Megumi Fushiguro Voice Impression Guide
Megumi Fushiguro’s voice is one of the most distinctive in Jujutsu Kaisen — not because it is loud or high, but because of what it withholds. That controlled, low-energy delivery, the clipped consonants, the near-flat affect that occasionally cracks under genuine pressure — it is a technically demanding voice to replicate precisely because its defining quality is restraint rather than expression. This guide breaks down the acoustic profile, the DSP settings that get you close, the AI voice cloning workflow that gets you closer, and the performance habits that make the difference between “sounds kind of deadpan” and “sounds like the Ten Shadows sorcerer.”
TL;DR
- Megumi’s voice sits near-neutral male pitch with slight gruffness and almost zero dynamic swell — the challenge is restraint, not a dramatic pitch transformation.
- Japanese dub (Yuma Uchida): –1 to –2 semitones pitch, neutral formant, slightly dry tone curve. English dub (Adam McArthur): near-zero pitch offset, minimal formant drop, cooler mid-range EQ.
- AI voice cloning captures his specific timber and dryness far better than DSP alone — target index influence 0.70–0.80 for natural-sounding deadpan.
- VoxBooster loads community AI voice models on Windows with sub-300 ms latency — no Python, no kernel driver, compatible with anti-cheat games.
- Performance matters as much as settings: slow deliberate pacing, clipped sentence endings, suppressed emotional peaks.
- Use cases: Discord JJK roleplay, gaming with cosplay friends, streaming anime content, VTubing with a stoic character persona.
Who Is Megumi Fushiguro?
Megumi Fushiguro is one of the three central characters of Jujutsu Kaisen, the manga series by Gege Akutami animated by MAPPA. He is a first-year student at Tokyo Jujutsu High, a Ten Shadows Technique user, and the character whose moral philosophy most deliberately contrasts with Yuji Itadori’s straightforward heroism. Megumi does not save everyone — he prioritizes the people worth saving, in his own estimation, and carries that calculation without visible guilt.
His personality shapes his voice entirely. Where Yuji is warm and reactive, Megumi is flat and deliberate. Where Gojo is theatrical, Megumi is economical. Every vocal choice — the low energy, the minimal pitch arc, the short sentences — communicates a character who processes emotion internally and almost never lets it reach the surface.
For voice impression work, this makes Megumi both easier and harder than most characters. Easier because you are not performing wide dynamic swings. Harder because the absence of those swings has to be active and intentional, not just dull.
The Acoustic Profile: What Makes Megumi’s Voice Work
Fundamental Pitch and Register
Megumi’s voice is low-to-mid male range — not dramatically deep, but grounded and unhurried. Yuma Uchida’s Japanese performance places the fundamental around 100–130 Hz in neutral dialogue. Adam McArthur’s English dub runs slightly warmer in the 110–140 Hz range, with a touch more natural resonance that suits Western dub conventions without losing the stoic edge.
Neither performance uses much vibrato. The tone is nearly straight throughout neutral lines — wavering would imply emotional engagement Megumi is actively suppressing.
The Gruff Quality
“Gruff” is the right word, but the gruffness is subtle. It comes from:
- Mild vocal fry on sentence endings — not a full creak, but a slight drop into fry register as sentences close, signaling finality without volume.
- Limited breath support — Megumi speaks with front-of-mouth articulation and moderate breath, not the chest-projected delivery of hot-blooded shonen heroes. This creates a contained, dry quality.
- Forward but narrow formant placement — the voice is not breathy or wide, just controlled and somewhat nasal-forward, emphasizing intelligibility over warmth.
Emotional Restraint as a Vocal Technique
The most technically interesting aspect of Megumi’s voice is how the actors handle moments of genuine emotional stress — Toji fights, the Sukuna vessel arc, late-season confrontations. In those scenes, neither Uchida nor McArthur explodes into a loud shout. Instead, the voice tightens: pitch rises slightly (2–3 semitones), delivery speeds up in short bursts, and there is a controlled roughness on key words. The intensity is compressed rather than released.
Replicating this compression is a performance skill, not a DSP setting. The software can shift and convert your tone; the deliberate suppression of release has to come from you.
