Can I convert my own voice to whatever I like?

Your own voice as the source needs nobody's permission — that half is always simple. The target is a separate question: it is somebody's voice too, and the same rule applies to it as to cloning it. You may only clone a voice you have the right to use — your own, or someone else's with their explicit, informed permission. Nothing about the upload can verify whose voice it is, so this is entirely on you.

Can I make myself sound like a famous singer or actor?

No. Somebody's voice, usually somebody's livelihood, and frequently a protected performance. That a clean recording is easy to find is not a permission, and a disclaimer in the description is not one either. A credit in the description does not change any of that, and neither does calling it a parody.

What if I only want to anonymise someone?

That is one of the better uses of this tool, and it still needs their agreement — converting a voice to protect a speaker is something you do WITH them, not to them. Tell them which voice you are using and let them hear the result; they are the only person who can say whether it still sounds like them.

So what is it actually for?

Dubbing and re-voicing your own material, playing several characters yourself, holding one voice consistent across a long project, and letting somebody speak in their own words and be heard in another voice. Four real uses, all of them ones where the consent question has a clean answer.

Voice Changer

Whose voice can you use

Q: I recorded an interview. Can I convert the other person's voice?

Not without asking them. Conversion involves two people, not one. The target voice is somebody, and so is whoever is speaking in the recording you upload — and their permission is a separate question from the target's. Recording a call, an interview or a meeting is not permission to put those words into another voice, and nothing about the upload asks you whether you have it. This is the case people most often get wrong, because permission to record feels like permission to use — and the two are not the same thing at all.

Conversion involves two people — the speaker in your recording and whoever the target voice belongs to — and the tool checks neither. What it is honestly for, and what it is not.

Lesson 3 · core · 7 min read

Open Voice Changer

This is the lesson worth reading before you run anything. AudioPod's voice conversion takes one person's speech and makes it come out of another person's voice, which means every job you run involves two people and at most one of them is you. Neither of them is asked.

Careful

Nothing in this tool checks any of this. There is no box to tick, no attestation, no warning on the tab and nothing in the upload that can tell whose voice is in a file. Our own marketing calls the feature consent-based; that describes what we expect of you, not something the software verifies. The whole of the obligation is yours.

Two people, not one

Cloning a voice raises one question: may I reproduce this person? Conversion raises two, and the second is the one people miss.

The target voice
You may only clone a voice you have the right to use — your own, or someone else's with their explicit, informed permission. Nothing about the upload can verify whose voice it is, so this is entirely on you.
The speaker in your recording
Conversion involves two people, not one. The target voice is somebody, and so is whoever is speaking in the recording you upload — and their permission is a separate question from the target's. Recording a call, an interview or a meeting is not permission to put those words into another voice, and nothing about the upload asks you whether you have it.

Tip

If the target voice is a clone you made, the cloning rules apply to it in full — how to make one, and whose voice you may make one from, are taught in the Voice Studio track at /guides/voice. This lesson is about the second question, which that track does not have.

The cases, plainly

The situation: Where it stands
Your own voice, converted to a voice you cloned from yourself: Go ahead. Both people are you. This is the case the tool is really built for, and it is the only one with nothing to check.
Your own voice, converted to a voice somebody licensed to you: Go ahead. Fine, within whatever that licence says. Read what you were actually granted — permission to use a voice in a product is not always permission to use it in an advertisement.
Your voice, converted to a voice you cloned from someone else: Not without permission. The target's permission is required, and it has to be permission for a synthetic voice specifically. Somebody agreeing to be recorded has not agreed to be reproduced.
Somebody else's recording — an interview, a call, a meeting: Not without permission. The speaker's permission is required as well as the target's, and it is the one people forget. They agreed to be recorded saying that. They did not agree to be heard saying it in a stranger's voice.
A public figure, an actor, a singer, a character from a film: No. Somebody's voice, usually somebody's livelihood, and frequently a protected performance. That a clean recording is easy to find is not a permission, and a disclaimer in the description is not one either.
Making it sound as though someone said something they did not: No. This is the thing the technology is genuinely dangerous for, and there is no version of it that is fine because the clip was short or the intent was funny. Do not.

