June 10, 2019 at 11:19

Less ears or how not to spoil the sound in the game from the very beginning



    The article is about sound, but it is addressed not to sound designers (who know everything), but to producers, game industry PMs and novice developers. The errors collected here are our own experience from the time when War Robots was still a prototype in a new niche for the company.

    At that time, we wanted to quickly test the idea, which affected not only architecture or code , but also sound. Then the project metrics went up sharply and the first decisions could no longer support such rapid growth. The team urgently took up critical moments, and work on the sound design had to be postponed. Now we are actively developing in this direction, although many problems could have been avoided from the very beginning. Which ones - I will tell further.

    Often, developers do not pay enough attention to the sound at the start and set aside for later. A small saving of time and money at the beginning is fraught with the fact that at a certain stage you will have to redo everything. This can turn into a complicated and expensive process when the project is already running.

    Do not think that only beginners are faced with this. Surely many have heard the sensational story about problems with sound in Metro: Exodus. At gamedev conferences, they also often talk about "how we remade the sound."

    Why it happens? Let's go in order, starting with the obvious.

    Recruit specialists


    The option “our modeler plays the guitar, even if he cuts the sound” - will not work on a large project. To work on game sound, it’s not enough to be able to make beats in FL Studio or have experience working as a DJ at weddings. The development of non-linear audio is a specific area, and when searching for the right specialist, it is important to pay attention to work experience in game dev.

    Even a studio sound engineer or a person with extensive experience in the movie is not always suitable for such a task. They will probably be able to make great assets, but that’s not enough.

    Creating sound effects is only half the task (or even less). Sound designers need a good knowledge of game and audio engines (and related tools); the ability to correctly implement and optimize sound assets; understanding of how the game works; concept of standards, formats and limitations. In the end, the sound designer should speak the same language with the developers and not be stupid when asked to hold brunch, pull pull request or test task.

    When developing a project for PCs and consoles, the complexity of working on sound increases at times and we are not talking about any remote or part time, because one full-time person-orchestra is no longer enough (or he will soon go on indefinite leave due to professional burnout). An audio department of 5-10 people with its manager and allocation of responsibility areas on AAA titles is a completely normal occurrence. But this is a separate issue.

    A person without experience can voice the prototype, but mistakes made at an early stage can become a serious problem in the future. If you are just starting to work on a small game and do not see the point of hiring a special employee to work on sound, it is better to find a remote designer on a project basis. The task of this person is to correctly lay the foundation for further audio development.

    But even with remote work, the sound designer must be involved in the project and interact with game designers, artists, programmers; have access to repositories; work independently in the editor and check the reproduction of sounds in the game. Well, if sometimes he will participate in playtests. Ordering sounds based on text descriptions and screenshots is not the best solution.

    A major role is played by equipment. The sound designer needs really good monitors and, ideally, a specially prepared room to work properly. The famous Microlab Solo speakers, popular among young KRK Rokit beatmakers and Beats swag headphones, are not for this.

    Sound Design Basics by Riot Games

    Take full advantage of the engine


    Beginner developers often use only the most basic audio features without worrying about what's under the hood. I had to deal with projects where all the sound is output directly to the master without a mixer and the most basic processing and with the same (maximum) volume. In this case, everything sounds too “dirty”, and the details are lost in the general noise.

    In the meantime, even the bare Unity has the necessary tools and they work. You can and should use reverb zones, frequency filtering depending on distance, side chain, equalization and compression.

    In Unreal, an incorrectly tuned sound can also cause a lot of pain when finalizing, although the tools for working with audio are quite advanced and convenient here - there are sound classes and various templates. But, again, this makes sense only if you use them from the very beginning.

    Use middleware


    Put in third place, although this is Problem Number One - this is where the root of all evil lies. When developing on Unity, using a standard engine for sound is not a good idea, because it has fairly modest out-of-the-box capabilities. To implement many things, you need to write your own components, and when working with a large number of assets, difficulties appear.



    You can, of course, allocate programmers to finalize the audio, but, in my opinion, it is difficult and expensive. Everything has already been invented before us and has proven itself perfectly. The Unity audio engine is based on FMOD, and even Unity developers themselves recommend using FMOD Studio in the official manual.

    Using a third-party sound engine - Wwise or FMOD - greatly simplifies the process, removes almost all the sound work from the client team and opens up great opportunities for the sound designer, both in terms of working on sounds and in terms of optimization.

    All sound settings in Unity are tied to prefabs and scenes, and an external audio library with all parameters is stored separately - changes in it do not affect the rest of the project. In addition, 90% of the work takes place in a separate editor, which is more convenient for working with audio.


    FMOD Studio

    A simple example. Let's say there are 50 types of weapons in a game on Unity, each with its own set of sounds. You want to change the hearing distance of the guns. To do this, you will have to change the settings in each prefab (manually, by script or by AutoCorrect). But if someone else works on these prefabs (for example, artists or game designers who rule the balance), version conflicts, endless merges, additional work for QA and subsequent edits will immediately begin because 5 out of 50 prefabs are lost somewhere or who Something has already uploaded its version on top.

    And if you use FMOD, then all changes can be made once in one template in a couple of minutes. Using middleware, you make the process of working on sound more convenient, faster and cheaper. This saves time not only for sound designers, but also for programmers, game designers, and testers.

    Among the opponents of this approach, there is an opinion that third-party audio engines reduce performance and increase the build size. They are especially wary of some mobile projects where no more than 10-20 sounds are planned. But this is not entirely correct:

    • Firstly, ten sounds quickly turn into a hundred.
    • Secondly, the use of middleware affects the performance in the best way because of the “smart” compression and wide optimization options: you can separately prioritize for different classes of sounds, scale the settings for each platform and track everything in a separate profiler. It also makes sense to store various types of assets in separate banks and load them only at the moment when they are used - this will save us RAM.
    • Thirdly, the build size can just be reduced by using a small number of assets and the rich possibilities of combining them and processing in the editor.

