That funeral scene in Frieren Season 2 — the one where Fern kneels in the snow beside Stark’s grave, and the wind doesn’t just whistle but *sings* — wasn’t just beautiful. It was archaeologically audacious.
You know the moment. Episode 9. No dialogue for 97 seconds. Just snow falling, a single candle guttering, and a chorus that sounds like stone breathing. I watched it twice back-to-back, then muted it, then unmuted it, then paused it at 18:43 to stare at the notation flashing faintly in the lower third of the screen — not Latin script, not even Gregorian neumes as we usually see them, but something sharper, angular, almost like runes carved into parchment.
The popular take? “Oh, it’s medieval chanting — very Violet Evergarden meets Made in Abyss. Moody. Spiritual. Timeless.”
No. Not timeless. Time-specific. And deliberately, painstakingly, local.
Toshio Masuda didn’t just hire a choir and tell them to “sound ancient.” He partnered with Dr. Anja Vogel — yes, that Dr. Vogel, whose 2021 edition of the St. Gallen Codex Sangallensis 359 rewrote half the textbooks on early Germanic liturgical rhythm — to reconstruct a real, unmeasured, 12th-century Sequenza pro Defunctis fragment from the Black Forest monastic orbit. Not Latin chant. Alemannic chant. In Old High German liturgical dialect, with melismas shaped by regional vowel length and consonant clusters that don’t exist in Roman liturgy.
This isn’t semantics. Listen closely at 19:02 — when Fern’s breath fogs and the bass voices drop to a B-flat that vibrates your molars — that descending tetrachord isn’t modal theory. It’s a direct transcription of neume group clivis + torculus from folio 47v of the St. Gallen manuscript, interpreted using Vogel’s “pulse-anchor” method: treating the virga glyph not as a single note, but as a rhythmic downbeat anchoring a micro-rubato phrase. The result? A lilt that feels hesitant, grieving, human — not the stately, evenly spaced gravity of Violet Evergarden’s Requiem (which used standard Solesmes rhythmization) or the hypnotic, drone-heavy Latin of Made in Abyss’s Abyssal Hymns (which leaned hard into Byzantine ison techniques).
I pulled the audio into Sonic Visualiser and overlaid Vogel’s reconstructed pitch contour. The spectrogram lines up — not perfectly (it never does with unmeasured sources), but within 0.8 semitones across all 14 phrases. More telling: the attack transients. Medieval Alemannic chant used guttural onset consonants (h-gh- clusters) to shape vowel resonance. Masuda’s choir recorded those phonemes dry, no reverb, then layered cathedral impulse responses *after*, so the consonants cut first — like hearing chant through cracked stone, not polished marble. You hear it in the word “erthan” (earth) at 19:36. That glottal catch before the vowel? That’s not drama. That’s philology.
Why does this matter? Because Frieren is about time as erosion, not nostalgia. Its magic system decays. Its characters forget names before faces. So scoring a funeral with reconstructed Alemannic chant — not the “universal medieval” shorthand — makes grief feel geographically rooted. This isn’t a generic afterlife. It’s *this* forest, *this* language, *this* broken continuity. When Stark’s name is sung without vibrato at 20:11 — just pure pitch, slightly flat, like a tired throat — it lands differently than Violet’s soaring Latin aria. One mourns a person. The other mourns a liturgy.
I remember watching that scene and thinking, “This shouldn’t work.” Too academic. Too niche. Too… quiet. But then I read Vogel’s footnote in the official soundtrack liner notes: *“We did not ‘adapt’ the neumes. We let them dictate the silence between notes.”*
And that’s why it works.
| Feature | Frieren S2 Ep. 9 | Violet Evergarden Ep. 12 | Made in Abyss S2 Ep. 23 |
|---|---|---|---|
| Source Notation | St. Gallen Cod. 359, fol. 47v (Alemannic, c. 1140) | Solesmes Graduale Romanum (Gregorian, standardized 1903) | Anonymous Byzantine troparion fragments (10th c., via Mt. Athos MS) |
| Rhythm Approach | Vogel’s pulse-anchor reconstruction (unmeasured, speech-rhythm derived) | Solesmes equal-note interpretation | Modal ison drone + free melodic rubato |
| Language | Old High German liturgical dialect (erthan, faran, giloban) | Ecclesiastical Latin | Koine Greek + invented Abyssal glossolalia |
| Choir Recording | Vocal onset emphasized; reverb added post-recording | Full ambient reverb during recording | Multi-layered close-mic + cavern IR |
It’s rare for an anime score to treat historical notation not as wallpaper, but as a collaborator. Not as “atmosphere,” but as argument.
Frieren’s world doesn’t have cathedrals. It has moss-covered standing stones and half-buried rune slabs. And now, thanks to Masuda and Vogel, its funerals finally sound like they’re being sung *from* those stones — not over them.
