Looking at the most successful games like League of Legends, Elder Scrolls (series), Borderlands (series), Witcher (series), Evolve (or well, it use to be until they messed it up), Overwatch, Halo (series), Left4Dead2, Portal (series), etc, the characters generally have very flavorful and profound dialogue, or sometimes clever and comedic scripts. From what I’ve gathered working and talking with developers, a lot of game designers would like vocals to fit their game to take it to the next level and add a layer of uniqueness with more profound dialogue beyond basic tactical phrases.
The main obstacle with this is the amount of resources it takes to do custom voice acting: you need to buy or rent professional equipment, you need to own/rent a studio to record in, you need a voice actor, and you need an audio designer for post production work. I’ve recently taken a liking to producing vocals, I myself have all the tools necessary for it, but it doesn’t forgo the cost of production, and typically indie developers do not have the budget to hire a designer for $200+ for minimum quotas of only hundreds of words. However, when I found asset stores like this, I realized a good compromise for developers would be to make preset packs of vocals relevant to games, which is what I started doing with this SoundCloud - Hear the world’s sounds and submitted that to the Unity asset store. I still have a more advanced and profound script that includes actual dialogue for robotic characters, but I know there’s more than just scifi games out there.
But to make future products more suited to game designers, my question is: what other kinds of unique phrases and styles of vocals do you want? Could be a demon voice talking about ancient fantasy rituals, could be a ghost seeking vengeance, a cartoon chipmunk acting cute or whatever anyone can think of.
In order for any sort of pack to be useful, it’d have to be really generic, right? I mean, what are the chances that a line of dialog with any sort of specificity – recorded without any knowledge of my game’s existence – will fit?
Well, crap. My game uses warp, not hyperdrive.
But my player character is an Admiral!
Well, geez. I hope I don’t have to hear the same line each time we jump from star to star…
That’s a nightmare to produce, and would provide so much extra stuff people wouldn’t need.
Instead, I could see generic lines that aren’t intended for fully-voiced games to be useful. Generic dialog clips that play when a character’s text bubble pops up or something, that may or may not be what they’re actually saying, if that makes sense.
Totally just my opinion. I, admittedly, have no actual experience working with dialog in my games, though, so take it with a grain of salt.
I can’t imagine buying an asset for something like a VO. Not only does it have to be super generic as mentioned already, simply the nature of it would prevent the kind of editing you can do to 3D model assets, which means everyone who uses the pack is going to have an identical “Affirmative!” vocal in their game.
Additionally, I couldn’t give you any clear stat on this, but I doubt there’s a great percentage of people who want enough voiced dialog that they have these generic barks, but not enough that they don’t VO the whole game (or at least significant portions) and use their own actors’ barks.
Just doesn’t seem likely.
Edit: I also regret clicking your link while forgetting that my sound was on. I hear “Soundstorm Labs” with the background sound of a robot moaning in pain/pleasure…
I actually don’t think it has to be generic at all, as long as there’s enough variety, and I’ve already got some sales for the vocals. I’m talking like at least 400 words per pack typically. There’s two paths to go: the generic core sounds, which are like what you’re describing and what I usually have in the sound cloud link, things that build the skeleton of a character’s script that remain versatile. It’s okay for those types of sounds to be used many times because players expect them. Even in real life, despite that people have all their own unique personalities, everyone still says generic things like “Bye!” “Cya!” or “Thanks!” “Thank you!” “I’m doing good,” “Man…” “I can’t imagine…” “Take it with a grain of salt.” “:)” “It doesn’t seem likely” “I could see…” and etc. From here, a designer can leave it as a simple script where the game is not fully voiced, or they can take it a step further and develop a unique character from it with additional packs that specifically have unique types of personalities.
And then the other path I’m working on now is actually a variety of sophisticated script that will let designers develop the character into whatever personality they want, ranging from someone austere, someone innocent, someone philosophical, someone arrogant, someone psychopathic, someone tactical, someone elementary, etc or anything in between. This is the flavor that characters need to be unique and engaging. However, if there’s anything specific you want, now’s your chance to speak up.
