about 2 months
ago -
Coffee Stain Studios
-
Direct link
Transcript (by Youtube)
0s | like to really like drive that feeling |
---|---|
2s | of like that dread like oh I need to go |
4s | I'm about to sh my pants we've |
7s | made a lot of trailers here at Cain over |
10s | the years working on satisfactory and |
12s | I've always wanted to make a video sort |
13s | of explaining the process in how we make |
16s | trailers So today we're doing exactly |
18s | that I'm going to walk you through the |
19s | process from ideation to post- |
21s | production throughout the entire thing |
23s | and uh we're going to have jolly time so |
25s | let's dive into |
27s | it so in this video I'd like to |
29s | highlight like all the aspects to like |
31s | how we make these trailers like how we |
32s | make them why we make them and when we |
34s | make them to sort of give you insight as |
36s | to like how much work goes into these |
38s | trailers and we're going to be using the |
40s | release dates trailer for 1.0 or the |
43s | toilet flush trailer uh as sort of an |
45s | example or a use case uh for this entire |
48s | video uh the reason for that is because |
50s | I was the one directing that one so I |
51s | know the ins and outs of that trailer |
54s | specifically it's also a somewhat recent |
56s | trailer so hopefully most things will be |
58s | fresh in my mind as to like why and how |
60s | we did things and it's also a short |
62s | trailer with very few Cuts in it so it |
65s | should be sort of easy to sort of |
67s | demonstrate um so let's dive into it so |
70s | when it comes to making trailers the |
71s | first step for us is always ideation uh |
74s | the first thing really is to figure out |
76s | sort of what is the purpose of the |
78s | trailer that you're about to make |
79s | because most of our trailers serve very |
81s | different purposes and they also have |
83s | very different audiences some trailers |
85s | are there to build hype for new players |
87s | who've never seen satisfactory before |
89s | some trailers we made are there to like |
91s | explain the game for people who are have |
93s | heard of satisfactory but are interested |
95s | and want to learn more whenever we made |
97s | our update trailers those are more |
99s | catered towards our existing community |
101s | and to sort of highlight like all the |
103s | new things and give you you know make |
106s | you excited to want to play the update |
108s | basically and there's a ton of other |
109s | reasons as to why you'd want to make a |
112s | specific trailer like sometimes it's |
113s | there to uh announce a game and make |
116s | people interested in it it might be also |
118s | there to like kind of bring in |
120s | investment to a game that you're working |
122s | on or you know Get Hype and build hype |
125s | and all that stuff sometimes I get the |
127s | question of like why are game developers |
129s | making these cinematic trailers when |
131s | like there's no gameplay footage so we |
133s | have no idea what the actual game is |
134s | about and you know they have different |
136s | purposes uh cinematic trailers usually |
138s | like that are there to make you excited |
140s | about the concept and also to bring in |
142s | some investment to Studios or even bring |
144s | in Talent uh wanting to work on some |
146s | games so yeah there's many different |
149s | reasons as to why you'd want make a |
150s | trailer for your game and one thing when |
152s | it comes to our update trailers is that |
154s | we put a lot of effort into our update |
156s | trailers like sometimes they take months |
158s | to make uh the 1.0 release trailer took |
162s | more than a year to make I think um and |
166s | it might be strange when you think about |
167s | it because the trailer is just like a |
169s | one minute two-minute video like why |
171s | would you want to spend so many |
172s | resources on making such a like small uh |
176s | little thing but they actually have a |
177s | pretty significant impact on you you |
180s | know how we Market our game and like how |
183s | people are able to sort of you know take |
185s | a look at our game and and see what it's |
187s | all about and the thing for us is like |
189s | we put a lot of effort into our updates |
192s | and we want that to be reflected in our |
194s | trailers as well because if we have a |
196s | trailer that you can tell we put a lot |
198s | of effort into then that will also |
200s | signify that like oh you put probably |
202s | put a lot of effort into the update as |
203s | well so that's why it's so important for |
206s | us to have the sort of the quality that |
208s | we have for our trailers uh whenever we |
210s | release them because it's kind of wasted |
211s | if you put a lot of effort into your |
213s | game and you don't put a lot of effort |
215s | into showcasing that game or you know |
217s | putting that effort into your trailer |
218s | then people are going to see that and |
220s | they're it's not going to reflect good |
222s | on it so you really want to match you |
224s | know the quality of what you're making |
226s | and there's a lot of things you have to |
227s | take into consideration when you're in |
229s | this ideation process of like figuring |
231s | out like what is the purpose of the |
233s | trailer you know there are questions you |
234s | need to ask especially like you know do |
236s | people know about your game like who is |
238s | the trailer for and when you're figuring |
239s | out the purpose of the trailer you also |
240s | need to sort of figure out like what is |
242s | the context behind the trailer you know |
244s | like is this a thing where people know |
246s | about the game beforehand or is this |
248s | like a reveal trailer of sorts um you |
251s | know is this a a game that's been out |
253s | for a long time so people already know |
254s | about the title uh you know all these |
256s | kind of things kind of matter when you |
258s | decide like how you make the trailer and |
260s | like what's in the trailer and and |
262s | whatnot and context is also important |
264s | when it comes to the trailer in like how |
266s | is it displayed like is this supposed to |
268s | be at like a big event with like a ton |
271s | of other reveal trailers as well or is |
273s | it just supposed to live on your YouTube |
275s | channel or like on social media um these |
278s | also impact the trailer a little bit for |
280s | instance if a trailer is just supposed |
281s | to live on social media then it's really |
284s | really really important that the first |
285s | like 15 seconds really catch you so like |
288s | if you're posting on social media it's |
289s | harder to make like a cinematic trailer |
291s | for instance where you have like a |
293s | company logo that fades in for like 10 |
295s | seconds or whatever like on social media |
297s | people typically click away if they're |
299s | not hooked life within you know a short |
301s | amount of time but like at events show |
303s | and stuff like that that's a whole |
305s | different ball game and then you can |
306s | sort of create a different type of |
308s | trailer um you know depending on on who |
