menu
 

Behind the Sound of Headbangers: Rhythm Royale | Bringing the Pigeon Rhythm Game to Life

游戏音频

Headbangers: Rhythm Royale is the world's first competitive online music game. It pits you against 29 other players in intense musical and rhythmic battles. The challenges involved in creating this game were enormous. In this article, I'll explain some of them and show you how we managed to get around some of the limitations.


Online synchronization

The way we synchronize our 30 online players in Headbangers is unique. The game is not based on the characters' movements in space or their actions, but on their responses to musical events. We therefore decided to synchronize the client and the server to the music and its progress. 

Since time and validation are managed by the server, we can detect whether a player is out of sync with the network time. Above a 100 ms time lag, the player will be resynchronized, which in practice means that the playback head will move in the music to synchronize the player at the right moment. Thanks to extensive network optimization, we were able to ensure that these resynchronizations were extremely rare. 

When it came to implementing all this in Wwise, we were faced with a real challenge: Wwise is not designed for the method we wanted to apply. There's no simple way of moving the playback head of music and sounds forwards or backwards according to network time instructions.

We came up with a (largely random) trick: using the transition system to create specific rules. We noticed that by creating an immediate, fade-free transition from a song to itself in the Transitions tab of a Switch Container, we could now move the playback head as we wished. 

img1

This method was quite cumbersome to apply to the whole game. Headbangers contains thousands of musical segments (over 50 hours of music), and this procedure had to be applied to every musical segment in the game. Which brings me to my next point. 

SoundBank management and optimization

The sheer quantity of sounds and music contained in Wwise (tens of thousands in total) makes it certainly one of the most voluminous games in terms of sound content in the history of video games. With the game being released on all consoles and the PC, managing memory and SoundBanks (especially for the Switch) was a real challenge. 

My advice is to anticipate these issues as far in advance as possible. On my end, I had to deal with this concern late in the development process, which forced me to completely rework my architecture. 

There are 23 mini-games in Headbangers. In the end, we decided to divide the game into 24 SoundBanks: one SoundBank per mini-game and one general SoundBank, containing sounds common to all mini-games, as well as menu sounds and music. The decision to distribute sounds between these SoundBanks is a key element in memory management. We therefore had to think long and hard about the use and place of each sound in the game to optimize all this. 

In the same spirit of saving space and memory, we also had to optimize conversion. We therefore created several conversion parameters in Wwise, some of which are only used on certain platforms such as the Switch. So, by converting all sound effects to ADPCM on this console, we were able to save a great deal of space with very little loss of quality, while maintaining a more optimal result for music (the key element of the game) in Vorbis.

img2

Synchronize audio feedback with player actions

Most games accept a certain latency between the moment the player presses a button and the sound feedback of the action. In Headbangers, it was impossible to be satisfied with this latency. Indeed, for a musical game, if the time lag between the button press and the sound feedback is significant, the player's entire perception is blurred. 

We therefore decided to apply streaming parameters to certain sounds directly linked to gameplay. Defining that a sound will be streamed means that it can be called up more quickly, thus reducing audio latency. The key is then to set the Prefetch length for each sound, to avoid source starvation effects or untimely audio cuts. 

img3

In one of our mini-games, Piano Mezzo Forte, our pigeon has to sing over background music. So we created an event for each note in each nuance (Piano, Mezzo, and Forte) and set their Prefetch length according to their characteristics and durations. Even if the end result was totally ridiculous (that’s what we wanted), this near-perfect synchronization was very important, as you can see in the video below. 


Dynamic menu music

What I particularly like about Wwise is how easy it is to create dynamic music.

Even though Headbangers wasn't the game that needed most in this respect (it's a very scripted game), I had a lot of fun creating dynamic, ever-changing music in the menus, which is always appreciated by players.

Our main menu is made up of several tabs, each representing a different menu in the game: Home, Progress, Challenges, Customization, Shop, Options (see video below). 


The process for creating this dynamic music effect is simple. I created one tune per menu, with the constraint that all these tunes have the same tempo, duration and harmonic progression. Then simply use Wwise transitions to create an immediate "Same time as playing segment" transition rule, and you can move from a point in one piece to that same point in another piece smoothly and efficiently. 

img4

It's all very simple and quick to set up, and the effect is always satisfying. Wwise really helped us meet a lot of the challenges on Headbangers, and we consider it to be a really reliable and robust audio engine, yet accessible and intelligent enough for composers or sound designers. 

Charles Bardin

Director and Co-founder

Glee-Cheese Studio

Charles Bardin

Director and Co-founder

Glee-Cheese Studio

Charles Bardin is the director and co-founder of French video game studio Glee-Cheese Studio. For over 10 years, he has been creating sound worlds for a wide variety of independent games, including Stellar Overload, Overloop, A Musical Story, and the very recent Headbangers: Rhythm Royale.

 @GleeCheese

评论

留下回复

您的电子邮件地址将不会被公布。

更多文章

人人都能用 WAAPI(二)wwise.core 分支

大家好,我是溪夜。 在《人人都能用 WAAPI(一)概述》中,我们用思维导图对 WAAPI 进行了重新归纳,并在配置好开发环境后,一起用 Python 写了几个简单的小程序,体验了 WAAPI...

29.10.2020 - 作者:汪洋

使用Wwise制作拟真且泛用的多普勒效应

前言...

11.8.2021 - 作者:徐巍

赛博朋克世界的声景塑造 — 以游戏《底特律:成为人类》为例 (第一部分)

30.8.2021 - 作者:赵茉茉

在 Wwise 中构建环境声系统

简介 大家好!我叫刘易斯,是 Soundcuts 的声音设计师,同时也是 Airwiggles...

10.12.2024 - 作者:刘易斯•汤普森

《Alan Wake 2》背后的声音设计 | 第 3 部分 – 音频 QA 的作用

《<Alan Wake 2>背后的声音设计》系列博文 《<Alan Wake...

2.4.2025 - 作者:Adam Butterworth

《驱入虚空 (Into The Void )》的音频设计分享丨如何通过声音塑造沉浸之旅

简介 《驱入虚空(Into The Void)》是一款充满冒险与未知的游戏,故事背景设定在一个全新开始的世界。 庞大的建筑群从天而降,带来了新的人类和未曾预料的新使命。...

29.4.2025 - 作者:高庶人

更多文章

人人都能用 WAAPI(二)wwise.core 分支

大家好,我是溪夜。 在《人人都能用 WAAPI(一)概述》中,我们用思维导图对 WAAPI 进行了重新归纳,并在配置好开发环境后,一起用 Python 写了几个简单的小程序,体验了 WAAPI...

使用Wwise制作拟真且泛用的多普勒效应

前言...

赛博朋克世界的声景塑造 — 以游戏《底特律:成为人类》为例 (第一部分)