Featured Blog | This community-written post highlights the best of what the game industry has to offer. Read more like it on the Game Developer Blogs or learn how to Submit Your Own Blog Post
Sounds and music are one of the most important parts of the game atmosphere - as equally important as visuals. However, all sound works are postponed for the very end. And that's completely wrong.
This article will consist of 2 parts, one of which continues topic on sound design, which we started in one of our previous articles - Designing sounds. The other part will shed some details on the sounds in Rise Of Colonies: Uprising game. This time we’ll focus more on the creative process of how to achieve needed results.
I guess I’ll never stop saying this, but sounds and music are one of the most important parts of the game atmosphere - as equally important as visuals. However, all sound works are postponed for the very end. Especially I often see this for the smaller titles (indie, casual, etc.), and then, completed on with the approach: We did, what we could (or had time for). I highly recommend ditching this approach and start working on sounds as early as possible.
As listed in the previous article, there are two paths you can take:
a. Create sounds yourself
b. Outsource them. Soundworks planning depends on this choice.
If sounds are created within a studio:
Short answer: During the game design stage (provided you’re not making another 2048). You must then think about the atmosphere of the future game and pre-consider the soundtrack theme. You need to bring a soundtrack early in the game, so while developing it you constantly keep listening to it to understand if it is annoying or not.
Soundworks should start as soon game avatar is ready to make its first steps, first bullets begin to fly, etc. You’d be surprised, how much in the game should be synced to sounds - explosions, crashes, steps. You’ll be constantly adjusting your Sound Engine/Manager logic towards various playback scenarios.
If sounds are created by 3rd party providers:
Often, the main goal here is to get just a good result within as little iterations as possible, because every iteration costs real money. So, make sure that you have at least the first level ready (the more, the better) with art as close to final as possible, with all particle and visual effects before you present your game to a sound production studio. This is necessary for a sound producer to carefully catch the game’s mood and setting. There’s a chance that you’ll get stone button press sounds instead of metal ones (just an example) if you send just a UI prototype.
Same goes for the music score. For a music composer to come up with real matching soundtrack, he would need to feel the flow of the game, get understanding of characters, what they say, how they move, which events recently have happened or will happen soon to them. This might not be important from the first sight, but believe me, these slightest details find their reflection in tunes, chords, tempo, instruments. For example, the same six-note tune will have different impact when played on sine synthesizer vs saw or impulse one.
As many as possible. However, the game shouldn’t become a sound-like Christmas tree just because you had a lot of time to invest in sounds. This is very a game-specific question for advice, but make sure that each action (shots, steps, door openings, noises, etc.) has at least 3-4 differently sounding versions.
As for soundtracks - make sure that you cover all game main themes: I.e. if you have a simple casual game with menu and more or less the same levels, you’d have to consider making 2 soundtracks - menu theme and game theme. If you have a game that has city, dungeon, desert levels, the number of themes grows up to at least 4: menu, city/village theme, dungeon/creepy theme, desert/Arabic/African theme. Why do I say at least? Well, because if you have bosses in these environments, it’ll be rewarding to provide an additional boss theme for each region.
There are two possible paths: Content path and Software path.
Content path basically means that you have multiple sounds for the same event being played randomly (i.e. gunshots - slightly different sounds for gunshot make it sound like there’s different input from a player with each shot. Sounds for bullets hitting obstacles are also bringing the feel of variety when randomized. Otherwise, they’d be sounding much like a machine gun file.
Software path means that sounds are being varied during playback using post-processing capabilities of the sound engines. Most often, pitch randomization is used to achieve that effect. For example, you can play gunshot sound with varying pitch each time (I find threshold within 0.8... 1.2 intervals work best for most sounds), so it sounds a bit higher or lower.
Though I do not recommend using pitch alone as if a single sound is randomly pitched every time, players will quickly recognize it’s the same sound. Best results are achieved when different content is combined with randomized effects, giving your product a great sense of variety.
So, let’s talk about how you approach sound design from a creative point of view. Music design is not as easy as it sounds and might become a really impossible task if you don’t know how to approach the process.
As discussed earlier, the easiest way is to purchase one of the royalty-free sound effects collections and get your sounds from there. But, you need to be aware - almost 100% of sounds you find there won’t fit your project and will require additional processing. Just something to keep in mind, so you don’t think, those libraries are a magic wand (they are pieces of wood magic wands are made of).