Both people, in the situations readers are actually in.

Careful

Permission for a synthetic voice is its own permission. Somebody agreeing to be recorded, agreeing to be interviewed, or selling you a recording has not agreed to be reproduced — and the person who will need that agreement in writing later is you.

What it is honestly for

None of the above makes this a bad tool. It makes it a specific one, and the specific uses are genuinely good — every one of them a case where both people are accounted for.

Use: Why conversion is the right way to do it
Dubbing and re-voicing your own material: You recorded the narration, the read is right, and the voice is not the one the piece needs. Conversion keeps your timing — which means it stays in sync with the picture you already cut — and changes who is heard.
Character work from one performer: Perform every part yourself, then convert each take to a different target. You get distinct voices with your own timing and intent behind all of them, which is why this tends to beat casting a synthetic read for each line.
Consistency across a long project: A series recorded over months in rooms that do not match, or a voice that changed because you had a cold. Converting every episode to one target gives you a single consistent voice across material that was never consistent.
Accessibility and privacy: Someone whose own voice is difficult to listen to, or is tiring to use, or whom they do not wish to be identified by, can speak in their own words and their own rhythm and be heard in another voice. This is the use that most justifies the tool existing.

The four uses that justify the tool existing.

Note

Accessibility and privacy is the one to keep in mind when this technology's reputation comes up. Someone whose own voice is difficult to listen to, or is tiring to use, or whom they do not wish to be identified by, can speak in their own words and their own rhythm and be heard in another voice. This is the use that most justifies the tool existing.

What it is not for

Said plainly, because a tool's own guide rarely does and because the tool itself says nothing at all.

Putting words in someone's mouth. Making it sound as though a real person said something they did not is the thing this technology is actually dangerous for, and there is no version of it that is fine because the clip was short or the intent was funny.
Impersonating a performer. A recognisable voice is usually somebody's income and frequently a protected performance. That a clean recording is easy to find is not a permission.
Getting round somebody's refusal. If a person said no to being cloned, converting a recording of them to a voice you built from other recordings of them is the same act with extra steps.
Passing a voice check. Voice conversion against a bank, a helpline or an identity system is fraud, and it is fraud you have left a receipt for in your job history.
Shopping for a general quality. If what you want is "a warmer voice" or "an older voice" rather than one specific person, the catalogue already holds voices recorded for exactly that and reading typed text needs nobody's permission — see /guides/voice.

In practice

Two habits cover almost all of it. Before you upload, name both people out loud. And if the answer for either of them is "probably fine", treat that as a no — the whole reason this lesson exists is that nothing else in the product will.

That is the honest shape of Voice Changer: your performance, somebody else's voice, one file in and one file out, no effects and no consent check. Get the source right and it does a narrow job well. Get the permission right and it is a tool you can keep using.

Questions

What people ask about this

Free to start

Now go make one

Reading about a style description only gets you so far. The studio is free to use — write one sentence and hear what comes back.

Create a free account

1,000 credits every month. No card required.

All Voice Changer lessons

Two people, not one

Cloning a voice raises one question: may I reproduce this person? Conversion raises two, and the second is the one people miss.

The target voice

You may only clone a voice you have the right to use — your own, or someone else's with their explicit, informed permission. Nothing about the upload can verify whose voice it is, so this is entirely on you.

The speaker in your recording

Conversion involves two people, not one. The target voice is somebody, and so is whoever is speaking in the recording you upload — and their permission is a separate question from the target's. Recording a call, an interview or a meeting is not permission to put those words into another voice, and nothing about the upload asks you whether you have it.