DSP Settings for a Megumi Voice Mod
If you want a quick setup without AI model training, DSP pitch and formant adjustment gets you into the right range.
| Setting | Japanese (Yuma Uchida) | English (Adam McArthur) |
|---|---|---|
| Pitch shift | –1 to –2 semitones | 0 to –1 semitone |
| Formant shift | –0.2 to –0.3 semitone | –0.1 to –0.2 semitone |
| EQ — low shelf | Slight boost below 200 Hz (+1.5 dB) | Flat or +1 dB below 180 Hz |
| EQ — high-mid | Cut 3–5 kHz (–2 dB) to reduce brightness | Cut 4 kHz (–1.5 dB) |
| EQ — air | Cut above 8 kHz (–2 dB) for dryness | Cut above 10 kHz (–1.5 dB) |
| Reverb/room | Minimal — dry signal preferred | Minimal |
| Noise gate | –30 dBFS | –30 dBFS |
The high-frequency cuts are counterintuitive compared to most voice changer setups, but Megumi’s voice lacks the bright presence typical of shounen heroes. Reducing 3–5 kHz takes out the forward excitement energy; trimming the air frequencies increases the dry, controlled quality. Add those cuts gradually and test by ear.
The pitch drop is modest — you are not doing a deep villain voice, just grounding an already low-mid register slightly further. Overdoing the pitch shift produces a voice that sounds artificially lowered rather than naturally settled.
AI Voice Cloning for a JJK Megumi Mod
DSP gets the register right. AI voice cloning gets the specific character — the dryness, the particular resonance of Uchida’s or McArthur’s performance, the subtle characteristics that distinguish “stoic young male voice” from “Megumi Fushiguro specifically.”
Finding a Pre-Trained Model
Search weights.gg or community voice repositories for “Megumi Fushiguro” or “JJK Megumi.” A quality model trained on clean JJK dialogue will have:
- Clear training notes listing the source material (season, episodes)
- Multiple emotional registers represented (tactical commands, terse responses, a few emotionally loaded lines)
- Download counts and user feedback indicating the model has been tested
Download the .pth model file and the .index file — both are needed.
Loading the Model in VoxBooster
- Download and install VoxBooster from /download. It injects into the Windows audio pipeline via WASAPI — no kernel driver is installed.
- Open the Voice Clone tab. For Megumi, this produces a significantly more convincing result than DSP alone because of how closely his voice ties to a specific performer’s timbre.
- Import the model via Voice Models → Import Custom Model, pointing to both the
.pthand.indexfiles. - Set pitch offset. For typical male input, –1 to –1.5 semitones is a reasonable starting point. Measure your own speaking fundamental (a free spectrum analyzer app works) and compare against Uchida’s ~115 Hz calm-scene average.
- Set index influence to 0.70–0.80. Megumi’s deadpan quality comes partly from the model’s formant representation — keeping influence in this range preserves the character voice character without over-processing unusual phonemes.
- Apply post-chain EQ. Even with a good model, adding the high-mid cut described in the DSP table (–2 dB around 4 kHz) in VoxBooster’s post-processing chain increases the dry, restrained quality that is Megumi’s signature.
- Enable noise suppression. Clean input is especially important for a quiet, low-energy voice — background artifacts are more audible when the voice itself carries little masking energy.
- Route to Discord or OBS. VoxBooster appears as a standard Windows audio input. Select it in Discord under Voice & Video → Input Device, or in OBS under Audio Sources.
Training Your Own Model
If you want better quality than community models provide, or you want to target a specific performance range (e.g., only tactical dialogue, more emotionally varied), training on curated data produces stronger results.
For Megumi specifically, build training data in three tiers:
- Neutral tactical lines — short, flat, mission-related dialogue. These set the baseline deadpan register.
- Terse emotional responses — single-sentence reactions to team members, clipped affirmatives and negatives, dismissive deflections. These add range without breaking stoicism.
- Compressed intensity lines — confrontation scenes where the voice tightens but does not break. These are the hardest to find cleanly (music beds are common in fight scenes) but they make the model convincing under pressure.
Exclude any lines with heavy background music or sound effects. A model trained on contaminated audio will produce conversion artifacts, particularly on the quiet, sustained resonances that define Megumi’s register.
For the full AI voice cloning training workflow, the AI voice changer guide covers data sourcing, model quality evaluation, and export from start to finish.