310s | you are and and you know what the |
313s | context of it is I guess so it's really |
314s | important to figure out the purpose like |
316s | is this a reveal trailer is this an |
318s | announcement trailer is it supposed to |
319s | Showcase your game do you want people to |
321s | be like Blown Away by some technical |
323s | element of it or is it like a cinematic |
325s | trailer where you're more aiming to like |
327s | evoke a feeling or you know make people |
330s | interested or excited so at the end of |
333s | the day it's really really really |
334s | important in the adiation process that |
336s | you really figure out like what is the |
337s | main purpose of the trailer because like |
340s | I said context and effect matters a lot |
342s | for trailers like if you want people to |
343s | talk about your trailer or talk about |
345s | the game then you need to sort of hit |
348s | certain beats for it to make sense so |
350s | once we figured out like the purpose or |
352s | know why we're making this trailer or |
355s | yada yada yada we start concepting sort |
357s | of uh and start coming up with ideas as |
360s | like what will happen in the trailer and |
361s | like what is the format of the trailer |
364s | um and yeah like based on the purpose of |
365s | the trailer that really dictates what |
367s | happens in it if we're making a trailer |
369s | where we want people to sort of see |
370s | something cool and be blown away uh we |
372s | might do it in one way and if we want to |
374s | make a trailer where we're like |
375s | explaining the game and you know talking |
378s | about the game it's it's one thing for |
379s | instance we made a trailer a long time |
381s | ago uh where it's basically me just |
383s | explaining the game from start to finish |
386s | in in in like a goofy way uh that was |
389s | one way to make that trailer and that |
390s | trailer was developed in a certain way |
393s | uh because we wanted to do that and it's |
395s | up to you to utilize all the resources |
397s | you can find on that planet to create |
399s | multi-story factories and connect those |
401s | suckets with conveyor bels to create |
403s | tasty conveyor Bel spaghetti so that you |
404s | can produce all these cool items like |
406s | that thing and that thing and whatever |
408s | that is and some other trailers which |
410s | are more cinematic um like the update 3 |
413s | trailer for instance um you know that |
415s | one we really wanted to evoke a feeling |
417s | and and you know bring the hype in but |
419s | also to sort of showcase it in a real |
421s | setting so yeah based on all these like |
423s | parameters and and and whatnot that sort |
425s | of drives the ideation and the |
427s | conceptualization of your trailer and |
430s | sometimes these ideas come immediately |
432s | sometimes it takes a long time to figure |
433s | some of these things out um and you know |
437s | we usually sit down and brainstorm uh |
439s | these ideas and and discuss this a lot |
442s | uh and and sometimes it's external |
444s | people coming in and pitching to us uh |
446s | for certain ideas so this varies a lot |
449s | for us but usually we just sit down and |
451s | talk about the purpose and talk about |
453s | like what could we make and sometimes it |
456s | stays in the ideation face uh and |
459s | sometimes that brings into the next step |
463s | so in the case of the announcement |
464s | trailer uh the purpose was obviously we |
466s | wanted to announce the release date for |
469s | satisfactory uh we actually also |
471s | originally were going to announce |
473s | console in this trailer but here's the |
475s | thing too that also kind of affected the |
477s | trailer a bit um we didn't know 100% if |
480s | we could announce console at at when we |
483s | made this trailer uh because of various |
486s | different factors so we needed to design |
488s | a trailer where we 100% were announcing |
491s | the release date and we were like |
493s | potentially announcing console uh so we |
496s | really need to fit the trailer in such a |
498s | way where if we needed to cut the |
500s | console part portion of it then it |
502s | wouldn't really affect the trailer that |
504s | much and for the announcement trailer we |
506s | originally were going to have it at an |
508s | event uh we were originally pitching to |
511s | have this trailer be showcased at an |
513s | event the reason why we wanted to have |
514s | an event is because we wanted to catch |
517s | people that may have heard of |
518s | satisfactory but didn't want to play it |
520s | because they were waiting for 1.0 uh so |
522s | we really wanted to like hit the old |
524s | drum and let everyone know that like hey |
526s | 1.0 is coming out soon and it's coming |
528s | out at this time uh so get your |
532s | together and when you have a trailer at |
534s | an event it's really important to think |
536s | about the context here right because |
537s | we're showcasing this trailer in a block |
540s | of many trailers so there's a ton of |
542s | other games U that are doing sort of the |
544s | same thing that you're doing uh and some |
547s | games are really really big like if Call |
549s | of Duty was at the same event that we |
550s | were you know most people would probably |
552s | talk about Call of Duty or maybe not but |
553s | like you know some other big game you |
555s | know if Naughty Dog announces their new |
557s | game then who will remember the |
560s | satisfactory announ date reveal trailer |
562s | you know so it was really important for |
564s | us to like hook people uh in a block of |
566s | other trailers to make you know this |
568s | trailer kind of memorable so this event |
570s | we pitched two trailers actually we |
572s | pitched the regular 1.0 uh trailer that |
575s | we released with 1.0 and we pitched this |
578s | announcement trailer and uh we made this |
582s | pitch for the release the the the one we |
583s | actually made with the toilet flush uh |
586s | after the event was like it's too |
588s | generic like the the it's a cool trailer |
590s | and all all that uh but there's no |
593s | exciting new you know unveil or anything |
596s | like that so they didn't think it was |
598s | too interesting for us to have that |
599s | there so that's why we sort of came back |
601s | with them with this pitch we're like hey |
602s | but what about it this stupid trailer |
605s | and in the end we didn't end up having |
607s | the trailer on that at that event but we |
609s | really liked this idea of this trailer |
611s | so we decided to pursue it on our own |
613s | instead and the idea for like the |
614s | announcement trailer like came like that |
617s | uh actually so me and lenus the lenus is |
620s | like our main uh trailer guy uh and I'm |
623s | just like sort of the standin trailer |
624s | guy that comes in here and there and |
625s | make trailers uh we basically came to |
627s | each other at the same time with the |
629s | same idea and we're like we should do |
632s | this so like for the announcement |
634s | trailer like the idea of this trailer |
637s | you know came to be immediately like the |
639s | thing with the announcement trailer that |
