Session: From Sounds of Chilling Winter to Voices in Outer Space - Audio Behind Frostpunk 2 & …

[Music] I would like to invite the next person on stage which is um GR melac for voices and outter space audio behind the altars hey guys well Simon delivered such an amazing talk that I think I will have to borrow him to our office 11 B Studios so that our budgets for audio will grow uh hopefully so today I will talk about the AIO for the Alters uh little bit about me I'm M which is a typical polish name so that's

going to be easier if you call me simply Greg uh I've been working in a game audio for about 10 years working for a vast the spectrum of different projects from very small Indie projects and mobile projects all the way up to the Tria uh I had an opportunity to work on um Sniper Ghost Warrior series on uh hard West two series or and then I also worked for CD project red uh working on the quent the throne breaker and cyberp punk 2077 uh so let's talk about the Alters what is Alters Alters is an upcoming sci-fi uh narrative sci-fi strategy not really strategy but more of a management game with an element of survival the game is about our protagonists yeah Oli who is um who is

stranded in space because of the certain events that happened he's stred in space and he has to work the way out to survive there on a on a hostile planet and the way to survive for him is to alternate himself and Alternate is to clone himself in a way that he clones him with clones himself um making alternate copies of himself when we were designing the game we were thinking about one question that each and every one of asks which is what if and that what if question is what if we choose this over that what if we didn't leave our H Hometown or what if something else happened and with that underlying concept of what if what happen the the the game concept grew what if you can meet your alternative alternative versions of yourself from

different alternate life path uh would they have different experiences would they have different skills different abilities and how would it would be if you meet them and would they be the same person or a different person and that concept grew into into our game in a sense that uh our protagonist has to clone himself to survive on this hostile planet and use the skills of his different different alternate life paths to to overcome the challenges so I can talk a lot about the Alters but uh let's see a short clip [Music] [Music]

that gives you much more of a overview what the game is about so let's introduce our very small uh audio team um we have uh our sound designer Philip stanowski whove been working previously on movies uh we have our music composer P mus you may know him from frost Punk series and from Witcher 3 blood and wine DLC and there's also me as you can see we do really have a very small audio team at uh at that project and the project is obviously growing in scope as every project is growing in scope and with such a small team we had to be as efficient as possible and we we

we have to use the resources wisely we don't have well wisely that's a keyword uh we have to use those resources wisely so that we we deliver the the the vision that we had for audio of a game and that is very important especially here with so many developers from Indie projects from uh from mobile uh mobile projects it's very important to know that in in that case wise for us was something amazing tool that really helped us to realize our vision um in a sense you for example see we don't have um audio coder we don't have Pro Audio programmer in our team we do borrow sometimes programmers from gameplay team or from other um code teams but we don't have our own pure technical audio coder which makes things difficult and that is why we from day one decided that we need a robust audio uh solution that will

help us to realize our vision and that's why how why we went with the wise so what were our audio goals for um for the Alters how do we wanted to approach the audio in that project the main idea was that in our game we have two vastly different environments we have planet and we have BAS the player is constantly jumping between planet and base and doing different things and starting with base we want something to we want the base to be a very comfy space a safe space for our character a place where you are feeling like a home a place where you can have activities with your um with your Alters with your clones a place where you have a rather chill Music and on the other

hand we wanted planet to be something harsh dangerous something uh other worldly ominous and something that makes you you don't want to go get outside of our your safe base and go to the planet but you have to to overcome the challenges that you meet on your way also in a base um also in a base we wanted the audio for the base to grow uh with you as you grow your crew with new altars you clone yourself you get new crew members and we wanted the base to grow with every new crew member and we wanted the base as a whole to grow because you're building new modules and new elements to the base uh on the other hand the planet is as I said something dangerous something otherworldly we also wanted on a planet to maintain the feeling that you're not sure whether the sound of the

sound of the planet is music or is it sound design wanted to blend blend those two worlds uh on a planet so that it feels Lively and different uh on a planet so let's see how we how we realize that Vision first let's talk about the Bas and the base is as you see a huge greed Donuts uh and in circular circular base with grid system and that grid system allowed us for example for use of shared reverbs we started with a Bas because this is the most important place for for player in Bas there's a lot of things happening there's in Bas there are mostly mostly all the dialogues all the emotions that happened and a lot of narrative is taking place in a base so we wanted to make it home like fing of

the base and we decided to use shared reverbs for our um for the GD system of Bas we fortunate have the GD fortunately have the GD uh on a base and decided to sh to create the reverbs for each of those modules based on how much space they take in the grid system that's why we have uh the verb for Reverb for to U for U which is all un like two units four units six units eight units and and big storage which is the largest module uh we also wanted those modules to be something that you feel like you're in a room that you hear uh for example bleeding of a sound from different from different rooms from different uh uh adjacent modules that you have we did use spatial on that also we did use the rooms and portals system from one

on um on the on the grid system of the base although we had to uh we had to use our own system of um cutting of the rooms because you can freely move those modules you can make any sort of arrangement of that and you can expand the base making it Al all hard on performance so that when we have a lot of rooms and you can freely move them around on the grid we had to call all of them cut cut some of them so that the performance wouldn't be uh the performance would be better and it wouldn't be so taxing on the performance another element of the base is is the lifein based system lifein based system is a system to populate the uh populate the uh base with things that are happened there with activities with elements that alters are uh doing in our our in our base and how it works this is

a lot of there's a lot of Animation there and with our very limited resources of a small team uh small AO team we couldn't deliver all the animations that we wanted we we we had the trouble to delivering all the animations that animation team provided and we couldn't put a sound to all of them so what we've done is we decided to use to create sort of a a modular system for a Foy so that we create uh different events that represent different type of fly movement and then we dropped it to our designers H we gave them all the names of the events and say hey use those events uh obviously not not all of them were n of them were perfect and we cannot expect from designers to to uh to hear as much of detail as sound designers do but still it in most cases it was fair enough and it only required

a little bit of tweaking from our side which helped us immensely to make it to to to deliver such aot such amount of animations emerging events in our base uh depending on what is happening in a base it uh often some events are happening such as uh food poisoning of the Alters because something terrible happened in in in kitchen module or some are fighting because they have different opinion and with that we approach it similarly in a way that uh we also try to use as much of reused modular elements of of folly sounds and uh Folly events next in base is also a use of music to convey the concept of the to convey the concept of Base being safe

and based being uh a place where most of the emotions and and are and dialogues are happening we used the underscore music system for for all the our dialogues how we did that is we we created U four different main element four different main uh emotions that can be supported by music and those are four different tracks of underscore for dialogue uh this way we work with narration team and ask them hey if you want to certain dialogue to have more of a emotion for a certain aspect of a certain aspect then you can simply call a state from Wise and all the rest is done all the rest is set up on in a wise and all they had to do is just called the called a certain state so it's a state based system for uh for the underscore during the dialogue

underscore music during the dialogue we also on a similar on a similar approach we also had the bad aftertaste music what we call uh it's a situation where base which is normally safe uh and safe and and uh and something uh something comfy for player it's changing because something happened for example the Basin is in a bad condition or the altered Smo their Morales are low and when that happens we use the bad aftertaste music which is very similar to our normal Ambiance track but it's bit of Out Of Tune uh and we play it as a stinger that is ducking the rest of the the rest of the uh Ambience music in the base voiceovers I I I connected voiceovers to the base because because most of the voiceovers are happening in the base and

in Alters we do have 16,000 lines for uh for all the Alters it's supposed to be all easy because we basically have just one actor the same guys is playing um playing all the all the characters uh fortunately our voice actor Alex Jordan delivered astonishing performance and we were able to uh distinguish in a tone own all the different characters otherwise having the same voice actor play all the 16,000 lines it's very hard to it's very easy to get lost in in in so many voices uh we use uh offline post processes and also use a lot of wise post processes and that case uh it is very it was very important for us to use those wise post processes mostly because we don't have resources uh to to to to

do it all offline and we're having such a small team again we had to be efficient with our tools with our resources that we have tsv import system we what we done is we're thinking about how we can work with all those lines and how we can uh use all those lines to how we can connect all those lines to uh to our game engine and how we can Implement them and we were trying uh two different approaches one was obviously external sources we tried external sources uh on our project but there was one major problem for us for our team we don't have audio coder we don't have audio programmer and without audio programmer uh having uh external sources it's external Source tool is is a is a is a best way for such a large amount of

voiceovers where you just have folders and you drop the correct naming convention you drop it there and it works with connects directly with uh with events but in our cases we couldn't support it constantly not having the the audio code but we do have tsv Import in Wise and what we decided is that tsv import is much more in a sense flexible for us because we could simply uh designers or sound designers could simply create a spreadsheet formulas right creating spad formulas could then easily uh translate to the tsv system and we could drop all the all the needed voice lines to to to wise that's why we decided that tsv will be flexible enough for us and and wouldn't uh wouldn't tax our other code

departments that we that we would have to request when implementing external sources planet planet is a second part of our game where where players can explore the planets uh and explore its its hostile environment and the Bas and the most important element of the planet is the ridium ridium is uh in a sense a a new particle a fictional particle that stimulates and accelerates the growth which allows on one hand for our protagonist to clone himself so quickly and create so many copies and on the other hand it's bit of a holy grail and a corporation that sent our protagonist in the space really needs it and it's a precious resource

