Sakurai on Creating Games: Sound (14 Topics)

Masahiro Sakurai, the creator of the Kirby and Super Smash Bros. series, shares a wealth of game-development know-how on his YouTube channel “Masahiro Sakurai on Creating Games”. This article summarizes and restructures the content of that channel by topic.

That said, a summary is only an entry point. So much of the value — Sakurai’s own words, his real-world examples, his pacing, and the footage itself — can only be gotten from the videos. So please don’t stop at reading the article; I strongly encourage you to also watch the original videos embedded under each topic.

This category covers ideas for supporting a game with sound — from sound-effect priority to “no masterpiece has bad sound.” Here we bring together the key points of both videos in the “Sound” category (14 individual topics), structured in 2 parts following the 2 source videos. Each part opens with its explanatory video, so please watch along.

Part 1: Tempo, SFX priority, listening across environments, voice recording (#01–#07)

Sound from a game includes music, sound effects, and voice. Usually we don’t think much about it, but sound greatly sways a game’s feel and sense of reality. The knack of its design and production is discussed.

https://www.youtube.com/watch?v=pW_m_Nc5css

1. A tempo that suits the game

Tempo is the measure of how fast beats are struck in music, also called BPM (beats per minute). One beat per second is about BPM 60; two per second is about BPM 120. By a game’s content there’s a tempo that fits and one that doesn’t.

Stage tracks arranged for Smash matches are often a bit faster than 120, around 130–170, some about 200. Even in the same versus game, the right tempo differs by the speed of offense and defense. Even within one game, the right tempo and rhythm change as the scene changes. It’s a realm of strong personal taste, but it’s good for more than just the composer to think about a tempo suited to the game.

Sakurai cites Dragon Quest I. There’s a trick where the deeper into a dungeon you descend, the slower the music’s tempo, which he says is very effective. Perhaps because it’s hard to implement, it’s an idea rarely seen in other titles.

2. Balance SFX by importance

It’s no exaggeration to say sound effects always play in a game, so you need to think hard about their volume, flashiness, and prominence. For example, between a walking sound and a sword’s whiff sound, which should take priority? Sakurai’s advice is clear: make sounds that relate strongly to the player louder.

The loudest is the KO sound. As the big event that decides mind games and victory, it’s most important. Next is the moment your own action’s result returns directly — a hit, a perfect shield. Then specials, voice, explosions. Voice stands out over SFX when spoken clearly, so balance it. After that: dodge and shield block → whiff and dash → jump and landing → walking. Landing is deep: it’s remade by clothing and boot material, even down to differences in Cloud’s landing sound by costume.

Things like footstep sounds get too mechanical when the same sound repeats, so they stage variation. Note that walking can have higher priority in a 3D game; not hearing it feels like hovering. You need to change the sound by ground type (attribute/material), which is a pain, but genuineness creates a firm sense of feel, so you have no choice but to work at it. Watching only the size of an action won’t give a sensible volume balance.

3. Listen across various environments

For sound checks, he never skips listening across multiple environments — headphones, PC speakers, and importantly a cheap TV monitor. A small built-in TV speaker loses most of the bass and lacks power, but this is the most widespread and what the most people hear.

Fine-tuning carefully in a studio’s soundproof room or on good speakers is necessary too, but if you fuss too much over sound that’s delicate there, it won’t sound good at all on consumer speakers. Preparing different environments and making each fine is tedious, but catering to the customer’s setup matters. Note that with this method it’s hard to replace headphones — if the baseline shifts, it no longer matches past checks, which is trouble. He doesn’t weight the Switch’s built-in speaker much, since it sounds clear and many lower the volume in handheld mode. He recommends not finishing on the studio’s big speakers alone.

4. Game music as ambient music

“I can remember and hum old game music, but today’s is vague and hard to grasp” — you hear this a lot. Try playing Famicom music over a realistic modern game and it feels noisy. As graphics evolved and visual information increased, foregrounding the music too much stopped fitting.

A realistic screen needs ambient sound too — pulling grass, treading earth, gear clinking… so many components that you must consider each one’s presence. Meanwhile, Sakurai also feels too many tracks are made for one work. If there are needlessly many tracks, you can’t remember them, and can’t get into the groove. Holding them down moderately and using them well is good.

Sakurai himself often makes a main theme and applies arranged versions of it to various scenes. A track whose main phrase you remember is more effective. Either way, building it in step with the player’s emotions is good, he says.

5. Sound is fiction, and nonfiction

If a fighting game’s SFX were truly realistic sounds, it wouldn’t have been accepted this much. The Shoryuken sound in Street Fighter II is enormously fake — by Sakurai’s sense, exaggeration like a “Showa-era tokusatsu film.” Japan originally had a culture of exaggerating sound, anime included.

The more realistic the visuals, the more an exaggerated sound feels off. Yet a fully realistic sound makes a punch go “pat,” which is unsatisfying. Without the awareness that a fighting game is something far removed from real fighting, you miss the essence.

On the other hand, realistic-ish sound has merit too: it gives meaning where there is none, supplementing information. For example, a “tap-tap” sound while crafting lets you picture what object you made and how. Even with simple graphics, sound can make things feel like they have the texture of wood or stone. Sound is both exaggeration and realism, and taking a balance suited to the game increases the sense of reality.

6. Voice recording

Making one fighter takes considerable time across staff, but voice recording finishes in just about an hour per fighter. Including the reveal trailer and scenario, it’s under two hours. The involved time is short, yet since the voice sometimes goes in as-is, you strongly feel a human presence.

Recording starts with casting. Many characters already have voice actors set, so there’s nothing to decide. He weights staying close to the original, but matches if a series entry refreshes its cast. Next, write the script — usually horizontal, but he writes vertically to match a general script. He drafts the lines including words like “Come,” “Ha!,” “Hmph.” If the original has the same lines, he prepares reference audio.

Smash’s lines are overall short and infrequent. Long or frequent lines during battle reduce immersion. He also deems voice clashes bad, so keeps Bowser and Donkey Kong’s voices beastly, considering overlaps too. He usually directs the recording himself; overseas voices proceed led by Nintendo of America/Europe based on the Japanese voices. Pokémon is tough since the words differ by country. Surprisingly, after recording is the hard part, with editing repeated to fit the game. He listens to the voices the whole time during production, so maybe developers hear them more than players, he says.

7. Arrangement methods

In Smash, very many composers arrange very many tracks. First Sakurai picks plenty of arrangement candidates from various works, gathers the musicians, and explains the key points: match tempo and tension to the match; an intro during the start countdown is effective; mixing in other tracks after the main one is effective; a length of 2 to 2.5 minutes ending so the loop head is faintly heard is the basis, and so on.

After the explanation, he has them choose tracks, because it’s better to have them choose tracks they’re attached to or good at. Even versatile people have favorable genres, but since it’d be unfair not to talk with everyone, he keeps “who does what genre” hidden. Each takes it home, checks the original, and creates, and Sakurai checks every track. Supervision goes through the sound director — he gets the feel but can’t use musical jargon, so this is best.

Sakurai weighs that the main melody comes out roughly the same as the original if you hum it. This comes from being disappointed by game-arrange CDs remade into something else. “It’d be creepy if the Star Wars theme were different every time” — customers hear it as one among many, so the sameness should be kept. He hopes you’ll enjoy the vast sound library created that way.

Part 2: The original is always right, strong attack, the Famicom sound chip (#08–#14)

Where the first half was about “designing” SFX and voice, the second half goes deeper into sound-making — the mindset of arrangement, crafting SFX, and the Famicom sound chip. It closes with the conviction “no masterpiece has bad sound.”

https://www.youtube.com/watch?v=93lE-ELzQu4

8. The original is always right

A continuation of the prior music-supervision talk. In an arrangement of SNK’s Athena, supervision unusually got into a quagmire. The cause was a misunderstanding of what Sakurai called the main melody. When it came to which of the completed track’s A-melody or B-melody was the main, it felt like A when heard in the game or on the Famicom, but by a musical view it was B.

Whatever the original is, even if it strays from musical technique or convention, the track you heard in the original at that time is the right track. A difficult thing. Likewise, SNK’s Psycho Soldier is said to be the first game in the world to fully sing a BGM, in 1987. Smash Ultimate did a vocal arrangement — the Japanese version through verse 3, the English version only verse 1 (because the original switches to the internal sound chip after verse 1) — recreating even the original’s structure. A beloved original — the original is wholly right. You need to avoid harming the experience each player holds.

9. Strong attack, moderate reverb

Look at the waveform of Smash’s SFX and the sound rings firmly at first, with reverb. But the sound that comes up from the sound team before Sakurai’s check fairly often has reverb that’s long and too grand. Smash plays with 4 or 8 people, so SFX appear all over the screen, becoming a flood of sound, and long reverb blends with the next sound and gets buried.

So he has them keep two points in mind. One is a strong attack — a fast onset, an instant peak at the start. A hit’s sound mustn’t be heard slightly late. The other is moderate reverb. Long reverb feels grand, but if the next sound rings while it’s still going, it blends. A weak attack’s hit sound should reverberate only enough to keep the sense of mashing when mashed. A strong attack should ideally have its main sound finished while the screen is frozen in hit-stop. The resulting policy is “strong attack, moderate reverb.” Not just whether it’s a good sound, but whether it’s a sound that suits the game’s purpose, must be examined.

10. Directing sound effects

When SFX from the sound staff differ from your image, you need them adjusted, but since sound has no substance, directing that supervision is extremely hard. It’s not something a voice imitation settles. Concrete expressions like “strong attack” he gives each time, and for things everyone knows, he can use comparisons like “like a jet’s sound.”