641s | we thought would make it stand out was |
643s | the fact that we're kind of spoofing and |
644s | paring how other trailers were made um |
648s | and that's sort of how we wanted to |
650s | stick out or stand out at this event uh |
652s | because like the whole sort of premise |
656s | of this trailer was really that like we |
658s | wanted to highlight something really |
660s | stupid which is the toilet flush |
662s | functionality in in the game right we |
664s | wanted to highlight something really |
665s | really dumb and then have like all this |
667s | cool stuff in the background if you know |
670s | satisfactory you will know that like |
673s | you're not really interested in the |
674s | toilet flush you're interested in |
675s | everything else around surrounding this |
678s | solil flesh announcement right uh and |
680s | for people who don't know about |
681s | satisfactory they will just go like what |
684s | the was |
686s | that so like that was really the goal |
688s | with this trailer is like we we had this |
690s | running joke with our community uh and |
693s | you can argue that like yeah but only |
694s | people in your community will get this |
696s | joke but I really thought that like no |
699s | people will see the trailer and be |
700s | confused and if people don't know about |
702s | satisfactory they will be interested in |
704s | learning about more about satisfactory |
705s | and it will be easy to write about this |
707s | because it will make dumb |
709s | headlines uh because it'd be like |
712s | satisfactory is adding toilet flush it's |
715s | like that's easy to write about it will |
716s | make for Sensational news and it will |
718s | draw people in and make people |
719s | interested and everybody will know that |
721s | 1.0 is coming out and all these like |
722s | other implications that come with that |
725s | uh the announcement trailer I in my |
726s | opinion was such a good like instigator |
729s | of the path to come for 1.0 uh because |
732s | it was it had all these elements uh that |
736s | really played into it the other thing |
737s | with this trailer is that we had a |
738s | really short amount of time to make it |
740s | so we had to sort of reallocate |
743s | resources that were really already |
745s | working on a lot of other stuff uh for |
748s | the main trailer and also the game |
749s | itself self so it was it was rough |
753s | making this trailer um we made it I |
756s | basically made this trailer in like |
758s | full-time in two months time uh while |
761s | doing everything else at the same time |
763s | uh don't recommend doing this but |
765s | sometimes there's like opportunities in |
767s | gameevil constantly and like we we |
769s | really like this idea um and it it in |
773s | the end it was really worth it because |
775s | it it really paid off like the the |
777s | amount of Interest we got from this |
778s | trailer uh and it was like the fastest |
781s | YouTube video we've like in terms of |
783s | views we've ever had on the YouTube |
784s | channel so like and I think there were |
786s | over like 150 articles written about |
788s | this announcement trailer so it really |
791s | paid off making this trailer um but I do |
794s | not recommend uh doing it the way we did |
798s | it because holy it was rough but we |
801s | got it done so yeah for us the the |
803s | trailer would convey like the release |
804s | dat for satisfactory it would show new |
806s | content with the joke being that like |
808s | we're highlighting something that's you |
810s | know kind of useless which was the |
811s | toilet flush but everyone would know |
813s | about all the cool stuff behind it um |
815s | you know all the new parts and all the |
817s | new buildings and people would be really |
819s | excited about that you know our |
820s | community is really really good at uh |
823s | you know picking up the pieces and |
824s | figuring things out so we didn't feel |
826s | that we needed to really explain too |
828s | much with this trailer because people |
829s | were going to see you know the new |
831s | buildings and see the new parts and be |
833s | able to piece things together and they |
835s | really did there was so many videos made |
837s | uh from Community Based on that |
840s | trailer uh where people were figuring |
842s | things out and it came in waves too like |
845s | people figured some things out and they |
847s | were like wait a second and then they |
848s | made like follow-up videos on those |
850s | followup videos and it was it was one of |
853s | my favorite times working in this game |
855s | ever seeing just like how people were |
858s | were unpacking this trailer and creating |
860s | all these like conversation pieces and |
861s | it's it sparked so much joy and so much |
864s | uh interest uh in the community for this |
867s | trailer so that was really really cool |
869s | so now that we have an idea we have a |
870s | concept for trailer we start going into |
872s | the pre-production phase of things uh |
875s | and this is sort of where we tried to |
877s | figure out like how are we making this |
878s | trailer so is it an in-house trailer as |
881s | in like we're making this trailer is it |
883s | or is it someone else making this |
884s | trailer for us uh I think every single |
887s | trailer we've made have been inhouse um |
890s | and been created like by the team so to |
892s | speak there been a couple of external |
894s | partners that have helped us out with |
895s | some scenes and they've been super |
897s | helpful so for instance the opening |
898s | shots for the 1.0 trailer was made by an |
900s | external studio and that was really nice |
902s | like offboarding some of that work to |
904s | them um so that we can focus on like |
906s | more the the knitty greedy things that |
908s | only we at cof stain and working on |
910s | satisfactory would know uh for the most |
912s | part lenus is the one who's making most |
914s | of the trailers here at cof stain but |
915s | sometimes I also jump in and create and |
917s | make some of them so like the update 6 |
918s | trailer for instance and and the |
920s | announcement trailer so one thing that's |
922s | really really important for us uh for |
924s | trailers is audio design uh almost every |
927s | trailer that we made have always been |
929s | like audio and music comes first usually |
931s | what we do is we have we set a tone or a |
934s | Vibe with our music and then we sort of |
936s | build a trailer behind that and trying |
938s | to like you know make the trailer fit |
941s | the music so to |
942s | [Music] |
950s | speak but sometimes we have composers |
953s | like Dennis uh whove made a lot of songs |
955s | uh for from ghostwood Empire uh write us |
958s | custom songs and there's there's a bit |
961s | of a trade-off there when you license a |
962s | song you can usually take a song that |
964s | people already know about uh so they |
966s | already have like a connection to that |
968s | uh but the issue with that usually is |
969s | that uh they can be really expensive and |
971s | you can also only use them in like the |
974s | situation that you're you're licensing |
975s | for so like if we license a song we can |
