Feature
Voice over for AI avatar video: record it, generate it or hire it
An AI avatar needs a voice-over, and its mouth movements are made from that audio. You can record your own for free, generate one from your script, or hire a voice actor.
In short
- You have three main options for an AI avatar voice-over: record your own, generate one from your script, or hire a voice actor.
- Your own voice is free and sounds like you, but echo from the room you record in hurts the lip sync.
- A generated voice costs 70 credits a minute in Cutroom, about 47 credits for a 40-second ad.
- A hired voice actor gives you a real performance, but you pay the actor's fee and wait for the recording.
- Some tools can also clone your voice from a sample, such as HeyGen from its Creator plan. Cutroom does not clone voices.
- Credits per minute of generated voice-over in Cutroom
- 70
- Recording your own voice
- Free
- Voice-over length a Cutroom avatar accepts
- 3 s to 5 min
The three voice-over options compared
The voice-over drives the avatar's mouth, so choose it before anything else.
- A Cutroom avatar accepts a voice-over of 3 seconds to 5 minutes, and the finished ad is a vertical 9:16 video.
- Avatars need the Basic plan at $39.99 a month for 2,500 credits.
Voice-over options for an avatar video
| Option | Cost | Best for | Tradeoff |
|---|---|---|---|
| Your own voice | Free | Ads with your own face, and lines that need feeling | Echo from the room you record in |
| Generated voice | 70 credits a minute in Cutroom, about 47 for 40 seconds | Offers, announcements and product facts | A very even tone, which shows more in long videos |
| Hired voice actor | The actor's fee, which the actor sets | One important ad, like a launch ad that will run for months | You wait for the recording, and each change goes back to the actor |
Option: Your own voice
- Cost
- Free
- Best for
- Ads with your own face, and lines that need feeling
- Tradeoff
- Echo from the room you record in
Option: Generated voice
- Cost
- 70 credits a minute in Cutroom, about 47 for 40 seconds
- Best for
- Offers, announcements and product facts
- Tradeoff
- A very even tone, which shows more in long videos
Option: Hired voice actor
- Cost
- The actor's fee, which the actor sets
- Best for
- One important ad, like a launch ad that will run for months
- Tradeoff
- You wait for the recording, and each change goes back to the actor
Option 1: record your own voice
Your own voice is free and sounds alive. It also picks up the sound of the room you record in.
- Record in a soft room with carpet, curtains or a wardrobe full of clothes. Kitchens, bathrooms and bare offices echo.
- Record speech only, with no music or background sound under it.
- Talk at about 150 words a minute, a normal speaking pace.
- Leave a short pause at the start and the end, so the first and last words are not clipped.
- Save the file in good quality as an MP3, WAV, M4A or OGG. Low-quality files lose sounds like p, b and m, where the lips close.
- Listen to it on its own before you make the avatar video. If the words are hard to make out, the lip sync will be off too.
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Option 2: a generated voice
Cutroom can turn your script into a voice-over for 70 credits a minute. You pick a voice from a list, and the audio comes back with no room sound at all.
- It has no echo, no hum and no background noise, so the lip sync gets clean audio.
- You do not have to record anything. You paste the script, pick a voice from the list and Cutroom makes the audio.
- The tone is very even. That is hard to notice in a short ad and easier to notice in a long one.
- It suits offers, announcements and product facts, where the information matters more than the voice.
- ElevenLabs is a well-known text-to-speech tool on its own. Its free plan does not allow commercial use, and Starter at $6 a month does.
Credits for a 40-second avatar ad in Cutroom
The voice is a small part of the cost. The avatar video is the big part.
- Your own recorded voiceFree
- Generated voiceAbout 47
- The avatar video itself, at 350 credits a minuteAbout 234
Option 3: hire a voice actor
A voice actor gives you a performance: stress on the right word, changes of pace and pauses where they matter.
- You can hire one on marketplaces such as Voices.com or Fiverr.
- On Voices.com, posting a job and hearing auditions is free. The actor sets the fee, and Voices adds a processing fee of 4 to 6%.
- For a paid ad, the fee depends on where the ad runs and for how long.
- Choose a voice actor when one ad matters a lot and the stress and pauses have to be exactly right, like a launch ad that will run for months.
- Ask for a dry recording in high quality, with no music, no effects and no room echo.
Can you clone your voice for an avatar?
Voice cloning copies your voice from a recorded sample, so a tool can read any script in your voice. Cutroom does not clone voices.
- HeyGen offers voice cloning from its Creator plan, $29 a month or $24 billed yearly.
- ElevenLabs offers Professional Voice Cloning from its Creator plan, $22 a month.
- In Cutroom, you record your own voice for free, or pick a generated voice from the list.
- On TikTok, an ad with a cloned voice needs TikTok's AI-generated label or a clear disclaimer.
How the voice-over affects the lip sync
The avatar's mouth is made from the sound alone. The photo tells the AI nothing about how the mouth should move.
- A clean generated voice often lip-syncs better than a real voice recorded in an echoey room.
- Pick a voice that fits the person in the photo, in age and in tone. A deep, older voice on a young face looks wrong.
- Make a 20-second test video before you pay for the full length: about 117 credits with your own voice, or about 141 with a generated voice.
Which voice-over should you choose?
It comes down to the script and the room you have to record in.
Which voice-over should you choose
Choose your own voice if the ad shows your own face or brand, and you have a quiet, soft room to record in.
Choose a generated voice if the ad is an offer, an announcement or product facts, or you have no quiet room.
Choose a voice actor if one ad matters a lot and the stress and pauses have to be exactly right, like a launch ad that will run for months.
Choose filming instead if the script is a personal story or a product demo. An avatar cannot tell a personal story believably or use the product, whatever voice it has.
Questions people ask
- Does a generated voice lip-sync better than my own?
- Usually, unless you record in a quiet, soft room. A generated voice has no echo or background sound, which is exactly what the lip sync needs.
- Can the generated voice speak other languages?
- Yes. Cutroom's voice model also reads scripts written in other languages, such as Dutch or Spanish. You pick a voice from the list and paste the script in that language.
- Can I clone my own voice?
- Not in Cutroom. HeyGen clones voices from its Creator plan, and ElevenLabs from its Creator plan at $22 a month. In Cutroom you record your own voice for free or pick a generated one.
- Can I use the audio from a video I already made?
- Yes, if it is clean speech with no music or background sound under it. Save the sound as an MP3, WAV, M4A or OGG file first, because Cutroom takes an audio file, not a video.
- How much does the voice-over add to the cost?
- Not much. In Cutroom, a generated voice for 40 seconds is about 47 credits, against about 234 for the avatar video and 100 for the edit.
- Can I reuse a voice-over with another avatar?
- Yes. Save the audio file and upload it again with another photo. You only pay for the new avatar video.
- Which Cutroom plan includes avatars?
- Basic at $39.99 a month or higher. Lite and the free first video do not include avatars.
Record clean speech in a soft room, or generate it from your script. Avatars come with Basic at $39.99 a month, and your first filmed video is free.