But he wants to avoid casually citing examples not everyone knows (they don’t land). Sometimes he specifies a YouTube link of a similar sound. Partly due to telework, he doesn’t direct much with onomatopoeia. Even writing “gaki” or “boko” in an email — though it rings in the writer’s head, it doesn’t come across on the page.

But he heard a surprising thing from the sound leader: “sometimes such onomatopoeic expressions help.” Perhaps it’s wanting more info, more clues — Sakurai senses something beyond logic. As a director, he intends to OK a sound that suits the game’s purpose even if it’s not exactly the one in his head, and a sound exceeding his imagination does come out at times. Vague and hard, but since it conveys directly to the feel of play, you’d want to work to get close. Of course, having a clear image from the start is the prerequisite.

11. Balancing ambient sound

Smash has ambient sound too, though it’s hard to hear. In importance and priority, ambient sound is lowest, so it’s a problem if it’s excessive and drowns the crucial SFX. Yet it’s not inaudible either. In an environment surrounded by good speakers, it ringing properly matters — dead silence when you pause has no atmosphere.

A game also needs to make you feel as if you’re in that world, especially in first- or third-person. Unlike other SFX that demand exaggeration, ambient sound is one where a fairly realistic sound works. Even a comical game can layer realistic ambient sound without issue, and the know-how of video works applies easily. Set the environment without slacking, to increase immersion.

12. When you cite another track as an example

Thinking an existing track fits your game ferociously well, you sometimes order “I want something like that track.” It can be very useful as reference, but it can be dangerous. First, if the director’s own head gets dominated by that track, you start feeling everything else is fake and can’t accept it. The composer can get dominated too, and making something similar but different is painful. If you don’t know what to change, you lose sight of the goal.

God Hand’s stage-1 track was taken by composer Mr. Takada on an order to “make a track imagining The Ventures.” Even without leaving in The Ventures’ signature guitar, it ended up evoking a surf sound. On the other hand, there are cases of negotiating with rights holders to use a good track as-is. The track “Hana” from Oreshika fit perfectly as the theme of protagonists cursed with death who pass their will to descendants to live on, and planner Mr. Masuda reportedly negotiated directly. What you’re making anew isn’t another company’s work, so proceed flexibly with what’s needed.

13. The Famicom sound chip

As a bit of a chat, the Famicom sound chip. No other single piece of hardware evolved this much, before or since. The Famicom’s sound is made of two square waves, one triangle wave, one noise, and one DPCM.

The square wave is a clear “pee” tone, whose timbre changes by duty ratio, used for the main melody. The triangle wave is soft and suited to bass (the Famicom’s is strictly a pseudo-triangle). Noise is used for hits, explosions, and wave sounds. With these 5 sources they juggle both music and SFX, sometimes muting part of the music while an SFX rings. DPCM is a sampling source that records and uses sound (famously Super Mario Bros. 3). It ate capacity so was modest early on, but evolved remarkably in late works.

The Disk System added a wavetable sound source, and expansion audio with an IC in the cart appeared often too. But because the overseas NES had the terminal for cart-built-in audio removed, several works have a different musical flavor in Japan and overseas (later, increased DPCM capacity closes the gap). Sakurai recommends the NES Silver Surfer, and Naoki Kodaka’s Rough World (Journey to Silius) and Battle Formula (Super Spy Hunter). Focusing on the technique of juggling within limited functions is genuinely interesting.

14. No masterpiece has bad sound

Lastly, a conviction. From his feel touching countless games, Sakurai thinks “no masterpiece has bad sound,” and says the more he does it, the more sure he is. Aside from the ancient exception of having no sound at all, even SFX of only noise — a masterpiece’s sound is well made.

There are works that are masterpieces despite poor graphics, or despite gameplay that doesn’t come together. But “a masterpiece with bad sound” doesn’t come to mind (here “sound” includes music and SFX). Taste varies so you can’t say flatly, but recall — when you look back on a game you enjoyed, doesn’t music or SFX usually come to mind first?

Sound strikes the heart more directly than the game itself, more than visuals or characters. So never underestimate it. Note that smartphone games, or handheld games you played on the sly, may be exceptions — it’s possible they were fun even without hearing the sound. But if you can hear good sound without reserve, the joy you can feel rises all the more, Sakurai concludes.

Summary

We’ve now gone through both videos and 14 topics of the “Sound” category. Finally, let’s recap the key points of each part (= each source summary video) by individual topic.

Part 1: Tempo, SFX priority, listening across environments, voice recording (#01–#07)

Part 2: The original is always right, strong attack, the Famicom sound chip (#08–#14)

There’s a lot here, but I hope you’ll revisit it starting from whichever part caught your interest. Please also check out the related categories.