All use cases

From Ideas to Complete Songs

MIDI, Lyrics, and the Making of an AI Vocal

Use a vocal synthesizer in ACE Studio to turn MIDI and lyrics into an editable sung part. Review notes, phrasing, and voices against your arrangement.

A vocal synthesizer is useful when you already know the melody and need it performed with lyrics. In ACE Studio, a vocal track combines an AI voice, MIDI notes, and words you can refine. This differs from asking ACE Music to develop a complete song from a broad idea: here, the vocal part itself is the editable focus.

Start with a song in ACE Music

For this example, Custom with ACE Music v1.1 Preview generated two songs from the original lyrics and Style below. We downloaded one complete MP3 and its MIDI score. The MIDI contains separate musical parts; inspect them rather than assuming every track is a vocal or that a score perfectly matches every detail of the audio.

[Verse]
The last bus turns around the square
A quiet window waits somewhere
My coat still smells of winter rain
I count the stops and start again

[Chorus]
Leave the light on, I'll be home
I don't have to find my way alone
Let the evening take its time
There is still a road from yours to mine

[Outro]
Leave the light on, I'll be home
There is still a road from yours to mine

In Create → Custom, enter the original English lyrics in Lyrics. The verse develops the scene and the chorus returns to its central thought.

In Create → Custom, enter the original English lyrics in Lyrics. The verse develops the scene and the chorus returns to its central thought.

Intimate English acoustic pop with a clear, singable lead melody, gentle piano, fingerpicked guitar, warm bass, and restrained drums. Keep the verses conversational and give the chorus a simple, memorable lift. Leave natural spaces between lyric phrases.

Describe the musical direction in Style, check the title and lyrics, then select Generate.

Describe the musical direction in Style, check the title and lyrics, then select Generate.

The completed versions are now available in Library. Select each one to compare how the same input develops.

The completed versions are now available in Library. Select each one to compare how the same input develops.

Choose MIDI in the song’s download dialog, then inspect the downloaded parts.

Choose MIDI in the song’s download dialog, then inspect the downloaded parts.

Listen to version 1

Prepare the melody as a singable part

Check that the MIDI represents the lead vocal you intend, not a piano reduction containing simultaneous accompaniment notes. Listen for note lengths, rests, and the position of important syllables.

Keep a copy of the original. If a melody was transcribed from audio, review it before assigning lyrics; a mistaken note can become harder to notice once a convincing singer performs it.

Build the vocal in ACE Studio

1. Load a voice and prepare a short clip

Open ACE Studio's Voice Library, choose an available AI voice, and add it to the Canvas. Work on a vocal track with a clip containing your intended melody. Add or import a monophonic lead line rather than using a piano part whose simultaneous notes represent several instruments.

Begin with a short passage. An original exercise is “Leave the light on, I'll be home,” with the final word held longer than the opening ones. The exercise is a composition example, not a tested vocal preset. Choose pitches appropriate to the part rather than assuming that sentence supplies a melody by itself.

In ACE Studio, we imported only the first 17 notes from the downloaded Vocal Melody track, with the initial empty time removed. We supplied the first two original lyric lines, including a held syllable. The accompaniment parts were not assigned to the singer.

ACE Studio — Emma, same lead excerpt

ACE Studio — Emma, same lead excerpt

2. Align the words with the musical phrase

Enter the lyric using the line below the notes, or select notes and use the lyric-editing panel. Apply the text, then inspect the actual syllable placement. Words that read as one phrase may occupy several notes when sung.

Check which note receives the important word and where a vowel continues across more than one pitch. Do not add a new syllable merely to fill every note. Conversely, a long sentence needs enough musical space; extra wording can be lost or assigned somewhere you did not intend.

The lead-only MIDI excerpt has English lyrics aligned to the notes; the held ending uses a tenuto.

The lead-only MIDI excerpt has English lyrics aligned to the notes; the held ending uses a tenuto.

3. Audition and correct the part before decorating it

Play the phrase and compare it with the lyric and backing. If “home” arrives too early, change the underlying note timing or syllable placement. If there is no room to breathe, create musical space rather than trying to disguise the problem with effects.

Hear the line at its place in the arrangement. A convincing solo voice may still put a consonant on a drum accent or cover a guitar hook. Work on the musical interaction, not only how smooth the voice sounds alone.

4. Compare voices with the same material

Keep the notes and lyric stable while trying another suitable voice. Listen to the same low phrase, high phrase, and sentence ending at a comparable level. Write down the musical reason for the choice, such as clearer words or a more comfortable register.

Only after those choices work should you refine pitch or expression for an important word. This gives detailed editing a purpose. A curve cannot reliably compensate for an incorrect melody or a lyric assigned to the wrong notes.

The official Creating AI Vocals guide links the relevant note and lyric controls. Keep the editable project alongside any listening file you prepare. That way, a later lyric change can be made to the part instead of reconstructed from a mixed recording.

ACE Studio — Elirah, same lead excerpt

ACE Studio — Elirah, same lead excerpt

Real English audio examples

The two Studio files below use the same notes, words and export range of about 10.27 seconds. Only the selected voice changed from Emma to Elirah; the project effect settings were kept. They are new synthesized vocal excerpts, not isolated vocals extracted from the MP3.

Shape a singing synthesizer's performance

A MIDI-to-vocals workflow lets you work on the part directly. Adjust a crowded note, shorten a syllable, or refine a held ending and compare the phrase in context.

An AI vocal generator can sound convincing while still placing emphasis on the wrong word. Keep the lyric beside the piano roll and judge the sentence as speech as well as melody.

Use ACE Music for the broader song when appropriate

If your goal is to explore a full arrangement around a sketch, start with ACE Music's supported melody-reference workflow. If you want to create a separate singer for a melody you have already composed, use ACE Studio's note-and-lyric route.

Moving between products does not mean a mixed audio file automatically becomes a fully editable vocal score. Prepare or transcribe the part you need, then review that transcription.

Compare voices using the same passage

Keep the MIDI, lyric, and listening level consistent when auditioning singers. A voice that sounds impressive in a preset demo may not suit the register or diction of your melody.

Before completing the production, hear the vocal with the backing track. Check consonants, phrase endings, and whether the performance leaves room for the other musical parts.

Use ACE Studio when your song needs an editable sung performance built from notes and lyrics, with ACE Music available as the broader song-creation starting point.

Questions about MIDI-based singing

Does a MIDI file already contain the lyric?

Not necessarily. Review the melody and supply or check the words in the vocal clip. Treat the source file and sung text as separate materials until you have verified their alignment.

Can I use a full piano arrangement as one vocal line?

Prepare the intended singing part first. Simultaneous accompaniment notes are not automatically a singable lead melody, even if they are in the same MIDI file.

Is this full-song generation?

This workflow builds an editable vocal part. Use ACE Music when the starting goal is to explore a complete song from an idea; use ACE Studio when notes and words need direct production control.