978s | only really use that song for uploading |
980s | a video on our YouTube channel and |
981s | that's it uh whereas when you when we |
984s | have like composers that make something |
985s | custom for us uh we can sort of Make a |
988s | Deal better and like have it so that we |
990s | have more you know lenience in terms of |
993s | like where we can use it and how we can |
994s | use it and stuff like that so for the |
996s | 1.0 trailer for instance uh we |
998s | originally was going to use a license |
1000s | song uh but it was hella expensive and |
1003s | it really limited Us in terms of like |
1005s | where how we could use it and like this |
1007s | the terms of that license meant that we |
1009s | could only really use it on YouTube once |
1011s | and that was it so it was a lot better |
1013s | to just have ghost with Empire right a |
1015s | track and I I'm glad we we went with |
1018s | this route actually because I think that |
1019s | song is a banger and the process there |
1021s | is usually like that we you know send |
1023s | them the composer a couple of tracks or |
1025s | like some references that we have uh |
1028s | just to get the vibe going and all that |
1030s | stuff and they'll create something and |
1031s | then we'll iterate on that stuff uh and |
1034s | for the announcement trailer uh because |
1036s | we were so tight on time making that |
1038s | trailer from when we came up with the |
1040s | concept versus like when we actually |
1041s | needed to finish it um we had we decided |
1044s | to reuse the song that we had for the |
1047s | update 3 trailer um but we did ask |
1050s | Dennis to tweak it a little bit so it |
1052s | would better fit the timing and all that |
1054s | stuff for the uh announcement trailer so |
1057s | it was a little bit tricky to make it |
1059s | work uh but I think we got it done and |
1061s | it also ended up being like being that |
1063s | we had to reuse you know that trailer it |
1066s | was the same thing like we wanted to |
1068s | make a trailer that's dumb and like epic |
1070s | you know so we we we took like the most |
1072s | epic song that we could find and that |
1074s | was really that update 3 trailer song |
1082s | so once we have the song we start |
1084s | sitting down and making storyboards |
1085s | which is are basically uh uh one of our |
1088s | artists sitting down and drawing up the |
1090s | frames showing sort of like what is |
1092s | going to be in in each shot and which |
1095s | shots are going to be in the trailer so |
1097s | when we make storyboards we usually sit |
1098s | and discuss first like this is what we |
1100s | want to entail and then our artist sits |
1102s | down and draws it and then we go through |
1103s | all these shots and discuss like you |
1105s | know does it make sense in this order or |
1107s | should this be in the shot changeed this |
1109s | and this and this um so the goal of this |
1113s | really is that at the end of |
1114s | pre-production we need to know exactly |
1117s | how many shots we need and which shots |
1120s | we are making uh so that we can better |
1122s | estimate like how much time it takes for |
1124s | us to make this you know if we need a a |
1126s | lot of shots with a lot of big factories |
1128s | that's going to take a long time versus |
1130s | you know just close-ups of the pioneer |
1132s | or something like that um and also if |
1135s | like the shots are very intricate that |
1137s | also plays a big part in like how much |
1140s | budget we need to spend to be able to |
1141s | make these trailers and there's also |
1143s | elements like do we need custom |
1144s | animation and and like that so like |
1146s | storyboarding is such a powerful Tool uh |
1149s | to be able to figure all this stuff out |
1150s | so for the the announcement trailer we |
1152s | knew which songs we needed and when we |
1154s | make storyboards we also usually |
1156s | typically try and put them in the |
1157s | timeline in like Premiere or something |
1159s | where we can sort of have it rolling and |
1161s | like see and get timing and stuff like |
1163s | that and since we knew which song we |
1165s | were going to use for this trailer we |
1167s | could just put that song in there and |
1169s | sort of play it out and and see how it |
1171s | goes and like see how much you know each |
1173s | frame can be on screen uh at certain |
1176s | amounts of |
1180s | [Music] |
1184s | time so once we have that storyboard uh |
1187s | trailer we usually try and do a pass |
1189s | where we do like a little bit of a |
1190s | blockout uh in editor where we sort of |
1194s | try and like make a quick mockup of all |
1196s | the shots that we need to make and just |
1198s | Place some basic elements in there just |
1200s | to get a feel for what the framing will |
1202s | be like in editor if there are any |
1204s | limitations in game that we need to take |
1205s | into consideration uh and we can also |
1208s | take the Pioneer model if we have like |
1210s | characters doing stuff and place them in |
1211s | the world and and you know get some |
1213s | poses in here and there and you know see |
1215s | how things flows and then you play these |
1217s | animations in the engine and real engine |
1219s | also has this tool called control rig so |
1221s | what control rig does is it basically |
1222s | takes the rig that you created and it |
1224s | adds a rig on top of that so you can |
1227s | control the rig in |
1230s | engine control rig it makes sense |
1232s | doesn't it this is a really really |
1234s | really powerful tool um from my point of |
1237s | view because I'm not an animator uh but |
1240s | I am able to create animations in engine |
1242s | uh in the scenes and whatnot and this is |
1245s | really really good for um when we're |
1247s | making the blockout because I can just |
1249s | take the Pioneer and I can create the |
1250s | poses in the sequence itself uh really |
1253s | really really quickly in fact I actually |
1256s | made the entire intro sequence in the |
1259s | update 6 trailer with control rig so it |
1261s | is totally possible to make the entire |
1263s | trailer using control rig uh but it does |
1266s | have some drawbacks and whatnot which |
1267s | makes it more suitable for like these |
1269s | kind of basic blockouts things but it is |
1271s | possible to to make a full thing in |
1273s | sequencer uh in a real engine using |
1275s | control rig and yeah I can't tell you |
1278s | how much I love that tool because it |
1279s | makes things so so nice for me as a as |
1283s | like a creating these kind of cinematics |
1285s | in engine and in the storyboard trailer |
1286s | we can also use references to sort of |
1288s | indicate like we're going for with |
1289s | certain scenes when it comes to emotions |
1291s | or or you know movement and whatnot the |
1294s | running sequence in the announcement |
1295s | trailer for instance is heavily based on |
1297s | one uh scene from Sherlock Holmes where |
1299s | they're running through the woods uh we |
1301s | also had a lot of inspiration from Wes |
1303s | Anderson with like the zoom in and all |
1305s | that stuff and and how we compose the |
1307s | shots and whatnot so all whatever |
