How to Use Text to Speech in PowerDirector
Turn subtitles you have already written into a spoken voice-over, without a microphone.
Updated Aug. 29, 2026 · 4 min read
In a hurry? Use the text to speech quick guide instead.
No microphone, or a voice-over needed in a language you do not speak: both are jobs for Text to Speech. The thing to know before you open it is that it has no box to type into. It reads text that is already on the timeline, so the writing happens first.
- Put the words on the timeline as subtitles or text objects.
- Select them, then click Text-to-Speech above the timeline.
- Set the language to match the text before anything else.
- Pick a voice and a mood, then click Generate.
Steps to Generate a Voice-over
Prefer to watch? Here is the whole thing in a little over two minutes.
Step 1: Get the Words on the Timeline
There are three routes, and any of them works. From the Subtitles menu, Create Subtitles Manually lets you type the lines in.

If the script already exists as a file, Import Subtitles from File brings it in instead of retyping it.

Text objects count too, so a title you have already placed can be spoken without being retyped as a subtitle. The tutorial calls this the Text feature; the tab itself is labeled Titles.

Step 2: Select the Text and Open the Tool
With every line written, select the subtitles or text objects you want spoken. Only what you select is converted.

Then click Text-to-Speech above the timeline.

Step 3: Set the Language First
The language menu comes first for a reason: a language that does not match the text causes generation errors rather than an accent. Set it to whatever the words actually are.

Step 4: Pick a Voice, and a Mood
The window opens on My Voice Profile, and on a fresh install it is an empty dashed box. Clone Voice is the way in: it hands you to AI Voice Cloning, which builds a profile from a recording of your own voice.
Once a profile exists it sits here above the stock voices, so a whole series can be narrated in one consistent voice without recording every line.

Underneath sits Voice Profiles, the stock voices. The filters above the grid narrow it by gender, speaking style and use case, and Clear All resets them.
The grid is not simply a list of speakers: each card is a name paired with a style, so the same voice appears more than once reading in different moods. The heart icon at the end of the filter row keeps the ones you come back to.


Click a card to hear it. Reading the same line in a different mood changes the result more than changing speaker does, so it is worth listening to a few before settling.

Parameter Settings underneath adjusts speed and pitch, which is the fix when a voice is right but reads a little fast for the pictures.

This one draws on credits, and the button shows the cost. Beside a Total characters count, the Generate button carries a credit badge, so the price of the run is visible before you commit to it. A Get Credits link sits at the top of the same window.
The Result
Click Generate and PowerDirector produces the audio and returns you to the editor. The speech lands on the timeline at the positions the text occupied, so it is already in sync with the subtitles it was made from.


Haven't got PowerDirector yet? No microphone needed, and no box to type into either: it reads subtitles you have already written.
Download PowerDirector Free →FAQs
Because each card is a speaker paired with a speaking style, so one name can appear several times reading in different moods. Use the filters above the grid to narrow it by gender, style or use case, and Clear All to start again.
Yes. The Generate button carries a credit badge showing what the current run will cost, next to a running Total characters count, and there is a Get Credits link in the same window. Longer scripts cost more, so the badge is worth a glance before you generate.
Yes, through a voice profile. Clone Voice, at the top of the Text to Speech window, leads to AI Voice Cloning, which builds a profile from a recording of you. After that the profile appears in the same window as the stock voices and can read any script you put on the timeline.

