:49.333 --> 00:04:53.333 possible times that you can. In fact, there's a good fable about this, and Ben 74 00:04:53.333 --> 00:04:57.333 tells it incredibly well! Well I'll summarize it here but there's a fable that 75 00:04:57.333 --> 00:05:01.333 may be an approcraphle story, but the idea is that a teacher of a pottery class says guys, I'm going to 76 00:05:01.333 --> 00:05:05.333 divide the class in two this term. One half of you are going to be rewarded with a prize at the end 77 00:05:05.333 --> 00:05:09.333 of the term, for those who have made the most, the person who has made the 78 00:05:09.333 --> 00:05:13.333 most pottery, we're just going to put it on a scale and see who's made the most. On the other side of the class, 79 00:05:13.333 --> 00:05:17.333 we are going to reward the person who makes the best piece of pottery, and guess what, 80 00:05:17.333 --> 00:05:21.333 at the end of the term it was both the person that made the most pottery that also made the single 81 00:05:21.333 --> 00:05:25.333 best piece of pottery, which is kind of against what you'd think. You'd have thought that the person who 82 00:05:25.333 --> 00:05:29.333 labors over one supposed masterpiece would produce the best thing, but actually, the process of 83 00:05:29.333 --> 00:05:33.333 iterating and just doing more stuff, throwing more at the wall, 84 00:05:33.333 --> 00:05:37.333 and maybe more sticks. So that's the concept here, it's get 85 00:05:37.333 --> 00:05:41.333 busy, get active, and get the entire pipeline done. 86 00:05:41.333 --> 00:05:45.333 We're going to be showing you that the bowling pin that you create, for example, 87 00:05:45.333 --> 00:05:49.333 in a game engine needs a very different type of mesh 88 00:05:49.333 --> 00:05:53.333 for collision detection, and that mesh has some constraints on it. It must be convex 89 00:05:53.333 --> 00:05:57.333 and it must have 255 or less polygons, 90 00:05:57.333 --> 00:06:01.333 and that is very very different to the visual mesh that you're going to create. 91 00:06:01.333 --> 00:06:05.333 And in future sections we'll also be talking about things like normal maps 92 00:06:05.333 --> 00:06:09.333 or bump maps and how this ties into the process. So please get it through the whole pipeline! You know 93 00:06:09.333 --> 00:06:13.333 we showed you about handedness... Yep! ...and about left and right-handed coordinate 94 00:06:13.333 --> 00:06:17.333 systems and also about what's up. Yep! Yep! 95 00:06:17.333 --> 00:06:21.333 Y or Z, all of that you want and scales, that's the other thing. Get all that 96 00:06:21.333 --> 00:06:25.333 out early or the game guy may go, the scale is all wrong and it's pointing in the wrong direction! 97 00:06:25.333 --> 00:06:29.333 Obviously with a bowling pin that's not going to be an arduous task to rotate it, but if you 98 00:06:29.333 --> 00:06:33.333 just produced 1,000 assets for something and they're all the wrong way up, that's going to 99 00:06:33.333 --> 00:06:37.333 annoy, well it's going to annoy everybody in the production process. However, if you're 100 00:06:37.333 --> 00:06:41.333 running a Lean production, the first asset they receive, they could have given you that feedback 101 00:06:41.333 --> 00:06:45.333 and no more problem further on. So in your mind remember the progressive 102 00:06:45.333 --> 00:06:49.333 JPEG! You get to that website, you get a fuzzy image, and then, it gets sharper and sharper 103 00:06:49.333 --> 00:06:53.333 and sharper. Even Netflix does that when you watch it, and the thing is, you can make a decision 104 00:06:53.333 --> 00:06:57.333 about the complete product and say, yes I do or I don't want it in that form. You can 105 00:06:57.333 --> 00:07:01.333 change the film, before you have to wait for that high def stream to buffer; 106 00:07:01.333 --> 00:07:05.333 you can change the image, before you wait for the full thing to download. The same thing, get your models 107 00:07:05.333 --> 00:07:09.333 to the end user as quickly as possible, and then, improve and improve and improve. 108 00:07:09.333 --> 00:07:13.333 And that's where the box modeling, we've already covered, comes in, that's literally your 109 00:07:13.333 --> 00:07:17.333 placeholder asset, that you would move about to see whether it fits in your scene, 110 00:07:17.333 --> 00:07:21.333 to see whether it's the right size, scale, upright, and everything else. So 111 00:07:21.333 --> 00:07:25.333 that's why that's core to getting that simple model out there, straight away. And also, 112 00:07:25.333 --> 00:07:29.333 all the unexpected things won't catch you out either! There can be a million things you can imagine 113 00:07:29.333 --> 00:07:33.333 maybe going wrong, to do with lighting or reflections or scale or rotation, 114 00:07:33.333 --> 00:07:37.333 it goes on and on and on. And by progressively building up the entire 115 00:07:37.333 --> 00:07:41.333 pipeline and not just working on your model in isolation, but delivering it out, delivering it out, 116 00:07:41.333 --> 00:07:45.333 rendering it, post processing, the whole lot, then you're going to avoid 117 00:07:45.333 --> 00:07:49.333 those mistakes. Another brief example we gave, when chatting about it early, was maybe you've been 118 00:07:49.333 --> 00:07:53.333 asked to just do a close in realistic thing of somebody's face. 119 00:07:53.333 --> 00:07:57.333 And maybe you missed the part of the brief that they're going to crop everything but their eyes and you spent three days 120 00:07:57.333 --> 00:08:01.333 modeling an ear and it's not going to appear as part of the scene. Or maybe something that's reflecting in the 121 00:08:01.333 --> 00:08:05.333 eyes is very important, that you've included in your scene, so it matches some real footage. 122 00:08:05.333 --> 00:08:09.333 It goes on! The point is, iterate like they do in the films and you'll be 123 00:08:09.333 --> 00:08:13.333 golden! Excellent! Well let's get on with some modeling! Let's do it!