How to Change the Voice in a Video

Most people who ask this have a finished clip: a gaming video, a VTuber update, a TikTok, a school project. They don't want to record it again. They want the same words in a different voice.

What do you need to change the voice in a video?

• The video file. MP4 is the safest. Our Online Voice Changer reads the sound straight out of an MP4, up to 10 minutes long.
• A voice changer that takes files. Ours is free, runs in Chrome, Edge, Firefox and Safari, and processes the file on your own computer, so the video isn't uploaded.
• Something to put the sound back. Any video editor, or ffmpeg if you're happy with a command line.

How do you do it, step by step?

1. 01

   Open the video in the voice changer

   On the Online Voice Changer, click Upload audio and choose your MP4, or drag it in. You'll see its name and length, e.g. “voice-voiceover.mp4, 0:17”.

2. 02

   Pick a voice

   Click a preset: Deep Voice, Robot, Announcer, Girl, Monster and the rest, 26 in all. The file is changed to that voice right away. The Fine-tune sliders (pitch, tone, robot, echo, reverb) adjust it.

3. 03

   Listen and compare

   Press play and switch between Original and Changed under the player. After moving a slider, click Apply again.

4. 04

   Download the WAV

   Click Download WAV. You get audio only, at the same length as your video's sound.

5. 05

   Put the sound back under the video

   In your editor, mute or delete the video's own audio track, drop the WAV in at 0:00, and export. Or use the ffmpeg command below.

What does it sound like?

Our test clip is a 17-second VTuber update read by an AI voice from our Character Voice Generator, which took 4.3 seconds to make. Here it is before and after three presets:

How long does it take?

So a long video is fine on a laptop: nine minutes took sixteen seconds. The catch is size. The download is an uncompressed WAV at 48 kHz, about 5.8 MB a minute in mono, so a 9-minute file came out at 52 MB. Your editor compresses it again when you export the video, so the finished file isn't bigger.

Will the voice stay in sync with the lips?

Yes, with voice effects. They change the sound, not the timing, so every word lands where it did. Our Deep Voice and Robot files came out 0.05 and 0.1 seconds longer than the 17.2-second original, and the extra bit is the effect fading out after the last word. Announcer adds room echo, so its tail was 0.6 seconds. An editor, or ffmpeg's -shortest, cuts the tail to the video's length.

How do you put the new voice back with ffmpeg?

If you have ffmpeg, this keeps the picture as it is and swaps in the new sound. It took under a second on our 17-second clip:

ffmpeg -i video.mp4 -i changed.wav -map 0:v -map 1:a -c:v copy -c:a aac -shortest video-new-voice.mp4

What if you want a completely different person's voice?

Effects bend your own voice: deeper, higher, metallic, on the radio. Listeners can still hear it's the same speaker underneath. If you want the clip to sound like someone else entirely, an AI voice changer re-speaks it in a new voice. Ours, the AI Voice Changer, takes recordings of up to 5 minutes, runs on our servers and uses credits; the voices are original characters.

What are the limits?

• 10 minutes per file. Split a longer video's sound into parts, or export the audio and trim it first.
• Audio only out. The tool gives you a WAV, not a new video, so you need the last step in an editor or ffmpeg.
• Music and voice change together. If your video has background music, it gets the effect too. Change a clean voice track if your editor has one.
• Not tested on phones. Our timings are from an Apple M4 laptop. The tool works on phones, but long files take longer there.