1310s | information you can convey in these like |
1312s | storyboard trailers in the blockout |
1313s | trailers uh from you as a director to |
1315s | the rest of the team the better so that |
1317s | more people sort of understand your |
1319s | Vision as to what you're creating all |
1320s | right so now that we're done with |
1321s | pre-production you should have a list of |
1323s | all the shots that you need to make uh |
1325s | from there you can sort of pinpoint like |
1327s | who's doing what what do we need to make |
1329s | do we need make make custom animation |
1331s | for this shot uh can we make this in |
1333s | engine do we need something external |
1334s | blah blah blah at this point you should |
1336s | have gotten all your planning done and |
1338s | that's when we head into production and |
1340s | this is where the grind starts uh at |
1342s | least for us from making satisfactory |
1343s | stuff because for our trailers we |
1345s | usually have a lot of factories uh in |
1347s | our shots and making these factories |
1349s | take a really really really really long |
1355s | time H when we make trailers I kid you |
1358s | not I think like 80% of the time uh |
1361s | working on trailers is just building |
1363s | factories um it it there's so much |
1366s | effort that goes into that we also have |
1367s | a tendency to keep like upping ourselves |
1369s | I think in the in the beginning we |
1371s | always made like very simplistic |
1373s | factories and now we have like these |
1376s | Citywide factories uh and it's kind of |
1378s | your fault honestly because we see all |
1381s | these cool builds within the community |
1384s | and we see like wow that looks so cool |
1386s | we need to sort of try and emulate that |
1388s | um and yeah it takes so much time and as |
1392s | an example here for the 1.0 trailer the |
1395s | big Factory like at the end of that |
1397s | trailer that one took Lena's 242 hours |
1400s | to build just for those shots um so a |
1404s | lot of time and effort goes into this uh |
1407s | for those like seconds of of on screen |
1411s | [Music] |
1415s | time now you might be wondering like why |
1417s | don't you just reuse like save files |
1420s | from Community or you know like ask |
1422s | someone else to build a factory and then |
1423s | just you know use that in your trailer |
1426s | um and you can do that sometimes uh and |
1429s | we've done that in the past but for a |
1430s | lot of times when we're making something |
1432s | very specific you kind of want to build |
1433s | the factory around the shot rather than |
1435s | trying to build a shot around a factory |
1438s | if that makes sense like when you're |
1439s | composing these shots and we have a very |
1441s | like specific framing in mind or a very |
1444s | specific composition in mind it's very |
1446s | very very difficult to uh place that in |
1449s | an existing Factory usually what we need |
1451s | to do is we need to build a factory |
1453s | around the shot um and it's really |
1456s | frustrating because it takes a lot of |
1458s | time uh but we do get a lot better |
1461s | results in fact for the announcement |
1462s | trailer there's like four different |
1465s | factories for like the same scene like |
1467s | there's one Factory when then you see |
1469s | like the the establishing shot with the |
1471s | cof stain logo and then when we're |
1474s | closer to the ground that's a different |
1476s | Factory that I rebuilt and then when the |
1478s | Pioneer starts running that's a |
1479s | different Factory and then there's like |
1481s | a different Factory based on like which |
1483s | angle you're looking from |
1485s | like the when we have something very |
1487s | specific in mind we have to sort of |
1489s | build the factories around the shots so |
1491s | let's take a look at the announcment |
1493s | trailer shots and and sort of talk about |
1494s | them as well so here's what the game |
1496s | looks like in engine usually when we |
1498s | build factories for our shots we usually |
1500s | just boot up the game like regular and |
1502s | and you know have like Advanced game |
1504s | settings on and unlimited build mode and |
1505s | stuff like that and just build away uh |
1508s | sometimes it's nice to build an engine |
1510s | but the engine is a lot more expensive |
1513s | to run than just the game in general so |
1515s | we I at least try to avoid that as much |
1517s | as possible uh but sometimes you do like |
1519s | need references like sometimes it's nice |
1521s | for instance this character down here |
1523s | for instance I can pull I can whoops I |
1525s | can move this around and place it |
1527s | wherever I need it to instead of you |
1528s | know being in the game and trying to |
1530s | like place things down and figure things |
1532s | out and figure scale out and all that |
1533s | stuff so when we're working on shots we |
1535s | have we used a tool called the sequencer |
1537s | in Unreal Engine and the sequencer is |
1538s | basically you know being able to create |
1540s | a timeline and be able to have things |
1542s | move in time and whatnot um and this is |
1545s | what that tool looks like in unre engine |
1547s | it's basically like a it looks like any |
1549s | kind of timeline Tool uh where you have |
1551s | like a playhead and whatnot um and you |
1554s | can't see all the gizmos and whatnot uh |
1558s | when we're in play mode um so like |
1561s | usually you can see like where the |
1562s | camera is in 3D space and whatnot uh so |
1564s | you'll just have to imagine that um but |
1567s | basically I can like create the camera |
1569s | here you know and I can like uh see like |
1573s | what it looks like when I'm playing and |
1575s | moving it around um right now there's a |
1577s | lot of optimization running um so for |
1580s | instance the building lags a lot as you |
1582s | may notice but yeah here I can basically |
1584s | you know get an overlay of what the shot |
1586s | will look like and you know move the |
1588s | camera around and place it in the world |
1590s | I can even like move around and it will |
1592s | save this so like this will be a really |
1594s | strange animation here so like whoa look |
1597s | at that you know stuff like that so this |
1600s | is the tool that we use when we create |
1603s | all the shots for our trailers it is |
1605s | very nice to use a sequencer because |
1606s | it's very deterministic using the |
1608s | sequencer we can really like play in |
1610s | sequence and create these kind of shots |
1612s | uh sometimes we record straight from |
1614s | gameplay so like we'll play the game and |
1616s | record footage and that's really useful |
1618s | when we have like multiplayer sessions |
1620s | where we like want to showcase like |
1621s | first person shots of things happening |
1623s | and very specific things going on that |
1625s | are a little bit more complex to be able |
1627s | to create in the sequence ER because in |
1628s | the sequence you have to like manually |
1629s | create all these things uh for instance |
1631s | in the opening shot for the update 6 |
1633s | trailer you know the car coming in and |