Ultimately the key is just the numbers. If there’s enough VO samples then people are bound to find what they’re looking for, just like with regular sound effects.
I’m also just not very concerned about duplicates immediately because of the sheer number of games on the market, unless the sound pack gets like 100 sales. But if it does, then that’s an indicator to me to make a wider variety of sounds.
Another thing to keep in mind is that the characters from successful games like borderlands and league of legends and many other games are not fully voiced over with a fully developed conversational script like the NPCs in Skyrim, they just have a wide variety of phrases that they say while walking or attacking, the little flavor that’s been added develops into enormous fan bases for each individual character because of that. So far it sounds like what people want is something unique more than generic, which means you’re in luck because niche is exactly what I do.
What I could also do is take a sort of “pre-order,” where I see enough support for a particular kind of character, I can put all the work into developing that script and producing the variety of vocals, then upload it to soundcloud and submit it to the asset store.
I guess I just don’t get it, then. Is the idea that you just have a big ol’ catalog of words, and then I can somehow piece all these words together to make actual dialog? So you give me a recording that has the words:
And then I use that to build sentences like
I have a red apple.
You ate blue apples.
He eats five green apples.
If so, how does that not sound choppier than commentary from Madden '96?
Again, I think I just don’t understand, which is probably fine because I doubt I’m your target audience anyway, being a hobbyist and all.
The idea isn’t to have a catalog of words, it’s to have a library of unique phrases that can be combined to create a unique character, along with the common generic words to fill the gaps. You’ve used a level builder before right? You don’t have to change the textures of each block manually, you don’t have to change the shape of each block manually, but you can efficiently combine the the blocks however you want to make a level. That’s what I’m doing with vocals to make a personality builder.
And that’s cool that they make that, and I hope it has a large variety of voices beyond just an average voice that it can generate. But, people also said we’d be living on the moon by now too. The issue as always is economics. Even when and if they do release it, it’s not going to be cheap, and they still need core voice actors for the work, and it still needs a professional voice to work from. For instance, you “could” just make everything here in the asset store yourself in some way, but it would take a lot of time to learn how to professionally do all of it well and cost a bit of money to buy all the software.
What do you mean by more voices beyond average? Both techniques work on a small sample data to generate any intonation from any words, so it works for any voice. And both are currently worked out in a product with voco being soon to market. The thing is if you can crop clean sample from a movie (or any sources) that might be enough, that’s how advanced both technique currently are.
Of course talent is not given even with infinite voices.
The truth is that I haven’t checked if any is available RIGHT NOW
I’ve worked on vocal software like that, and the effectiveness really depends on what you use it for. There’s definitely really great post-production software that let’s you mess with the wave form similar to photoshop, but it’s mostly for things like cleaning up noise, nothing just lets you make any voice you want from scratch. The thing they use in their demonstration looks good, but what you don’t know is that’s like the best of the best possible outcomes, and even then you can still hear some less than ideal inconsistencies with it.
It’s basically like photoshop’s “content aware,” it’s another tool that has its place for being useful, but it definitely has very clear limitations, you can’t just say “Photoshop, make me a sunset” and have a sunset with clouds and an ocean suddenly appear from scratch. Similarly, you can’t just say "Adobe Audition, make a robotic voice saying “supercalifragilisticexpialidocious.”
And, just like with content aware, it needs a lot of good source material to create anything near a viable interpolation. I think one day it will be possible when AI become more sophisticated, but that day is not today, nor any day in the near future, it probably won’t happen for at least 20 years.
Combat audio is probably the most ubiquitous and generic voice work that can be done. Think the canned responses in an RTS, or the background chatter of other soldiers in a FPS.
And now for my semi rant for all audio asset developers…
Whatever you do, build a good text index for it. I’ve checked out countless audio assets. And the big down side with almost all of them is a lack of an index. I don’t have the time or skill to listen to hundreds or thousands of hours of sound, just to figure out the ones I want to use.
