Testissä MiniMax Music 3

Testasin MiniMax Music 3 -musiikkimallia, joka tekee tekstikuvauksesta ja sanoituksista valmiin kappaleen — laulut, soittimet ja miksauksen, kaikki kerralla. Malli pyörii omalla koneella ComfyUI:n kautta, eli mitään pilvipalvelua tai kuukausimaksua ei tarvita.

Miten se toimii

Työnkulku on yksinkertainen. Ensin täytän lyhyen kyselyn (Fable kielimalli oli tänään käytössä): genre, tunnelma, tempo, laulaja, soittimet, tuotannon aikakausi ja kesto. Yksi tai kaksi sanaa per kohta riittää, ja kohdan voi jättää tyhjäksikin. Sitten kielimalli laajentaa vastaukset ”captioniksi” — parin sadan sanan tuotantokuvaukseksi, joka kertoo mallille millainen biisistä pitää tulla: miltä rummut kuulostavat, milloin kitarasoolo tulee, minkälainen laulaja on äänessä. Lopuksi caption ja sanoitukset liimataan ComfyUI:hin, ja hetken päästä ulos tulee valmis kappale, enimmillään viisi minuuttia pitkä.

Sanoitukset voi kirjoittaa itse — malli laulaa ne sanasta sanaan, suomeksikin. Tosin lausunto on hieman Hullujussia välillä. Caption kirjoitetaan englanniksi, koska malli ymmärtää musiikkisanastoa parhaiten sillä.

Alla kolme hyvin erilaista lopputulosta ja niiden luomiseen käytetyt Captionit.

Vaiti

Suomalainen laulelma, jossa on kevyt calypso-keinunta. Juicen sanoitukset, sointukiertokin annettu kyselyssä (G#-duuri hookina mollimaisemassa).

A calm, serene Finnish laulelma with a gentle calypso lilt, around 100 BPM in C minor, built on a circling progression of Cm, Fm and G, where a warm G# major chord returns again and again as the song’s harmonic hook — a brief lift of sunlight in the minor-key water.

Genre tags: Finnish Laulelma / Acoustic Calypso / Folk Ballad.

The song stays emotionally even and unhurried, like still water in the morning: it opens in quiet acceptance as a boat slips from the dock, swells gently and warmly in the middle verses where memory turns tender and sensual, then settles back down and glides into a hushed, contented homecoming. Perfect for a summer evening at sea or a slow archipelago film scene, evoking imagery of a small boat pulling away from a wooden pier, a figure waving on the dock, a long wake unfolding across a sunlit bay, and islands drifting past in warm haze.

The production is modern acoustic and clean: an organic, naturally wide stereo image, warm and uncluttered low end, airy but never brittle highs, gentle transparent compression — polished yet human, every instrument clearly placed and given room.

Vocal Gender & Timbre: a soft, gentle male voice, light and tender, almost spoken in the low phrases, completely unforced.

Vocal Style: patient, storytelling phrasing that floats slightly behind the calypso sway, rising only a little in the middle verses where the memory grows warm.

Harmony/Backing Vocals: sparse — a quiet, close harmony doubling only the repeated ”Vaiti, aivan hiljaa” chorus lines.

Vocal FX: a light, natural room reverb and nothing else; the voice stays intimate and true.

Arrangement: piano is the harmonic heart, opening the song alone with a two-chord vamp and carrying warm, rounded voicings throughout; a soft round bass enters with the first verse, walking gently under the calypso sway; light calypso percussion — rim clicks, congas or bongos, and a soft shaker — establishes an easy rolling groove that rocks like a hull in low waves. In the middle verses the percussion fills out subtly and the piano voicings grow richer for the gentle swell, then everything thins again for the final chorus.

Embellishments, Textures & Spatial FX: occasional high piano sparkles like light on water, a distant gull-like ambience far back in the mix, and a slow fade on the outro as the boat reaches harbor.

Puutarhassa

Hidas, leijuva jazzballadi — kuiskaava naisääni, saksofoni liittyy mukaan vasta keskivaiheilla. Kysely: ”jazz ballad flavor / hushed → sensual swell → dissolve / very slow / duet”. Tommy Tabermanin runo.

Global Metadata.
Basic Attributes: A very slow, floating jazz ballad, rubato and almost pulseless, in a dark, warm minor tonality.

Genre tags: Jazz Ballad / Vocal Jazz Duet / Finnish Art Song.

Global Emotional Progression: The piece opens hushed and suspended, two voices barely above a whisper in a dark room, all breath and nearness. Through the middle stanzas the intimacy deepens into a slow sensual swell — never loud, only warmer and closer — as the voices begin to overlap. At its peak the music is dense with closeness rather than volume, and in the final minute it dissolves: instruments fall away one by one until only breath, bass and room remain.

Application Scenarios & Imagery: Perfect for the last hour of the night, evoking imagery of an empty jazz club at 2 a.m., two figures in the half-dark, smoke curling through a single lamp, a slow river moving unseen beneath the floor.

Sonics & Production Profile: An intimate late-night club recording — small room, close mics, the singers almost touching the diaphragm, audible breaths and lip sounds kept in. Narrow, warm soundstage with the voices centered and slightly interleaved, natural room reverb only, soft analog warmth, generous low-end roundness, no polish or loudness.