1635s | stopping I had to I had to manually |
1637s | animate that sometimes it's nice to just |
1639s | like play the game and just have it come |
1641s | in and break but when you do that it's a |
1643s | little bit more tricky to think get |
1644s | things timed properly so working in the |
1646s | sequencer is really like the preferred |
1648s | way of doing things because when you're |
1649s | work in the sequencer you know that you |
1651s | get the shot um otherwise we'd have to |
1654s | like get recording sessions running and |
1656s | if you have multiplayer you need |
1657s | multiple people so obviously that a |
1659s | little bit tricky we actually have an |
1661s | example of that uh with the 1.0 trailer |
1663s | when we made the shot for the um trucks |
1667s | uh coming into frame oh that looks good |
1670s | oh this was oh this was the take guys |
1672s | this was the take this was the best take |
1674s | we've done holy so yeah here here |
1677s | you have it here's the scene uh we can |
1679s | actually fly around a bit you may not |
1680s | have noticed that the logo for coffee |
1682s | scene is an actual physical thing in the |
1685s | world uh we decided to do this instead |
1687s | of superimposing it imp post because we |
1689s | felt like this saved a lot of time we |
1692s | also had more it was easier to plac it |
1695s | and it's also stuff like this where like |
1696s | the Manta for instance over here like |
1698s | that's also animated so I can control |
1701s | exactly where it is in the frame and |
1703s | whatnot and same thing with like all the |
1704s | planets and all that stuff I can |
1706s | position them perfectly in the world you |
1707s | can see that like I have this planet |
1709s | over here this is actually scripted so I |
1712s | can actually like move this around so |
1714s | that it's perfectly inframe so like here |
1717s | I I set the actual like position of it |
1720s | and whatever like how it's rotated in |
1722s | the screen and where it is and and |
1725s | whatnot so like all these things are |
1727s | very meticulously planned out for the |
1729s | sequencer uh to get the shot that we |
1730s | want and the main thing I want to draw |
1732s | your attention to here is like the |
1734s | factory looks like this for this opening |
1736s | shot and in the next shot it looks |
1738s | completely different like suddenly |
1739s | there's whole this thing here conation |
1742s | wise it doesn't make much sense um but |
1745s | the shot is like really nice uh you know |
1749s | and uh yeah it's really important to for |
1751s | us to like get the composition down |
1754s | because like in the Heat of the Moment |
1755s | you're not going to like see these two |
1757s | shots and be like actually the factory |
1758s | doesn't look exactly the same I mean you |
1760s | will uh but that's not really important |
1763s | the important thing is that the shots |
1765s | tell what they're supposed to tell uh |
1768s | and in this case it's the the Pioneer |
1770s | going oh my little tummy um and |
1773s | then we have this closeup of the pioneer |
1775s | going like that and then we switch back |
1777s | and the Pioneer like looks around and |
1779s | obviously we have the shot where it's |
1781s | does the zoom in maybe I'll load that in |
1785s | and then the Pioneer runs away and like |
1787s | here's the block out of that shot where |
1788s | we just get like poses and whatever and |
1790s | try to figure out and and you know |
1792s | they're kind of similar like they have |
1793s | the they don't have the exact same |
1795s | composition um but you know you get the |
1798s | point and especially this run |
1800s | cycling perfect and there's a lot of |
1802s | stuff going on as well like for instance |
1804s | there's we've added a bunch of Lights |
1806s | here to like make the background look |
1808s | nice and like this purple light here |
1810s | which like comes from the I already |
1813s | forgot what this product is called but |
1815s | like we try to Spruce it up a little bit |
1817s | and and you know add a little couple of |
1819s | Lights here and there to to make it uh |
1822s | look nicer and you know add a little bit |
1824s | of uh depth of field and that kind of |
1827s | stuff um to make it look nice once we |
1830s | finished all these shots and we have |
1831s | them ready uh what we do is we just plug |
1834s | them into this like kind of batch |
1837s | program where we then uh run them and |
1840s | render them so we take advantage of the |
1842s | Unreal Engine render pipeline which |
1844s | means that it will sort of like boot up |
1846s | the game and then it will render out the |
1848s | sequence now the the cool thing with the |
1849s | render pipeline is that it is possible |
1851s | to make sure that the game you know you |
1853s | always get perfect like 60 FPS uh this |
1856s | trailer was made in 24 FPS because 24 is |
1860s | usually like a little bit more gets you |
1862s | a little bit more like cinematic feel |
1863s | rather than like gameplay feel but the |
1866s | uh 1.0 trailer was made in 60 FPS uh |
1869s | because we wanted that to like actually |
1870s | look like the game and and play sort of |
1872s | like the game in a way it wasn't as |
1874s | important that it was cinematic as much |
1876s | as you know showcasing the game itself |
1879s | and the thing with the render pipeline |
1880s | as well is that like you know everything |
1882s | that we do for our trailers they're all |
1884s | ingame like that we're using in-game |
1886s | assets uh it's not like a it is is |
1888s | pre-rendered but we're not like adding |
1890s | things that you wouldn't be able to have |
1892s | yourself in the game now granted we are |
1895s | you know using the render pipeline to |
1897s | build to have like no laoding for |
1899s | instance so like everything looks crisp |
1900s | in the background making sure that all |
1902s | animations run perfectly you know we |
1904s | might have humongous factories that are |
1907s | going to run perfectly smooth in 60 FPS |
1910s | when you render but maybe not on like |
1912s | the same machine and this is a little |
1914s | bit of like a |
1915s | tricky balancing act that we have to |
1917s | play with constantly because we want the |
1920s | game we want our trailers to sort of |
1921s | reflect what you're going to get when |
1923s | you play the game uh but you obviously |
1925s | won't have like this all these depth of |
1927s | field and all all the Cinematic stuff uh |
1930s | so it's a yeah it's a it's a line we |
1932s | have to draw somewhere you know as to |
1934s | how detailed we can make it look but |
1936s | theoretically it is possible to make the |
1938s | game look exactly like they do in the |
1940s | trailers but you know realistically most |
1943s | people probably won't have you know as |
1945s | nice looking uh you know settings for |
1949s | all this stuff and have it run this |
1950s | smooth there's an argument to say like |
1952s | oh this is a little bit deceiving and I |
1954s | totally get that um this is something |
1956s | that we constantly have to think about |
1958s | especially also when we're doing like |
1959s | post production and doing color grading |