Tip

The cases, plainly

The situation: Where it stands
Your own voice, converted to a voice you cloned from yourself: Go ahead. Both people are you. This is the case the tool is really built for, and it is the only one with nothing to check.
Your own voice, converted to a voice somebody licensed to you: Go ahead. Fine, within whatever that licence says. Read what you were actually granted — permission to use a voice in a product is not always permission to use it in an advertisement.
Your voice, converted to a voice you cloned from someone else: Not without permission. The target's permission is required, and it has to be permission for a synthetic voice specifically. Somebody agreeing to be recorded has not agreed to be reproduced.
Somebody else's recording — an interview, a call, a meeting: Not without permission. The speaker's permission is required as well as the target's, and it is the one people forget. They agreed to be recorded saying that. They did not agree to be heard saying it in a stranger's voice.
A public figure, an actor, a singer, a character from a film: No. Somebody's voice, usually somebody's livelihood, and frequently a protected performance. That a clean recording is easy to find is not a permission, and a disclaimer in the description is not one either.
Making it sound as though someone said something they did not: No. This is the thing the technology is genuinely dangerous for, and there is no version of it that is fine because the clip was short or the intent was funny. Do not.

Both people, in the situations readers are actually in.

Careful

What it is honestly for

None of the above makes this a bad tool. It makes it a specific one, and the specific uses are genuinely good — every one of them a case where both people are accounted for.

Use: Why conversion is the right way to do it
Dubbing and re-voicing your own material: You recorded the narration, the read is right, and the voice is not the one the piece needs. Conversion keeps your timing — which means it stays in sync with the picture you already cut — and changes who is heard.
Character work from one performer: Perform every part yourself, then convert each take to a different target. You get distinct voices with your own timing and intent behind all of them, which is why this tends to beat casting a synthetic read for each line.
Consistency across a long project: A series recorded over months in rooms that do not match, or a voice that changed because you had a cold. Converting every episode to one target gives you a single consistent voice across material that was never consistent.
Accessibility and privacy: Someone whose own voice is difficult to listen to, or is tiring to use, or whom they do not wish to be identified by, can speak in their own words and their own rhythm and be heard in another voice. This is the use that most justifies the tool existing.

The four uses that justify the tool existing.

Note

What it is not for

Said plainly, because a tool's own guide rarely does and because the tool itself says nothing at all.

Putting words in someone's mouth. Making it sound as though a real person said something they did not is the thing this technology is actually dangerous for, and there is no version of it that is fine because the clip was short or the intent was funny.

Impersonating a performer. A recognisable voice is usually somebody's income and frequently a protected performance. That a clean recording is easy to find is not a permission.

Getting round somebody's refusal. If a person said no to being cloned, converting a recording of them to a voice you built from other recordings of them is the same act with extra steps.

Passing a voice check. Voice conversion against a bank, a helpline or an identity system is fraud, and it is fraud you have left a receipt for in your job history.

Shopping for a general quality. If what you want is "a warmer voice" or "an older voice" rather than one specific person, the catalogue already holds voices recorded for exactly that and reading typed text needs nobody's permission — see /guides/voice.

In practice

Whose voice can you use

Two people, not one

The target voice

The speaker in your recording

The cases, plainly

What it is honestly for

What it is not for

In practice

What people ask about this

Does AudioPod check whether I have permission?

Can I convert my own voice to whatever I like?

I recorded an interview. Can I convert the other person's voice?

Can I make myself sound like a famous singer or actor?

What if I only want to anonymise someone?

So what is it actually for?

Now go make one

Whose voice can you use

Two people, not one

The target voice

The speaker in your recording

The cases, plainly

What it is honestly for

What it is not for

In practice

What people ask about this

Does AudioPod check whether I have permission?

Can I convert my own voice to whatever I like?

I recorded an interview. Can I convert the other person's voice?

Can I make myself sound like a famous singer or actor?

What if I only want to anonymise someone?

So what is it actually for?

Now go make one