When working with audio people in person, I describe the purpose and feel of the sound, and the audio guy magically gives me what I need. I don’t have to sit there listening to hours of sound to figure out what I want. I want that same experience when I download your asset and try and look through it for sounds.
I will keep that in mind when submitting assets. To be honest I suspected just the opposite, the Unity store told me they want a large audio preview of all audio files like through soundcloud, but if people prefer to simply read the indexed list, I can provide that too, unless Unity tell’s me I can’t. If you have the time I would recommend you voice your issues to Unity itself since they make this listening a required policy for audio designers when submitting to the store.
I see your point about combat audio though, that it is ubiquitous which leaves people wanting something else, but if you think about it, that’s a good sign in some ways. That means it’s so popular and people like it so much that they just can’t enough of it, which also means I should probably make some combat audio just to ensure I am providing a unique take on something that game designers are most likely to need. But it does seem like the general feeling in this thread is to have vocals that are more than just generic phrases, which is completely fine and I agree that game vocals can be more sophisticated, that’s what I was planning on making.
At a certain point though, you don’t want everything to be incredibly unique because it makes the character unrelatable and incoherent. “supercalifragilisticexpialidocious” is certainly something unique to say, but is there really a good context for it in a game? “Hamburger shirts break-dancing contrasts fickle pickles” isn’t something you’ve ever heard and it would definitely attract attention if you put it in a game, but that’s also just nonesense, so you do want some common phrases to fill the gaps that the unique phrases can’t, it helps make things less overdramatized which makes the character and story more believable. I might provide only unique items, but then people just like you would criticize me for just the opposite by saying “Can’t you just have some basic combat sounds? All I want is are simple tactical commands…,” not that there’s anything wrong with that. So it’s a fine balance that needs to be struck between generic and unique.
What I am interested in though too is specifically, what do you want for your characters? What kinds of characters do you like making, and what would you want them to say?
For the Asset Store page, you definitely need a preview in something like SoundCloud. But after customers buy your audio library it’s very helpful to include a text index in the unitypackage so they don’t have to dig through the SoundCloud page every time they want to find an audio clip. And if you’re making a text index anyway, you might as well put it on your website and add a link in the Asset Store description.
I wonder if there’s a market for non-human fantasy voices, like Warcraft’s orcs and trolls (“skobu!”, “dabo!”, “zug zug!”, etc.).
There are some sounds like that scattered across the internet, but random orc noises are even more niche than an actual orc character. Still, it’s good to keep in mind for when I make medieval vocals.
I’m not quite sure I understand that. If they bought the package, wouldn’t they simply have the files on their own computer and browse through them however they want? I was thinking more for on the asset store itself.
I search the asset store once a month for usable voice work and have never seen anything particularly good.
I’ve been desperately looking for some good gibberish voice work that I could at minimum use for placeholder. These would need to be gruff, harsh voices with very high quality gibberish (perhaps based on real language, like hungarian or even stuff that sounds like english but is unintelligible). Existing gibberish is all toon focused and styled after the sims.
Gibberish sounds should be based on things you would expect in games. “He’s over there”, “How much for the sword”, “Help!”, “They have reinforcements!”, etc.
I’ve also spent forever looking for good combat exertion fx and have been constantly disappointed. I own a good number of these and none were really usable. I want something that sounds like a life and death fight, not a guy going “ugh” “hrugh”, or “yeaaw!” but real desperation, passion, fear, rage.
Finally, I’ve been looking for good walla or small talk. Which is entirely impossible to find.
I’ve really spent forever looking for good general purpose voice and have never found anything usable.
Small talk is something I’m working on as well as quips and curses, so I’m glad someone appreciates that in a character. As for the gibberish though, would you be able to expand on what you mean exactly? “They have reinforcements!” doesn’t seem like gibberish to me, I would consider that a tactical phrase.
If you mean a new language, or a consistent structure of words that has never been heard before, that would require actually making a new language, usually by a language expert. The best fantasy writers in the world do this or Hollywood hires language experts for just for that reason.