Vocal Details.
Vocal Gender & Timbre: A duet — a breathy, airy female voice and a deep, warm male voice, air against earth.

Vocal Style: Very slow, speech-like phrasing that floats free of strict time; the male voice carries the opening stanza, the female voice answers with the second, and from the third stanza onward the two entwine, overlapping and finishing each other’s phrases.
Harmony/Backing Vocals: No separate backing vocals; the harmony is the two leads weaving in close intervals during the later stanzas.

Vocal FX: None — dry, natural voices colored only by the room.

Arrangement.
Instrument Lifecycle: Upright bass and brushed drums are present from the first second to the final fade — the bass round and unhurried, the brushes more texture than timekeeping, slow circles on the snare. A breathy tenor saxophone enters only at the sensual swell, answering the voices between phrases, then slips away for the dissolve.

Groove & Foundation: No fixed groove — a floating rubato held together by the bass’s long tones and the continuous whisper of brushes, like water moving under everything.

Embellishments, Textures & Spatial FX: Faint room creaks and breath as texture, an occasional soft cymbal wash panned wide, and a long final decay where bass and brushes fade into the noise floor of the room.

Kaarila tanssii ja soi

Ja sitten jotain ihan muuta: 70-luvun manserock-shuffle, Love Records -henkinen soundi, Hammond-urut ja kitarasoolo. Sanoitukset muokattu Juicelta, vaihdettu Kuopio Kaarilaan ja Tampere aiheisia lyriikoita lisätty.

Global Metadata.
Basic Attributes: A warm, good-natured 70s Finnish rock shuffle — classic manserock — at a relaxed medium tempo in the 95-110 BPM range, swinging rather than driving, built on a friendly major-key three-chord foundation with a bluesy seasoning.

Genre tags: Manserock / 70s Finnish Rock / Shuffle Rock.

Sunny from the first downbeat to the last — a contented, grinning stroll that never breaks a sweat. The verses amble along with a wry storyteller’s calm, each chorus opens the doors a little wider into a loose communal singalong, the bridge leans back to admire the view, and after the guitar solo the final chorus lands as the biggest, happiest crowd moment of the night without ever losing its easy swing.

Application Scenarios & Imagery: Perfect for a light Nordic summer evening or a neighbourhood dance night, evoking imagery of a lakeside industrial city at dusk, an open-air dance floor in a working-class suburb, denim jackets and long hair bobbing in time, someone’s transistor radio carrying the same song across the water.

Sonics & Production Profile: A warm early-70s analog LP sound, like a Finnish rock record cut live in one room — big round bass and thick dry drums right up front, gentle tape saturation, a swirling organ bed glued into the mix, short plate reverb on the vocal only; unpolished in the friendliest way, all mid-range warmth and glue-compressed togetherness.

Vocal Details.
Vocal Gender & Timbre: A single male lead — a warm, relaxed storyteller’s voice with a dry wit and an audible half-smile, more talker than belter.

Vocal Style: Laconic, conversational verse phrasing that sits slightly behind the shuffle, tossing off punchlines deadpan; in the choruses the melody widens and rises and he leans in with real joy, riding the swing instead of pushing it.

Harmony/Backing Vocals: A loose, good-humoured gang of male voices joins the chorus hooks in rough unison and easy thirds — friends around one microphone rather than a rehearsed choir — swelling bigger with every chorus until the last one is a full singalong.

Vocal FX: Warm plate reverb and a touch of tape slap on the lead; backing voices drier and slightly off to the sides; otherwise natural.

Arrangement.
Instrument Lifecycle: A fat, melodic bass guitar and a heavy, laid-back drum kit dominate from the very first bar and never leave. A Hammond organ — with a warm Wurlitzer-like electric-piano colour in the quieter moments — pads and stabs underneath throughout, while a clean electric guitar comps easy chords and short licks between vocal lines. The guitar steps forward only for the solo section, singing and bluesy over the churning rhythm section, then tucks back in. Tambourine and handclaps join the final chorus as everything plays at once.

Groove & Foundation: A rolling medium shuffle — swung eighths everywhere, the bass walking and bouncing between root and fifth, kick drum heavy and round, snare fat and dry on the backbeat with lazy triplet fills into each chorus; the whole band breathes together like one instrument.

Embellishments, Textures & Spatial FX: Organ swells rising under the ends of chorus lines, cheeky little guitar answers after the vocal punchlines, a drawbar-flick organ flourish into the bridge, claps and tambourine saved for the last chorus, and a final held band chord with the organ spinning slowly out on top.

Lopuksi

Tykkään siitä, kuina ohjausmahdollisuuset alkaa lisääntymään. Sanoituksissa voi käyttää seuraavia osia, jotka yhdessä Captionin kanssa antaa mahdollisuuden suhteelliseen tarkkaan tuotantoon:

[Intro]
[Verse]
[Pre-Chorus] — build-up lines before the chorus
[Chorus]
[Post-Chorus] — a hook/refrain right after the chorus
[Bridge] — contrasting section, usually before the last chorus
[Instrumental] — a section with no singing
[Solo] — instrumental solo section
[Outro]

Jos ei muuta, niin hauskaa tämä on.

ps. täytin vuosia ja pyysin minimaxin tekemään mulle synttärikappaleen. Tämän voi laittaa kokonaisuudessaan.