Megumi vs. Other JJK Characters: Voice Profile Comparison
Understanding where Megumi sits relative to other Jujutsu Kaisen voices helps you calibrate settings and avoid overlap.
| Character | Pitch Range | Energy Level | Key Acoustic Marker | DSP Offset from Neutral |
|---|---|---|---|---|
| Megumi Fushiguro | Low-mid male | Very low | Dry, restrained, gruff endings | –1 to –2 ST pitch, –0.2 ST formant |
| Yuji Itadori | Mid male | High | Warm, open, dynamic peaks | +0 to +1 ST pitch, neutral formant |
| Gojo Satoru | Upper-mid male | Theatrical | Bright, spacious, performative | +1 ST pitch, +0.3 ST formant, added room |
| Nanami Kento | Low male | Controlled | Deep, professional, even | –2 to –3 ST pitch, –0.5 ST formant |
| Sukuna (in Yuji) | Deep, wide | Menacing | Resonant, slightly layered | –3 to –4 ST pitch, –1 ST formant, EQ boost low |
Megumi is the most acoustically neutral of the group — the character closest to a “clean” stoic male voice with minimal processing. That sounds easier, but neutrality is hard to perform without going flat. The gruffness and deliberate pacing are what differentiate him from simply speaking in your own unprocessed voice.
Performance Drills: Training Your Megumi Impression
The DSP and AI settings do the timbre work. Your delivery is the other half.
Drill 1: The Clipped Sentence Ending
Take any line from the series — “I don’t care about that.” “Understood.” “Stay back.” — and practice ending each sentence with a slight drop into fry register rather than maintaining the vowel. Let the last syllable trail into a dry close, not a sustained note. Record yourself and listen for whether the ending drifts upward (Western speech habit) or holds flat with that slight fry.
Drill 2: Deliberate Pause Placement
Megumi rarely rushes. Between clauses, there is a beat — not hesitation, but consideration. Practice inserting a one-count pause before secondary clauses: “I’ll handle this.” [pause] “You don’t need to worry about it.” The pause signals that the words were chosen, not reflexive.
Drill 3: Compressed Intensity
Take a high-stakes line — something from a fight scene — and practice delivering it at 70% of the volume you would naturally use for intensity. The compression itself creates the tension. Your voice should feel like it is tightening rather than releasing. This is physically counterintuitive; budget several sessions before it feels natural.
Drill 4: Vowel Shortening
In Uchida’s Japanese performance, vowels are relatively short and closed. In natural casual speech (and in McArthur’s English performance), vowels often lengthen for emphasis. Practice shortening every vowel in an intense line by about 20%, then check whether the result sounds more clipped and contained.
Setup for Discord, OBS, and Gaming
The routing for a Megumi JJK voice mod is standard regardless of which tool you use.
Discord: Settings → Voice & Video → Input Device → select VoxBooster Virtual Mic. Enable push-to-talk if you are using AI voice conversion mode to handle the sub-300 ms processing window.
OBS: Audio Sources → Add → Audio Input Capture → select VoxBooster. For streaming, add a Sync Offset under Advanced Audio Settings equal to your measured AI conversion latency. Clap test: record a single clap with mic and webcam simultaneously, measure the gap, apply as video delay.
In-game voice chat: Any game that supports custom input devices will accept VoxBooster as a standard microphone. Because VoxBooster uses WASAPI without a kernel driver, it does not interfere with Riot Vanguard, EAC, or BattlEye anti-cheat systems.
For a complete Discord voice filter setup walkthrough, the Discord voice filters guide covers routing, noise gate placement, and latency compensation in detail.
Ethics of Fan Voice Content
Megumi Fushiguro is a fictional character from Jujutsu Kaisen, a property owned by Shueisha and animated by MAPPA. Fan voice impressions and AI voice mods fall under fan creative tradition with a long history in anime culture.
What is generally fine: personal roleplay, Discord sessions with friends, non-monetized streams, cosplay video content clearly labeled as fan-made.
What requires care: any commercial use, content that could be mistaken for official material, or using a voice mod to impersonate official voice actors personally (rather than the fictional characters they voice). The distinction matters — the character Megumi is fair game for fan creativity; Yuma Uchida and Adam McArthur as real people are not.
When publishing content online, label it clearly as fan-created AI voice content. Do not represent it as official audio or as representing the voice actors’ views or performances.