Your comment about emotions is something I bring up with voice actors too. Sometimes they put the method acting into it, and sometimes they don’t, it’s mostly a dice roll, and I don’t usually have the resources to deal with a lot of retakes. But, I’ll do my best to accomplish what you’re looking for and make it a higher priority.
A thing to keep in mind though is the kinds of characters that would say what you’re looking for. It doesn’t really make sense for good character to say “surrender or die!” and it doesn’t really make sense for an evil arrogant character to say “help!” so context is helpful too. Who is the warrior saying “surrender or die!”? What motivates them to fight or kill people? Where are they from? How did they get to where they are now? These are all questions that need to be anticipated to create a coherent character script.
I felt like Blizzard got a lot of reuse out of those noises by using them in different contexts. “Zug zug” might mean “I’m ready” in one context, and “On my way” or “Following orders” in another.
If the files have long, descriptive names, then yes. But I bought, for example, a sound effect library of hundreds of miscellaneous sounds that are all named something like NA/WE_TH001, short for Nature - Weather - Thunder clip 001. (That isn’t the exact naming scheme; I’m going from memory.) I’m not about to play through hundreds of clips to listen for the right one. Fortunately it includes a huge text file with one filename per line accompanied by keywords, so I can just search the file for “thunder”.
In terms of voice work - other than exertion and stuff - if you want polished results you really need highly custom work (and maybe even in the case of exertion).
So realistically with voice, what I’m looking for is a set of placeholders or stuff I can at least work with.
“They have reinforcements” isn’t as important to me as a set of gibberish that sounds like someone saying “they have reinforcements” if that makes sense. The same kind of delivery, the same kind of feeling, only unintelligible words. Maybe I could use the same effect as a “surprised” sound or a “they’re over there” sound, or other cases that don’t actually involve reinforcement but still involve the same “feel”.
If you look at actual games, especially really polished AAA ones, and you look at the voice work (even generic ‘quips’) its super different from anything currently on the asset store. Listen to NPC commotion in the background of an assassins creed or whatever, all those background voice clips or reaction clips.
These are the kinds of sounds that games need that are fairly generic.
You cannot use “evil arrogant” character lines, because a character that specific needs to be voiced specifically to the game. I need “started peasant quip 01” not “arnold the evil wizard monologue 01”. For Arnold, I need to hire voice talent.
That that definitely works in some circumstances, I mean they completely reused the physics engine in Oblivion and I believe they did that with Skyrim as well. But the kinds of games Blizzard makes are far too big to re-use pre-recorded scripts with the diversity of vocals and characters in their games; they have a team of writers improvising the scripts for every single project. However, it is common for a sound designer at a game company to modify or combine sound effects like explosions from previous projects to make new ones.
I have so far only seen one audio product on the unity store actually successfully show an indexed list of audio files where you can search through everything in organized folders. I am interested in this myself, but I am still pretty new to Unity and this asset store so it will take some figuring out. For now I can simply list all of the files manually, and categorize them in the description too.
You are definitely right that it takes good quality, but the problem isn’t the quality itself so much, I can remove noise and most of the puffs and harsh sounds manually in post production. It’s more of the voice actor’s experience. Advertisements are far more common and versatile than game characters, and consequently the bulk of voice actors respond by specializing in commercial voices, or things like podcasts, news broadcasting and journalism, audio transcriptions and youtube narrations. Because of that, they have less experience and motivation to do game voices well, so it is harder to find voice actors who can dependably handle the authentic emotion desired for game characters.
It seems like there’s maybe a little bit of a language barrier with the terminology we use, but from what I can tell, what you’re saying does exemplify what I said earlier to someone else, which is that game designers also want generic phrases that can fill common and expected roles for things like NPCs to distinguish from phrases that are unique and define a playable character.
So basically, you also want background phrases or things that aren’t terribly important but still are beyond one-liner grunts, things that maybe NPCs say that don’t draw too much focus from the gameplay or character’s script like what was in those videos. Perhaps even phrases that you can combine with other phrases later on for specific situations like, “I am…” and then having combination options of “awaiting orders” “on my way!” “going!” “forsaken…” “unstoppable!” and so on.