uh and for apidium we really wanted something else and we wanted to get back to the concept of having blend between sound design and music and from a from a from this concept we created uh a music that is partially partially sound design and partially partially just the music but we used the distance parameter to the to our ridium to define whether how it how it sounds whatever you are further from the resource uh it's more of a uh sound design when you get closer to it it's more of a music so we basically created more or less a little bit of a blend container within uh Within the Music uh uh music interactive music um part of the wise and thanks to that we

could easily uh use it as a partially sound design partially music what we also done is we using the distance to to distance to ridum as a parameter we had to somehow work on the music that is constantly playing on the planet already we do have ambiance music on the planet already and when you get closer to this to this particle of ridium you should hear this oddly partially music partially sound design and what happens is it was clashing a lot with our normal Ambience music of a planet what we've done is we we divided both of ours we divided our uh Ambience music to two parts to tonal and to main main is basically bit of an underscore really based in in bass base wi underscore and tonel is a tonal part

of the of the music uh having to having those divide into two elements were're able to using the parameter of distance we were able to either for example cut the tunnel part we were cutting the tunnel part almost fully with the uh voice volume and also we were putting a low pass filter on the uh tonal part on the other hand on a main on a main part which was much more Bassy and white we were only doing uh we were only doing the low pass volume because otherwise when we tried to do it in a way that uh we were cutting also the voice the the issue was that we were losing the the size of a planet we're losing the the the base Ambience of the planet as a music that's why we kept the kept the low base and only cut the tunnel apart uh let's hear

how it [Music] sounds fl's getting closer so there was clear blank and from from music that was on the planet was Ambience music to our ridium which is uh the strange other worldly uh element with partially music partially sound design we also use this parameter of distance trapesium to uh to modulate the effects on the uh environment elements here is an example of just the distance Trum changing the dynamic effect on on I think it's a flinger I think it's a

flinger yes on a flinger of the environment elements and we could use it on on many elements of the environment so that we had the feeling that ridium is a world bending world changing something unknown something that is that is uh reacting to us and we the world is reacting to it spatial on a planet we did use a lot lot of spatial on a uh on our on our levels on a planet and we we really had to learn a lot about the spatial when doing when doing that project well we when we started the project we were really extremely detailed in spatial extremely detailed in in rooms and portals and further we go we found out that often simple rooms and simple portals work as well as good enough as complex and that about the complex really require a lot of performance

issue and a lot of memory in that case or CPU to compile them so in that case the for spatial it was very important to do the right naming convention for all the for the elements and it was very important for example to learn and to start that to start from portals rather than from rooms it's counterintuitive but when you start from finding the the places where rooms end being the portals which is sort of a gate between all the between the rooms it's much easier then to connect connect the rooms rather than creating all the rooms meticulously and then uh and then trying to fit in the the portal so spatial really helped us again a lot we the fact that the the fact that spatial and rumor portal system is Straight Out of the Box in Wise uh made it possible otherwise we would use the old way of trigger box

that would cut the sound getting outside or inside which didn't really work and didn't really fit didn't felt like something modern and if we didn't have the wi we wouldn't be able to to to to approach it this way and another element is uh cutcenes on a planet we do have some cut scenes on the planet and uh we were really struggling in mixing the cuts scenes for the planet uh because we would have to do all the all the cuts scenes uh on the one offline render them and then put them into the sequencer in unreal which requires a lot of a lot of work and we don't have those that team that could that could deliver that so what we've done is we decided to use a lot of elements that we already head in game but um but use them in a special way so

that we can use from the footsteps to seen specific elements in within our sequencer we made a little bit of of a DW from the AL sequencer and and uh what that allowed us is to what we wanted to to to do is we wanted to have a parameters That Could That Could easily for example dark elements that we don't want to hear at certain moments uh and we to do that we we took the voice volume of uh of buses and then we transcribed them to the parameters so that we were able to do plus 12 for neck 20 neck 64 for each of for each of the elements such as environment crucial emitter uh foot steps cut scene specific elements and we had the parameters for all those

elements and thanks to that that we had those parameters we could easily mix within the uh within the sequencer we could do it with State system and we tried it with the state systems so that we get to a certain State um when starting the starting the cut scene but the issue was that cutscenes often uh require gradual change Cuts scene of certain volume and states are much more strict binary when on off right and we wanted something gradual between for example having a wind in a certain moment of cut scene go um go further or or or or be lowered and next to that we were able to do it all in the sequencer rather than mixing it offline and putting one uh event for for the sequencer and that's pretty much it guys uh we

have much more time for for question and answers I guess done done Simon's thought we do have time we do have time so does anybody have any questions at all no okay brilliant well oh there was one sorry hello thank you for the informative presentation uh I was curious about the team size uh I've I've never seen a team size so small to have an audio coordinator so I was wondering like what is the role of each individual and how did you cooperate within this small team uh let's start with our sound designer our designer Philip he takes care of uh we divided certain domains of