Also, sound libraries tend to contain many thousands of sound files. So, it quickly becomes frustrating when you try to find the one you need (obviously, you can’t listen to all thousands of files). This is worsened by the fact that most of these libraries are poorly organized and alongside the audio you get PDF catalog (yes, a text-based catalog) that you have to refer to. This is so big PIA, so first, try to remember terminology your library uses and use CTRL/CMD + F a lot. But use the shortest words possible. I.e. query “rocket blast explosion loud” will get you 0 results, because keywords search has to be used, so forget your “google-fu” - it doesn’t work here. Good queries example - “gun”, “shot”, “blast”, “kick”. Then, you further refine the search.
Even with advanced search powers, you’ll still be listening through hundreds of files, trying to find “the one”. My advice here is: Do not let this time go to waste. While listening to the sounds, imagine where it could be applied else, even if it completely doesn’t match your game/expectations. What that gives you is that you memorize relations between file name, keywords and effect category. Surprisingly, “blast” and “kick” effects can be used for button presses and bullet hits. This takes a bit of additional time, but it pays outs in the end.
Sound design is a creative work and, depending on the game, sounds might not exist at all. For example, how should sound 4-leg robot walking on the concrete highway? This sound never existed in the world and any library, so what does it mean? Right, that you have to create it. And you can only create it using multiple source sounds. But what those sounds should be? That’s where your talent of imagining the things comes into play, which could invent something as following:
A leg put on the floor - it is a metal leg, it is not hollow, then probably we need to search heavy metal object falling on the ground sound and take part in it.
Then, before the robot starts moving, there should be some activation, like something is clicking, so we search “click” sound and can even use processed “keyboard click”
Now, when the robot moves, there are some gears/bearings working. OK, let’s see if battery-powered drill or screwdriver would work. Surprisingly, it fits just well after some processing.
Now, legs are moving back and forth. OK, let’s play previous sound backward and we’re done.
Now, we just have to align source sounds in a multitrack audio editor and sync them to animation and you’re done. If your robot has complex behavior (like, i.e. it can stop in the middle of the action or walk with different speed), then single sound fx won’t work for you and you’ll have to deal with smaller separate files, timing them to animation parts.
Some notes to keep in mind:
Often “the one” sound lies inside of the bigger SFX (i.e.flight jet pass by sound can be split into the following: ascending sound (close up), blast sound (when jet is nearby), descending sound (as the jet flies away). The moment when the jet is near the lister can be used as part of an explosion or white noise sound.
Actually, most of the sounds are multilayered and are “assembled” from shorter source sounds.
Apart from the creative process, there are some basic technical rules to follow that will ease your life:
Equally, normalize and compress all sound effects to ensure volume and dynamics consistency. Normalization is necessary to give your sounds equal volume. I just normalize sounds to -3dB, but that can be any value around that - it just has to be consistent. I’d recommend not to go higher than 0dB. If you go above, you will most likely get unpleasant clipping. Normalization alone, however, won’t be enough, because, often sounds (especially splash ones - like cracks, hits, explosions, etc.) have loud spikes (i.e. the moment of explosion) while the rest of sound remains relatively subtle. If you use only normalization, the volume with be adjusted by the loudest sample, equally adjusting other samples. You’ll get your normalization level that only spikes sample, while to make other samples of audio effect audible, you’d want to raise other samples’ volume as well. This technique is called sound compression. In short, it makes the sound volume more constant by lowering the volume of loud samples and raising it for quieter ones. There are multiple types of compression plugins available, you can choose one to your liking. I use, standard Hard Limiter plugin, bundled with Adobe Audition and usually, increase volume by 6dB, with threshold -3dB. But, in the case of compression - it really depends, because the sounds are different by nature, so the best judgment should be used here. Compressing sounds will save you a lot of time adjusting the volume in the game and make it sound as a whole.
Try to have as short sound files as possible. Don’t leave silent areas at the beginning and at the end of a sound file. Having silence at the beginning will get the sound play lag, so you’ll have a hard time syncing it in your code. Silence at the end will increase file size, memory consumption and, more importantly, will lock your sound source for longer periods, because silence also plays. And you’ll want to keep your sound sources number to a minimum from a performance perspective.
Master final audio files in the format, that will be used most often. Believe me, there’s no and will never be a difference in gunshot, mastered at 16Bit/48kHz and 24Bit/96kHz. Sounds will be mixed with others during gameplay, players will be concentrated on the gameplay, so it should be there. However, what might get by mastering in 24Bit/96kHz are sample rate conversion artifacts, because mobile devices only play 16Bit/48kHz sounds, so they will be converted by a plugin of an unknown quality during import into the game engine. Bigger doesn’t mean better in this case.
Finally, we have come to the part, where we can discuss a bit of sound created by our game designers. The sounds in the game are divided into three main categories: menu sounds, game sounds, and dialogue sounds.