1961s | uh we also try to you know we want to do |
1963s | a little bit of color grading but |
1964s | sometimes it looks too off and then |
1966s | people notice that and you know yeah it |
1969s | it's you need to like look at our |
1971s | trailers and be like yeah that is the |
1972s | game mostly we don't want to deceive |
1974s | anyone but we do want to make it look as |
1976s | nice as it can possibly be and like like |
1978s | I mentioned before like all these |
1979s | animations that I'm playing right now um |
1981s | they've been created in Maya by our |
1983s | animator and then I created like the |
1985s | mockup animation uh in control rig but |
1988s | at the end of the day it's our animator |
1989s | sitting with Maya and then importing it |
1991s | into Unreal Engine uh to make it fit and |
1993s | whatnot and then we have like this shot |
1995s | with the Pioneer running and all the |
1996s | stuff in the foreg and whatnot um and |
2000s | yeah this there's a lot going on in this |
2002s | shot um for this I've added like |
2005s | particles in the background so like when |
2007s | the Pioneer jumps over over the the |
2009s | thing for instance you can see the |
2010s | particles in the background fly off |
2012s | they're they look really nice with the |
2014s | bokeh here the depth of feel and it's |
2017s | there's like slow motion going on um and |
2020s | then like this this get like the |
2022s | particles from the hand playing here so |
2025s | you can you see the smoke here going |
2028s | off her hand is on fire one really nice |
2032s | thing with the sequencer is that it's |
2033s | really easy for me to compose with all |
2036s | the elements that we have in the game |
2038s | and sort of put them together so for |
2039s | instance uh this shot where the Pioneer |
2041s | is running um like the regular Sprints |
2045s | animation here is is just the Running |
2047s | Animation from the game that I've just |
2048s | sped up a little bit it's been sped up |
2050s | by like 15% basically and then when she |
2054s | jumps that's like a custom animation uh |
2058s | that was made with just the jump in and |
2060s | then I just like tweak it and time it |
2062s | into the Run |
2064s | sequence um in the sequencer as you can |
2067s | see so you can see see here like right |
2069s | how it weighs into um you know and I can |
2072s | I can move this around so like if I |
2074s | wanted the Pioneer to jump earlier I |
2076s | could move this over here and you can |
2078s | see her like Fading Into the jump here |
2081s | and completely missing the mark so so |
2084s | it's really easy to like tweak these |
2087s | things uh in engine and like you know |
2089s | pin pinpoint animations and whatnot um |
2093s | and the same thing here like she's |
2094s | running and then we have a separate |
2096s | slide animation here so B Bally just |
2098s | time take like all these different |
2100s | animations and and just like take what I |
2102s | need and put them in the timeline and |
2104s | time them up and and just pick and |
2106s | choose what I need for the animation in |
2108s | the whole sequence when you're playing |
2110s | the game in general we do have a bunch |
2112s | of effects here like going off when you |
2114s | know it's building stuff but we needed a |
2116s | bit more we needed to juice it up a |
2118s | little bit more for the Traer so I added |
2120s | a ton of like these Sparks effects um |
2123s | that I made myself custom by the way uh |
2126s | to like better high and like different |
2128s | angle so it shows up better in the shot |
2131s | uh you you'd be surprised how much you |
2132s | need to cheat to make it look good and |
2134s | like for this shot here um I've added |
2137s | like VFX here to the right so you see |
2138s | like the little Sparks um there's also a |
2141s | ton of lighting change here so there's a |
2144s | ton of like lights that turn on as she |
2146s | gets closer to it uh and like there's a |
2150s | spotlight on the Hub that also like |
2152s | emphasizes the exposure so I turn down |
2154s | the exposure a bit so that like the Hub |
2157s | stands out a bit more |
2158s | um I add like more chromatic aberration |
2160s | as she gets closer to the hub like to |
2162s | really like drive that feeling of like |
2164s | that dread like oh I need to go I'm |
2166s | about to my pants you know |
2169s | um just like drive that further in and |
2172s | also doing this like Hitchcock effect uh |
2175s | where I am panning the camera back but |
2177s | I'm zooming in and that gives this like |
2179s | weird feeling where like things in the |
2182s | background becomes smaller so you see |
2184s | like these refineries in the background |
2186s | um and it's sort of a similar situation |
2188s | there like I had to build these |
2190s | refineries in such a specific location |
2192s | so that they would show up perfectly in |
2194s | frame uh so I I I legit had to set up |
2197s | this composition and then like build and |
2200s | move these factories around to make sure |
2202s | that when you get this Hitchcock effect |
2205s | that like they would move properly and |
2207s | and look weird and all that stuff um so |
2210s | so yeah this this shot there's a lot of |
2212s | stuff going on in this shot and then |
2214s | like as she runs into the the frame we |
2217s | have this other pioneer reacted to it |
2218s | dead p uh fun fact this like running |
2222s | thing here that's from the blockout we |
2226s | didn't change that at all we F we |
2228s | thought it was funny if like the turn |
2230s | and this all looked so dumb so we just |
2233s | kept it uh it kind of plays into the |
2236s | whole goofiness of it uh because like |
2238s | really like the whole trailer is so like |
2241s | dramatic up until this shot and like |
2243s | this shot is just like this is where you |
2245s | know the punch line is so it's fine if |
2247s | it looks weird and Goofy and and |
2249s | whatever and like really when we're |
2250s | making all these shots like we're just |
2252s | going back and forth constantly uh with |
2254s | everyone in the in the in the trailer |
2256s | Task Force Team uh you know looking at |
2260s | the shots you know having artists like |
2262s | draw on it on the frame and be like it |
2264s | would be nice if you move this uh the |
2267s | composition is a bit weird uh we're just |
2269s | like constantly iterating on these shots |
2272s | not only animation but also composition |
2274s | and lighting and you know all these |
2276s | effects and just like constantly render |
2278s | out shots and updating like our little |
2281s | blockout video and eventually you know |
2284s | our little temp video that we made with |
2286s | storyboards are replaced with you know |
2288s | shots that are in work work in progress |
2291s | and then eventually they're replaced by |
2292s | new shots and eventually you have a |
2294s | trailer and just like keep it iterating |
2297s | all these things um and eventually you |
2300s | you get to the point where you just have |
2301s | a trailer and and you know make |
2303s | adjustments to make sure that the |
2305s | lighting works for the shot so for |
