Find a picture in the lyric
Choose a portrait whose setting fits the song. A singer beside an empty kitchen table can suit a lyric about leaving home. Include the setting in your uploaded photo and keep the face clearly visible.
Give your song a face. Upload an audio clip and a portrait, then generate a singing video with lip sync. Preview the result and download it from the same workspace.
Choose the detail a listener might picture after the song ends. It could be a face, a familiar room or the way someone moves on the first beat.
Choose a portrait whose setting fits the song. A singer beside an empty kitchen table can suit a lyric about leaving home. Include the setting in your uploaded photo and keep the face clearly visible.
Start with a portrait taken in the room, stage or street you want on screen. The singing generator uses one image, so prepare the composition before uploading it.
A clear lead vocal gives the model a more useful performance to follow than overlapping voices. Try a short line first, then review the mouth movement and expression before using a longer passage.
For a sung line, frame the singer close enough to see their mouth and use a passage where the words come through clearly. Avoid photos where a microphone, hand or deep shadow hides the lips.
Start with a short vocal excerpt. Audio files can be 5 to 180 seconds long; the uploaded audio determines the video length and credit cost.
Sign in and upload an MP3, WAV, or M4A file under Audio. Trim your recording before uploading if you only want a chorus or verse. Leave Guide with transcript off for singing.
Upload one photo under Character. Use a clear face with the mouth visible. The photo supplies the person, setting and framing; no scene prompt or extra reference image is needed.
Choose a resolution and review the credit cost before generating. Follow the task in the result panel, then play and download the finished video. Your saved videos are also available in Assets.
Play the passage once for the performance and once for the picture. Small adjustments are easier to spot when you focus on one at a time.
Follow the first syllable and the pause at the end of the line. If the mouth moves through the silence, check where the vocal starts in your chosen excerpt.
Look at the singer at the intended viewing size. If a microphone covers the mouth, choose a portrait with the microphone lower before making another take.
For Reels or Shorts, start with a vertical 9:16 frame and keep the face away from the bottom controls. A wide 16:9 frame gives a band or a landscape more room.
Play past the final word. If your source excerpt cuts off a held note, adjust it in your audio editor before uploading a new clip.
Choose a rehearsal photo with the face clearly visible, add a vocal clip and bring the portrait into the performance.
Use these ideas when choosing or preparing your source photo. The uploaded photo sets the scene; the singing generator does not take scene prompts or camera directions.
A singer stands on a quiet train platform under a yellow station lamp, with the face well lit and turned toward the camera.
A waist-up portrait at a green kitchen table, with two cups of coffee and soft window light across the singer's face.
A singer faces the camera in a rehearsal room, with drums in the background and the microphone below the mouth.
A close portrait beside a parked car, with city lights behind the performer and enough light to make the lips clearly visible.
A chest-up portrait in a vertical frame, with a red theater curtain behind the singer and no objects covering the face.
One guitarist sits by a rehearsal room window. The instrument gives the photo context while the face remains the focus.
Using your own recordings, working with reference photos and keeping the performance in sync.
Start with the line your listeners sing back. Upload that vocal clip and a portrait to create a singing performance.