The short answer
Use Singer when you already have a song and want to change its vocal character. Upload a voice reference and a target song, confirm you have permission to use both, then submit the cover task. Singer is not the tool for writing a new song from a text prompt.
Before you start
You need two separate inputs: a voice reference that represents the voice you want, and the target performance you want to transform. Use your own recordings or materials and voices you have permission to use. A downloadable recording is not automatically authorized for adaptation.
The studio accepts MP3, WAV, FLAC and M4A files up to 100 MB each. Keep an untouched copy of both inputs. Sign in and verify your email before submitting; check the current allowance and points estimate in the studio rather than relying on a fixed price in a tutorial.
1. Prepare the voice reference
Choose a recording where the intended voice is clear and consistent. Avoid several singers taking turns, loud background speech or heavily distorted audio. A reference with an audible room echo gives the model less clean information about the voice.
Listen to the reference once with headphones before uploading. If the vocal is buried under accompaniment, start with a cleaner authorized recording rather than repeatedly submitting the same noisy file.
2. Upload the target and submit
Open the Singer studio and put each file in its matching input: voice reference and target song. Wait for the target duration to be read. Review the allowance or points estimate, acknowledge the rights requirement and submit once.
Open the task center to follow the result. Processing time depends on queue and service availability; this guide does not promise an instant render. If submission reports a balance or verification issue, address that message before trying again.
3. Compare the result with the target
Compare the same verse and chorus rather than only the first few seconds. Listen for intelligibility, phrasing and how the transformed voice sits against the backing track. Keep the original next to the new version so you can distinguish an input problem from a transformation artifact.
- Check quiet phrases as well as loud notes.
- Inspect transitions into and out of the chorus.
- If a problem is already present in the target, improve the source before another attempt.
When to use a different tool
Use Create if you need a new song from lyrics and a musical brief. Use Lyrics if the main change is the words of an existing song. Singer focuses on vocal character; it is not a substitute for permission to use a person's voice, and it does not guarantee an exact identity match.
INPUT EXAMPLE — NOT A GENERATED RESULT
Example input plan
Voice reference: my own clean solo recording. Target: an original song I recorded and own. Review goal: compare verse clarity and chorus tone against the original.
This is a planning example, not a published audio result or a quality guarantee.
Frequently asked questions
Can I make a cover without a voice reference?
The Singer studio requires both a reference and a target. If you want to choose a preset vocal character for a new song, use Create instead.
Why is my task waiting?
A submitted task can wait for an available worker. Check its status in the task center before resubmitting; a second submission can create a second task.
Does generating a cover give me permission to publish it?
No. The tool does not grant rights to a song, recording or person's voice. Use authorized inputs and check the intended use before publishing.