Using VoxBooster for a JJK Megumi Voice Mod
VoxBooster runs on Windows 10/11 and is designed specifically for real-time voice conversion in gaming and streaming contexts. Key characteristics relevant to a Megumi voice mod setup:
- Sub-300 ms AI cloning latency — usable with push-to-talk in Discord and in-game voice chat without the half-second delay that makes real-time conversation awkward.
- WASAPI-based audio routing — the application operates through the standard Windows audio API, with no kernel driver, which means no conflict with competitive game anti-cheat.
- Integrated noise suppression — runs before the voice clone stage, cleaning ambient noise that would otherwise cause artifacts in Megumi’s quiet, low-energy delivery style.
- Native custom model import — load community AI voice models directly from
.pthand.indexfiles without a separate Python environment.
A free trial is available at /download — you can test AI conversion quality on your own voice against the DSP baseline before committing to a plan. Pricing starts at $6.99/month (€5.99/R$29,90 in other regions). For a comparison of voice modding approaches, the anime voice changer guide covers multiple tools side by side.
Frequently Asked Questions
What makes Megumi Fushiguro’s voice distinct from other JJK characters? Megumi sits lower and drier than Yuji or Gojo. His fundamental pitch is near-neutral male, with minimal vibrato, tight formant placement, and almost no dynamic swell on declarative lines. The flatness is intentional — stoic restraint rather than boredom. That deadpan quality is what makes his rare emotional breaks hit harder.
Which pitch shift and formant settings approximate Megumi’s voice? For male input targeting Yuma Uchida’s Japanese performance, start at –1 to –2 semitones pitch shift and neutral to –0.3 semitone formant offset. For Adam McArthur’s English dub, keep pitch near 0 and lower the formant slightly to –0.2 semitones. The key is the delivery pace: slow, deliberate, clipped consonants.
How much clean audio do I need to train an AI voice model for Megumi? A usable model needs 15–25 minutes of clean, isolated Megumi dialogue — no background music or sound effects. Include calm tactical lines, terse combat commands, and occasional emotionally charged scenes for range. Community pre-trained models exist on repositories like weights.gg if you want to skip training entirely.
Can I use a Megumi voice mod in online games without triggering anti-cheat? Yes, as long as the software routes audio through WASAPI rather than a kernel driver. Kernel-level audio tools can conflict with EAC, BattlEye, or Riot Vanguard. VoxBooster uses only the Windows WASAPI API — no kernel driver installation — so it coexists safely with competitive game anti-cheat systems.
Is it legal to use a Megumi Fushiguro AI voice clone for fan content? For personal, non-commercial fan use — gaming sessions, Discord roleplay, streaming — enforcement against fictional character voice clones is rare. For any monetized project, consult MAPPA and Shueisha/VIZ Media character usage policies before publishing. Never represent AI-generated content as official.
What is the difference between a real-time voice mod and a voice generator for Megumi? A voice generator takes text input and outputs a Megumi-like clip — useful for pre-recorded content but not interactive. A real-time voice mod converts your live microphone signal as you speak, which is what you need for Discord calls, in-game voice chat, or live streaming where you are responding spontaneously.
How do I match Megumi’s emotional restraint in my performance? Speak from the front of the mouth with short breath support — avoid projecting from the chest the way shonen heroes do. Drop the ends of sentences slightly rather than rising. Pause between clauses. When you hit an intense moment, resist the urge to shout: Megumi’s “intensity” is compressed tension, not volume.
Conclusion
Megumi Fushiguro is a voice impression target that rewards precision over drama. The acoustic profile is subtle — near-neutral male pitch with slight downward gruffness, minimal dynamics, dry tone, deliberate pacing — which means every deviation from that neutrality is audible. Getting it right requires both the correct DSP or AI model settings and an active performance choice to suppress the expressiveness that comes naturally to most speakers.
The combination that works best: a community AI voice model loaded with index influence around 0.75, a post-chain high-mid cut to increase the dry quality, and deliberate delivery practice focused on clipped endings and compressed intensity. DSP alone gets the register; the model adds the specific character. Your performance adds the restraint that makes both worth having.
If you want to hear what AI voice conversion does with your input before building a full setup, download VoxBooster and run the free trial with a community Megumi model. The whole workflow — install, import, route to Discord — takes under 10 minutes with a pre-trained model. Check the pricing page to find the plan that fits, starting at $6.99/month.