2306s | instance here in this shot I have a |
2307s | little B of A Rim light on the Pioneer |
2309s | to just make her stand out a little bit |
2311s | better against the background uh and |
2313s | stuff like that and and we just go |
2315s | through scene by scene and add all these |
2317s | like little tweaks here and there to |
2318s | make sure that everything works and then |
2320s | once we have all these shots set up in |
2323s | sequencer we just put it in the render |
2325s | queue and then we set uh start the |
2328s | render and when we start the render it |
2329s | will sort of uh apply all these like Max |
2332s | settings to the render to make sure that |
2333s | it looks neat um and like it will do its |
2337s | own anti-aliasing pass so it kind of |
2340s | upscales the image and then uh renders |
2343s | it at a higher resolution and then down |
2344s | scales it and stuff like that so we get |
2346s | like maximum quality on the render and |
2349s | this takes absolutely forever and it's |
2351s | really stressful because when you're on |
2353s | a tight deadline and you constantly have |
2355s | to wait for renders uh it it it's really |
2359s | stressful you know running the game at |
2361s | these like high settings and like |
2363s | loading in so many things is quite |
2365s | taxing on the system and as an example |
2369s | uh the last shot for the 1.0 trailer |
2371s | took 21 hours to render and it failed |
2374s | twice during that render when we were |
2376s | doing that um and it was really clutch |
2380s | the final render for that shot was made |
2383s | it was finished on the same day as it |
2385s | released um 1.0 so |
2390s | uh yeah it's oh boy clutch so for a lot |
2395s | of stuff that we make for the trailers |
2397s | we usually just use Onre engines tools |
2399s | uh We've also made some custom tools for |
2401s | our trailer specifically so uh for |
2403s | instance in the 1.0 trailer we had this |
2405s | like time lapse running we actually made |
2407s | a custom time-lapse Tool uh to be able |
2409s | to like spawn factories and have them |
2412s | kind of play out as if they were made in |
2415s | and in the game and we were just |
2417s | recording it but actually those were |
2419s | made in the sequencer and they were |
2420s | created in engine and we also had a |
2422s | similar tool where we were able to like |
2424s | record ourselves running around in first |
2427s | person and then uh importing that into |
2429s | the sequencer and have it play out as if |
2431s | we were playing to make it you know seem |
2433s | natural and seem like actual gameplay um |
2437s | but in in truth it was all made in |
2439s | sequenc so after all that grind making |
2441s | all these shots in editor and creating |
2443s | all the sequences and rendering them and |
2445s | whatnot uh we take all these clips and |
2447s | we pull them into Premiere and we start |
2449s | editing and we actually uh add a little |
2452s | bit of leeway in every clip that we make |
2453s | so that we have some editing power uh |
2456s | because sometimes a lot of things can |
2458s | actually change in the edited and uh |
2461s | sometimes we realize that we need |
2462s | something else or another shot um once |
2465s | we start editing it so it's nice to like |
2467s | edit as you go uh in fact like we almost |
2470s | have like a an edit ready before we even |
2473s | have all these shots rendered um just so |
2476s | we know like timing wise what we want to |
2478s | do and and whatnot so we edit it up and |
2481s | tighten it up and once we get to the |
2482s | point where we can uh freeze the edit as |
2485s | in like say like okay we're not touching |
2488s | the edit anymore this is when we send it |
2490s | off to the audio people so that they can |
2492s | add audio design um and you know tweak |
2495s | up anything that may be needed with |
2497s | music and and whatnot and let me tell |
2500s | you this uh when you've made the edit |
2504s | and you send it over to the audio people |
2506s | when you get the audio back that's when |
2508s | it starts feeling like an like a trailer |
2510s | like audio design adds so much to the |
2512s | trailers they're so uh underutilized in |
2516s | a lot of stuff I would would say uh |
2518s | audio is so important in anything cuz |
2521s | without audio our trailers are kind of |
2523s | lame I would say like it really needs |
2526s | the music and the sounds and everything |
2527s | to make it all come together um so yeah |
2530s | it it's just it's such a gratifying part |
2533s | of the process when you've made all the |
2535s | stuff and you send it off to them and |
2537s | then you get audio passes back and it's |
2539s | like H now it feels like a trailer and |
2542s | then we have a little bit of an |
2543s | iteration there like tweaking up |
2544s | anything we want to change or any audio |
2546s | effects that we want to add sometimes we |
2548s | also do a little bit of a color grade at |
2549s | this point where we you know maybe tweak |
2551s | some saturation values in here and there |
2553s | or do something uh else that is needed |
2557s | like a post-processing effect or |
2558s | something but for the most part we try |
2559s | not to do too much because like I said |
2561s | before um we typically want the trailers |
2565s | to sort of look like the game so we |
2567s | don't want to you know change too much I |
2569s | I found sometimes when we made two harsh |
2571s | color grades in the past people have |
2572s | commented like oh the game looks |
2574s | different now and we don't want that we |
2576s | don't we want the trailer to reflect |
2577s | what the game looks like so that you |
2579s | know you know what you're getting and |
2580s | last but not least we add these like |
2581s | title cards or end cards uh that |
2583s | actually have a lot of effort put into |
2585s | them as well uh almost every update |
2587s | we've made have had some kind of like |
2589s | really neat like design to the title of |
2591s | each update to make them look good and |
2593s | they also animate well and and that kind |
2595s | of stuff and for the toilet flush we |
2596s | wanted to make this like dumb D toilet |
2599s | flush letters flowing in fly like |
2602s | letters flying in and crushing the thing |
2604s | and like cool sound effects and whatever |
2606s | and timing to the music music and uh |
2609s | yeah really neat and the titles and like |
2611s | all this like cool graphics they're |
2612s | usually made in After Effects by uh one |
2615s | of our UI artists and that is the entire |
2617s | process of how we create trailers here |
2619s | at Cain all the way from ideation up |
2622s | until the point where it reaches your |
2624s | screen so thank you so much for sitting |
2626s | down and watching this video and I also |
2629s | want to mention that this is actually |
2630s | going to be the last video that I make |
2633s | here on this channel so I want to say a |
2636s | big thank you to everyone watching |
2637s | watching my videos over the years on |
2639s | this channel and uh until I see you all |
2642s | again take |
2644s | [Music] |
2651s | care all right let's let's get out of |
2653s | here |
2661s | [Music] |