Tiny Desk AI Voice & Sound
Your Tiny Desk AI video keeps the original stage song and singing voice. Uploading a different photo changes the lead performer’s appearance; it does not change the vocals, clone your voice or select a new singer.
Listen to the soundtrack first
This is the sound in the real example below. The built-in stage uses the same audio. Listening is free and does not use a generation credit.
If this voice is not right for your idea, check the alternatives below before buying a video.
One real input, one real generated result
The original stage is an original synthetic performance. It is not NPR concert footage and does not contain the “Outstanding” recording. Compare the original stage and finished take to see what changed.
Why does a male photo still have the same singing voice?
The portrait is a visual reference. In this workflow, the video is generated around that reference and the stage soundtrack is kept. A male or female portrait does not select a corresponding voice. Generating the photo again will not change that soundtrack.
There is currently no male/female voice selector, voice cloning, song picker or MP3 upload in the built-in stage mode. The singing-photo guide compares this workflow with tools that animate a portrait to an audio input.
What can I do if I want different sound?
| What you want | What you can do here | What to expect |
|---|---|---|
| Keep the preview’s song and voice | Use the original stage with one adult portrait. | A 15-second vertical video with that stage audio. |
| Download without sound | Choose “Download silent MP4” on the completed result. | The same video without audio; no new generation credit. |
| Use your own recorded performance | Open Advanced mode and upload a permitted 4–5 second MP4 or MOV with supported AAC audio. | A five-second edit retaining that clip’s audio; a silent source stays silent. |
| Change the song or clone a voice from a photo | Not supported in the current generator. | A photo alone does not provide a recording of someone’s voice. |
You can add permitted music to the silent download in a separate editor. That does not automatically make the generated lips match the new song. If lip sync matters, review the full performance and choose a workflow designed for your intended audio.
Why can’t I hear the preview or my result?
- Press play in the audio player above, or use “Click for sound” on the original stage. The stage video initially plays muted.
- Check the player volume, browser tab mute setting and your device’s output volume.
- For a completed video, play the regular result or use “Download video.” “Download silent MP4” intentionally has no audio.
- In Advanced mode, check whether your source clip has audio. The service does not add vocals to a silent source.
If the normal download is unexpectedly silent even though the source has supported audio, contact support with the request reference from My videos. Do not spend another credit just to troubleshoot playback.
Check these details before generating
Listen to the stage, inspect the actual result above, and choose one clear photo of a consenting adult. Review the face, hands, microphone and mouth movements in the finished video; preserving the audio does not guarantee perfect lip sync. The pricing page shows the current one-time packs. Free playback of examples is separate from paid generation.