From Hum to Hook: How Melody and Chords Get Chosen for Your Lyrics

2026-09-23

How jhoom.baby turns one sentence into a finished song by choosing tempo, key, and mood — and why the same lyrics can sound completely different.

You type one line into jhoom.baby. Maybe it's "my cousin cried at her mehendi and we all laughed." A few minutes later, you have a song with lyrics and music you can play, share, and download. The lyrics are the visible part. The invisible part is everything underneath them: the tempo, the key, the chord progression, the mood. That's the part that decides whether your line feels like a slow dance or a wedding baraat. Here's how that part actually gets chosen.

Tempo, Key, and Mood: The Three Dials

When you submit a sentence, the service reads it for emotional signals before it writes a single note. Three dials get set first.

Tempo is speed, measured in beats per minute. A line about a farewell at the office — cake, cards, someone clearing their desk — pulls toward 70 to 90 BPM. Slow enough to sit with the feeling. A line about a friend's sangeet pulls toward 120 and up. The words "we danced till 2 a.m." will not get a ballad tempo unless you ask for one.

Key is the musical home base, and it carries emotional colour. Major keys tend to read as open, bright, resolved. Minor keys read as wistful, tense, or bittersweet. This is why the same lyrics can sound like two different songs. Give the line "I still keep your hoodie" a major key and it's nostalgic and warm. Give it a minor key and it's a wound. Neither is wrong. The service picks based on the dominant feeling it detects — or on what you tell it.

Mood is the broadest dial and it steers the other two. Words like "miss," "last," "goodbye," and "empty" push toward melancholy. Words like "finally," "together," "home," and "yes" push toward celebratory. Mood also shapes how the vocals are delivered: breathy and close, or full and forward.

What "Genre" and "Instrumentation" Actually Change

These two words do more work than people expect, and they're the easiest levers you have.

Genre isn't just a label. It's a bundle of rules. Choosing pop sets a verse-chorus-verse structure, a repeating hook, and a chord loop that stays put. Choosing folk often drops the drums, loosens the timing, and lets a guitar carry the whole thing. Choosing EDM builds toward a drop instead of a chorus. Your lyrics stay the same. The shape around them changes completely.

Instrumentation decides what you actually hear. Swap a piano for a harmonium and a line about your grandmother's kitchen changes temperature. Add a dhol and the same line about a cousin's mehendi becomes a procession. Take the drums out entirely and it becomes something you'd sing alone in a car.

Here's a concrete example you can try right now. Use this sentence: "My best friend is moving to another city and we said goodbye at the station."

• Version one: genre = acoustic folk, instrumentation = fingerpicked guitar and soft harmony vocals. You'll get roughly 75 BPM, a major key, and a warm ache.
• Version two: genre = cinematic pop, instrumentation = piano, strings, and a slow build. Same words, now it's 65 BPM, minor key, and it swells at the last chorus.

Two songs. One sentence. The difference is entirely in the dials.

How to Describe a Feeling So the Music Matches

Vague input gets a generic result. "Make it sad" gives you a default sad — the musical equivalent of a stock photo. Specific input gives you your sad.

Instead of an adjective, describe a scene or a physical detail. "It should feel like the last ten minutes of a long phone call" lands differently than "emotional." "Like the quiet after everyone leaves the party" tells the service more than "nostalgic" ever will.

Name the moment, not the emotion. Name who is in the room. Name what time of day it is. "A 6 a.m. train platform in winter" and "a rooftop at sunset in June" are both bittersweet, but only one of them gets a cold, sparse arrangement.

You can also just say what you don't want. "Not a happy song" is useful. "No drums" is useful. "Keep it under three minutes" is useful.

The service is doing the same thing a producer does in a studio: it hears your sentence, guesses the feeling, and builds a frame around it. The clearer your sentence, the closer the frame fits. And if it doesn't fit, you change one dial and try again — which is the whole point of getting a finished song in a few minutes instead of a few weeks.