A guide to designing compelling rhythm games based on our studies and prototyping.
Throughout the past four months, Jam Session, a semester-long project at the Entertainment Technology Center at Carnegie Mellon University, investigated rhythm game design. The team researched, experimented, and created rhythm games to figure out what were the key elements that make a rhythm game innovative and engaging. From this, we were able to discover key aspects that make rhythm games such a compelling game genre. We’ve compiled tenets and tips that we believe will help game designers create rhythm games that are both innovative and engaging.
When researching various rhythm games, and creating our own, we recognized three qualities that were in almost every good rhythm game that we played.
Rhythm games benefit from having simple controls. Games like Guitar Hero or Beat Saber only offer players a limited range of input, such as button presses or a slashing action. Since players already have a hard enough time dealing with rhythm, game designers shouldn’t add mechanics that increase the difficulty and take away from the already complexity of keeping a rhythm.
While researching and playing different rhythm games, we noticed how at the end of the game, we always wanted to play again. Whether it was just to try again or play it on a higher difficulty, there was always a reason to want to play more. To encourage replayability, play sessions for rhythm games tend to be short, with levels spanning the length of a single song.
Potential for Mastery
Tying into replayability, rhythm games often gives players an impression of how well or poorly they are doing, usually through a scoring system. This causes players to want to keep playing to achieve higher scores, honing their skills at the game.
With these three design tenets in mind, rhythm game designers can create more compelling experiences that help players achieve flow.
While thinking about these three traits for your next rhythm game, there are also a few more things to keep in mind.
Realism, Predictability, and Control
One tactic that we used to come up with avenues of exploration was to think about rhythmic activities we do in real life. This was relevant when we explored virtual reality prototypes, given that VR is intended to simulate reality, and games with elements of realism flourish on the platform.
Our first VR experiment, Chopin Beets, was intended to be an innovation on how note highways are displayed in VR. We wanted to situate the player in an environment with a clear activity: chopping food in a kitchen using realistic movements.
In this game, food fell to the beat onto a chopping board, and players needed to chop and swipe food away. To exploit the strengths of virtual reality, food objects were physics-enabled and responded to players’ actions as they would in real life. Players could push the objects into each other and they would collide with others.
We soon realized this created various problems. While it was fun and compelling to see these things interact in a physically-accurate way in VR, the objects’ unpredictable movements and collisions made it hard for players to succeed in the game rhythmically. Allowing the cues of our game to be physics-responsive made the actions players had to do more unpredictable and gave players less control over what was going on. This is opposite to how most rhythm games operate, as players are usually allowed a limited number of actions but have full control over how and when to execute the necessary actions.
Another thing we realized was that most players have preconceived ideas about real-life rhythmic actions that may not be accurate. Rather than attempt to make the game more accurate to real life, it’s often better to have game rules conform to what players expect.
For example, we started off programming Dutch Double as a 1:1 replica of what jump rope is, where players must be in the air for a certain period of time after hearing the slap of the rope against the ground. However, playtesters found it difficult to time their actions — their immediate instinct was to try to land on the slap sound. After making that adjustment, players had a much easier time. Therefore, instead of being a slave to realism, we chose to conform to players’ expectations.
Timing and Feedback
If there’s one thing we’ve learned throughout our time playing and creating rhythm games, it’s that feedback is extremely crucial for players, and should come in as many forms as possible, including audio and visual.
When we first embarked on this rhythm game adventure, we wanted to stray away from the most common type of rhythm game, which uses a “note highway” presentation, where notes come at the player to a certain point or “action bar”. Due to this, our first prototype, Gang Beats, a multiplayer fighting game where players hit each other on the beat in order to be the last one standing, focused more on audio cues rather than visual cues. However, we quickly realized that we needed both. Our playtesters had a hard time grasping the rhythm while playing because there wasn’t any visual feedback.
We then decided that it was important to identify the purpose of feedback, which could be played at three distinct timings, categorizing it into pre-feedback, immediate feedback, and post-feedback.
Pre-feedback refers to giving players enough look-ahead, which lets them know what upcoming actions they have to perform, in the form of a visual or audio cue. Some amount of look-ahead is always necessary, because players can’t react instantly. Almost all rhythm games have this, especially if they have a note highway.
When a player gives an input, they should get some kind of feedback, be it audio or visual. For example, if the player hits something, they should hear a sound confirming that they hit the object and should see the object change as well. Without this feedback, the action may feel unrewarded. Of course, this type of feedback is important when designing any game in order to make it feel “juicy” and responsive.
If the audio feedback needs to sound good with a background track (melodic feedback), it might be good to delay playing feedback for early inputs until the next available beat, and play feedback for late inputs as early as possible (i.e. immediately).
Post-feedback lets players know how well they’ve done with their rhythmic input. Not providing players with this information makes it difficult for them to self-correct when they go off beat, and results in confusion over their level of skill at the game. This can be seen in many games in the form of “Miss/Good/Perfect” messages that appear after players perform certain actions.
The concept of flow isn’t unique to rhythm game design, but is vital to keep in mind. Rhythm games tend to involve much shorter play sessions compared to other types of games, but players are required to be in a more focused state of flow and concentration.
One way to tell if players are in a state of flow while playing a rhythm game is to watch for physical indicators such as players bobbing their head, tapping their feet, or swaying to the music.
Identifying flow was imperative for our prototype, Laserfade, since players had to rely on their inner sense of rhythm. Laserfade is a VR rhythm game where players are tasked with shooting targets that appear in the wireframe -style world around them.
In Laserfade, the players who performed the best were those who found themselves in that flow state. Since the music in the game faded out to test players’ sense of rhythm, it became clear that players who performed the best were those that remain in flow even without music. We saw this most often when players counted the beat to themselves while the music faded out.
To ensure that players enter the flow state, it’s important to design the game’s actions and inputs such that they are intuitive and predictable. In games that use beatmaps, actions should flow into each other by having consecutive cues placed in adjacent positions, while actions that might put players in uncomfortable or impossible positions should be avoided at all costs. In games like Beat Saber, well-designed beatmaps allow players to make movements that flow naturally into the next one, while preventing players from having their controllers collide into each other or awkwardly moving between rigid poses. To put it simply, a rhythm game with good flow has well-choreographed actions.
We explored choreography and movement in our 3rd VR prototype, Kinectic, where we took a choreography-first approach to beatmap design. Players must push ball-shaped cues that appear in front of them with a corresponding hand, tracing a line through spherical targets with their movements.
In Kinectic, we created our beatmap entirely in VR in real-time, which was the best way for our beatmap designer to feel out the space. The designer was able to test out movements and
this made it incredibly easy to map out fluid movements that naturally flowed into each other. Real-time beatmapping is an invaluable tool for creating solid beatmaps. We suggest that this should be a rhythm game designer’s first line of defence when creating patterns of action for their own rhythm games (fine editing of the capture data can be done using other tools, such as spreadsheets).
The discoveries we’ve highlighted in this article are only the beginning. During the course of our studies and prototyping, we also developed a rhythm game taxonomy and dictionary that we used to better explain and categorize both existing rhythm games, and the new ones we created. If you’re interested, check out our more detailed writeup about our semester here. You can also check out our blog and learn more about our project here.