uh sound design for each of the person and in his case he take care of all the Folly he take care of scenes and uh and designing a lot of ambiances depending on Act of the game so basically levels then there's there's me with uh I'm as audio coordinators I have to take care of the managing the team and managing all the tasks doing called the jira work which everyone loves uh so managing the team doing all the meetings and all that and on the other hand I'm also a sound designer so I was taking part in most taking care mostly of uh UI sounds Cuts scene sounds uh and then all of the um optimalization for for wise and for unreal we had our composer who only does just compose music and we we want to keep it that way so that uh composer

only takes care of composing because otherwise when we we we tried it at the beginning the way that composer we wanted him to compose very specific segments or Loops or what we what we needed for a game but that didn't really work because he was limited by the concept of technology that's going to put it that that's going to be used further so what uh what we end up is we just let him write the music and then we in a toring tool overwise we try to as much cut it put it into loops and implement it uh and that's also on on on our small two three person audio team so yeah that's that's basically doing everything but that's I I think it's it's great in a sense that you learn so many so many many elements and I think that's also experience of many people working on smaller projects or Indie projects or or on mobile project s when you when you have such a small team and

have to deliver so much and as I said many times in in in my presentation that without such a tool as middleware it would be extremely extremely hard to deliver that project thank you for your question do we have any more questions hi thanks for your talk um this is just a short one I wondered um since it was just the three of you did you all use wise at some point I wonder if you could expand upon I'm especially interested in the the musician uh perspective whether they had an involvement with wise or whether it was mostly just yourself um the the people the people work working with wise were mostly me and the other sound designer in case for a

composer uh he was not implementing music as I said we didn't want to push it onto him to implement also music because that was bit of limiting of his creativity so all of the all of the heavy lifting of implementing and using wise was on uh on sound designers and also we do obviously have some support from from uh code departments but that's the major issue that we that we had with not having dedicated audio coder is that whenever new issue comes up with something that has to be programmed within the in the wise there was different programmer getting that because he had a free time and a free slot for programming and then he she had to they had to learn the wise again and the new programmers were always learning new SDK the documentation or that for wise which uh near the end we decided that it's going to be much better to just keep it for one programmer uh and so that always

he gets the the the same the same person gets the tasks that are related to wise uh so yeah mostly heavy lifting of wise doing was on just two people uh just a specific uh small question about usage of sound libraries when you're designing sounds uh how much do you relay on them and uh do you think this it's kind of limiting or it's expanding your like creativity do you prefer to design from scratch do you use them a lot what's your approach uh in in my case I perceived using using using Sound Lab is absolutely fine we we all know what a huge workload we have when delivering a

game and using sound libraries is is in many ways uh a go-to whether you you have to deliver the content in my case I like to use them as a as a sort of seasoning you you add bit of bit of the sound libraries to make it um to to make as ingredients to make uh sound just to layer it but in fact I'm I'm never really found of using strictly sounds just strictly from the library so that you hear that this is this library because we all know working on sounds that we we we constantly hear the same libraries from uh from certain from certain uh production House of libraries but on the other hand when when I was working on much larger projects uh it really depends what the game is about and in my I in my perception the best way is to obviously use the libraries but for the elements of the game that

are crucial that are most important do the recording do the field recording session we for example when we were working on on cyberpunk we were we had to record and we recorded all the cars there because we found that cars and sound of cars is extremely important so we wanted to record that uh input it using wise for all the we recorded more than 30 60 30d 30 30 something cars and that's why it was important for us and if you're making for example FPS game it's okay if the UI sounds are from library that's that's going to that no one's going to really care that much but if you're doing the FPS maybe it's good to field record uh a guns or or a fol cuz you are moving on a first person so it's best to pick your battles wisely and and use the libraries absolutely where where it it's uh needed and on the

other hand do the recording if you can it's a lot of fun one more one more question uh do you target multi Channel AIO as an end as an end device or it's only stereo no we do Target we do Target suround we do Target suround 5.1 that's our main main goal and we as far as I remember uh I'm not sure if right now it didn't change for PS5 and and and a new Xbox but you have to Target the surrounds for certification of the console on the consoles so down mixim is done automatically then or yes it's not automatically but do you check it yes we do yes at least we try we try to do it and try to try to check it obviously as everyone knows it's hard to find a time for for mixing that's that's the thing that I would love to have much much more

time for with every project to just sit play the game and listen how it works on 20 different systems uh but in real life and real development we all know that it's hard to do it's hard to sometimes even find time for playing the game yet play it on so many different systems but if you have chance do it and and play your game on every possible system um to to hear how it how it works the best results for me my case for example to finding out whether it works good the specific sound is to just play it for some horrible horrible speakers but I guess that many people know that [Music] and [Music]