Designing a musical boss fight

Hello everyone and welcome to this week’s article. Today, I want to share the design process of the final boss fight from my last academic project, which was one of the most unique challenges I encountered during my course. To briefly summarize, my last project was a rhythmic action roguelike that aimed to create fast-paced combat encouraging players to follow the beat. You can find more information about it here.

Boss first phase

Step 1: The Idea

Every roguelike needs an epic boss fight to conclude the game, so we had to create one for ours. I started by looking at our main references: Hades and Hi-Fi Rush. Hades has fast-paced boss fights where every point on the map becomes dangerous, and you have to carefully consider your position and timing. Hi-Fi Rush boss fights follow a similar philosophy but emphasize the rhythmic aspect—if you listen to the music, you know when it’s the right time to dodge. Of course, many of the attacks in Hi-Fi Rush weren’t feasible with our camera and combat system. I had to design attacks similar to the ones in Hades but make them musical. At the start of this project, I tinkered a bit with Metasound on Unreal, and I knew it could read cue points in the WAV file and use them as triggers. So, I decided to create the attacks in conjunction with the music, placing cue points in the melody and associating each one with a different part of the attack. To avoid making the music repetitive, I opted to design six attacks associated with six short melodies, all written to fade smoothly into one another and randomly picked each time. Finally, to give players some time to hit the boss, every two attacks, there is a rest melody without attacks. Finally, any good boss needs a second phase, and for ours, the answer was obvious: speed the music by 20 BPM to make everything more frenetic. In addition, we added a rectangular area damage going back and forth in the arena.

Boss second phase

Step 2: Designing and Balancing

The attacks were designed to:

• Be as clear as possible

• Be easy to link to the music

• Put the player in danger in a good portion of the arena

• Utilize concepts the player is already familiar with, possibly recycling assets already employed elsewhere

For example, I mixed the mortar attacks of one of our enemies with the status puddle of another. It made thematic sense to have the boss master the abilities of the enemies in the game, and it helped players familiarize themselves more quickly with the boss as a whole. For almost every attack, I used the combination of preview on cue point + attack on another, as it was very clear and easy to link to the music. 

After implementing the attacks, it was time to write the music, but for this design, it was important to nail down exactly all the balancing regarding the attack duration before writing any melody. Otherwise, the me-composer would have become really mad with the me-designer. So, I wrote a simple track, with just the drums counting the beats. As this was late in development, I already had a good sense, for each attack, of how long it should last and the duration of its anticipation to allow players to dodge. Of course, this intuition had to be tested: I put the cue points where it made sense and then immediately tried out the fight, using the simple rhythm of the drums to easily count the beats and understand how to balance durations in the next iterations.

Step 3: Writing the Music

After I had the cue points and the design of the attacks, it was time to write the melodies. I put a constant heartbeat to represent the fact that the heart of the forest is the actual boss that “haunts” the trees. Then, I assigned a high-pitched organ to the moments where players needed to really be careful and a crash plus a tympan strike on the cue points. In the same attack melody, some musical segments are repeated to help players memorize the patterns. To allow a smooth transition between attacks, all melodies had to last exactly one bar (to allow an entire chord progression). This set the duration of each sequence, as the BPM were fixed for gameplay reasons. For each attack, I tried to evoke a sense of epicness and fear, with the attacks reaching a climax but laying on a bed of scary instruments.

Step 4: Assets and Animation

Finally, we had to implement all the VFX and animations for the attacks. Our artists did a tremendous job bringing this boss fight to life, and I’m really grateful that the entire team supported this idea even though it followed a logic so different from the other enemies. Indeed, in this area, I lacked the experience to understand how the cue points logic would affect our artists’ job. For example, the game can only make the actions trigger on the cue points, so it was difficult to synchronize the attacks to the animations, as they all started with an anticipation sequence. As soon as this problem arose, I prepared a spreadsheet to calculate the right duration for the animations, and our poor 3D artists had to modify them accordingly. In the end, they managed to do it in time for the release, but I really wished I could have thought of the consequences of my design and involved them sooner in my decision. 

Bonus step: the future

I often like to point out in this blog all the things I learn and this boss fight really taught me a lot. I already expressed my biggest regret in the last point but I want to add that I also didn’t like how the expanding circular attack ended up. I designed it to make the boss fight less repetitive and so I liked the idea of an attack that didn’t make use of the previews, but I feel in the end it was too disconnected from the rest. Furthermore, if you are already near the boss it is very difficult to dodge.

Expanding circle attack

Another important thing I learned is this: not everything can be balanced. Sometimes I get caught up in the idea that the right number exists for each thing, but that isn’t always the case and this boss fight was a good example. As every attack had to respect beats, cue points, single bar duration and human reaction times, while contemporary working at both 100 and 120 BPM, some attacks had really strong limitations on balancing. For example it was impossible to make the expanding circular attack cover the whole arena without making it too fast to dodge. I tested the attacks as early as I could, but probably already in the design phase I should have really thought if balancing was feasible. For less complicated things than this, it’s obviously a good idea to prototype the things yourself to discover if any problems can arise. 

That’s all for today, see you next week with another article and good luck with all the bosses in your life!