Total number of SFX is: 92 Game sounds + 19 Menu sounds + 143 dialogue/speech lines (English only). And, also, we have three music scores for this game. This might seem like the massive amount of sounds for the title like this and, indeed, we initially planned that there will be fewer sounds, but when we started to play, it quickly became clear that that amount wasn’t enough.
Typically, the games of such size tend to have 30-40 SFX in their packages. This one has about 3 times bigger amount, but it pays off with player experience.
While the sound design lasted throughout the development process, it took us approximately 14 days of full-day workload to complete sound and music design and get source SFX in our hands.
Let’s not list every single sound we have here, but highlight some sound design decisions made.
These are mostly button clicks and if you can get away with single SFX for a button press, in my opinion, it is not enough. At least it is necessary to have a general button click, denied button click (i.e. when upgrade is not available, it is a different action and should be presented differently), purchase button click (when players spend resources on something, they need to be rewarded and different sounds outline that experience), level start button click (when player engage into battle, they definitely need to be engaged not with regular button click, it should be calling out to action).
The same goes for level complete fanfares - we have 3 of them: Level failed, Level Complete and Boss Level complete. Again, boss level complete fanfare is an additional reward as well as boss music score as the change of mood outlines the importance of events. This technique is often utilized in JRPG titles and I find it highly motivating.
The rest of the menu sounds goes to different popups (like bonus popups), treasures revealing, usage of items and finish screen SFXs while values are popping up and counting.
Well, as discussed in previous “How it’s made” articles, this game has a lot of content and every piece of this content has its sounds: 3 different weapons each have 3-4 shooting sounds + reload start (clip/shells out), process (clip/bullets pick up sounds), stuck (clip/shells hit sounds), complete(clip/shells in) + enemies weapon sounds (shot only) gets us 26 separate sounds. Bosses also have 22 dedicated sounds for shooting, jumping, landing, engine start/stop, walking, etc.)
Then there are environmental, items pickup, abilities, grenades, and other game entities sounds (like bullets barrier hits, barrier kick/punch sounds) brings us to the quantities listed above.
For the most part, all hi-tech and environmental sounds have been constructed from multiple source sounds or fractures of longer sounds. For the game category, there wasn’t a single sound, that was used from the source library as it is. NOT. A. SINGLE. ONE.
The only tip here will be: Invest more than enough time into preparing VO materials for voice-over artists. These can have any form, but at least have at hand:
Character description. Have a doc, which clearly states who the character is, how he speaks, what he does, describe his personal qualities - this will help to get proper intonation quicker. Also, don’t refer to vo artist other games, movies from pop culture, you can’t expect them to be familiar with these references and artists are usually quite busy studying all the media out there.
Scene description - described in detail scene setting, intensity, what happened before the scene, what happens next - this will be used as main guidance by artists.
If there are intonations that are hard to get laid on the paper - record them yourself and provide as a reference, but make sure you don’t have a lot of these files, because strict guidance limits artist’s creativity and that’s why you hire them in the first place, no?
Our game development company has chosen to outsource voice over services for this game. And while preparing for submitting the work, we looked for several options:
Fully managed Voiceover (VO) talents studio. This is the most expensive solution as you get personal manager, which supposedly will search for the appropriate talent that matches your requirements, by interviewing (how it was advertised hundreds of artists). Going this path you’ll have to understand that you’ll still need to prepare all the above docs and have full responsibility for selecting talents. It’s just you’ll get an expensive additional chain that will be bugging you with CVs, examples, asking for docs, saying that results are mediocre, because the docs are incorrect, etc.). This option might be suitable if you need a large number of VO talents for your project, but for several of them, it wasn’t my cup of tea.
Self-managed Voiceover (VO) talents studio. Pretty much the same as above, but 3rd party Manager’s role is limited towards sending you the applicants (who you could as well review yourself as their portfolio is publicly available.
Hire freelance VO talent. This is the cheapest option, but you’ll have to close the deal and communicate with VO talent directly (for me, this is an advantage) and provided that you got your docs right, this might even be better than the first 2 options. You’ll need to review hundreds of talents to find voice/s that suit your case, but you’ll have to somewhat do this anyway, so it's worth considering. Obviously, I'm not naming any platforms as they are easily googled. From the price point of view, in our case we saw the following picture: Option A: 20x, Option B: 12x, Option C: x. I was completely satisfied with Option C and have fully closed VO part of the game within 1 week.
Well, another longread is over, we hope it was useful. Till next time, folks!
Read more about:
Featured BlogsYou May Also Like