    There are many examples when using Wwise in games with very strict limits on the build size, they managed to make generative dynamic music using a microscopic set of basic samples.

    Both sound engines - FMOD and Wwise - are free under certain conditions (one game per year with a budget of no more than $ 500,000 for FMOD, and for Wwise - a budget of $ 150,000 and no more than 500 sounds in the project). Both are fully ready to work with Unity and Unreal and connect in a few minutes.

    A short but very important lecture about middleware

    Do not take sounds anywhere


    There are not many high-quality sound libraries in free access and they have been used more than once in other projects. Many sounds (on the same freesound.org ) were recorded with extraneous noise or reverb. And “free” services almost always have restrictions on commercial use, written in small print and valid for one of the ten downloaded files.

    In paid libraries, too, everything is not always perfect - you can pay a decent amount and use several sounds from the entire set, and then find that they were already in a dozen games and films. It’s easier to use subscription services that give access to the same libraries for $ 10-15 per month (for example, Soundsnap or Splice).

    It is also difficult to make high-quality sound from disparate free samples or ready-made theme sets. Ideally, sounds are made for animation and should be uniform in nature and dynamics, consistent with the overall style. Therefore, when choosing libraries, it is better to focus not on ready-to-use assets, but on correctly recorded sources. In addition, it is sometimes easier to record the sound yourself than to look for a suitable option in the bowels of the library.

    Evaluate deadlines correctly


    There is no need to explain. Situations “urgently need to voice ten new characters, tomorrow the frieze!” - better to avoid. Although someone thinks that working by sound comes down to simply copying suitable samples from libraries.

    It’s better to set tasks for the sound designer in advance, ask him to evaluate the timeline and notify as animations and FXs appear - everyone will be happy.

    At the start of a new project, it makes sense to start a library with empty placeholder events and add audio tapes to them as new content becomes available.



    Audiokinetic Wwise

    Think about the players


    Think of the players

    “All the same, everyone plays without sound” - this opinion is found among some developers of mobile projects. Indeed, some players turn off the sound (especially in public places), but many use headphones, play at home with the speakers turned on on a smartphone, Android set-top box, Apple TV or computer through the Gameroom.

    Sound information is processed by the brain faster than visual, and there are a lot of players who need audio feedback. When they click on a button, they want to hear a click; when they shoot from an imba-gun, they want to hear a big explosion, otherwise they will think that this weapon is not very.

    The discrepancy between sound timing animations, frequency conflicts, problems with loudness and resonance, the lack of sound response where it should be - these little things add up to the overall impression of the game. Faced with “minus ears-sound” at the very beginning, the player will turn it off at best, and at worst, decide that everything else in the game is done the same way.

    A satisfied user will simply keep silent. If the game is all right with sound, then in most cases no one will tell you specifically. But bad audio will definitely add negative comments.

    Eventually


    Mistakes made at the beginning of development will be more difficult to fix in the future. Use middleware even in the simplest projects - this will save time and money, as well as reduce the build size without compromising on quality. Refer to experienced industry professionals, and when working on large projects, take sound designers to the staff. Respect the feelings of the audio players. Quality sound is no longer a luxury, but an important and necessary aspect of games. And this also applies to mobile games.

    FAQ


    Q: What to read about game sound design?
    A: See the links at the bottom, I advise you to start from the first.

    Q: Which software is better to use?
    A: Not so important. If we are talking about DAW and plug-ins, use the kit that is most convenient and accessible for you. Opportunities are now roughly equal everywhere, the only difference is in habits and approach. If we talk about highly specialized tools, then you can pay attention to the products of Krotos and LeSound.

    Q: How to choose between FMOD and Wwise?
    A: The choice is not easy, since both candidates are very worthy. In short, Wwise has more features and a higher entry threshold, and FMOD is easier to learn and has an intuitive interface. The best option is to master both engines and use them depending on the requirements of the project.

    Q: The budget does not allow even a remote sound designer to be hired for a while, what should I do?
    A: In any case, from the very beginning use middleware, especially since for projects with a small budget it is free. You will not lose anything, but then it will be much easier.

    Q: But is it possible to make a good sound even without middleware?
    A: You can. It's just that it's longer and harder.

    Q: How important are monitors to work?
    A: Very important. You can save on any other equipment, but not on monitors. The task of household speakers and headphones is to improve the sound and mask its shortcomings, and monitors allow you to identify all the jambs. Without them, you can do everything right, but to adequately evaluate the result will not work. No one forces you to buy a Barefoot or Genelec 1238 for half a million apiece, but a decent Adam A7X level speaker is necessary for normal operation.

    Q: Do I need to build my own studio?
    A: Depends on the volumes. If you have sound designers in the state and they are constantly busy with work, then your own studio will be a good solution (well, or you already have one). If volumes are smaller, there is no studio and is not planned, and the sound designer has his own equipment and the necessary conditions, then it is better to give him the opportunity to work from home part of the time.

    Q: How to choose the right equipment for a beginner sound designer?
    A: Start with monitors, a good sound card, midi keyboard and recorder. For a start, the budget Zoom H1 is also suitable, but it is worth ordering wind protection and shockmount on Aliexpress. Equipment should be upgraded as it will generate revenue and when you encounter restrictions and understand what and why you need to replace.

    Q: What events are worth attending or seeing in the recording?
    A: GDC, GameSoundCon, DevGamm, 4C, Waveforum, AES International.

    References


    Tags:

    Also popular now: