Anomaly Detection in Rotating Machinery Using Self-Supervised Learning
A university research presentation turned into a narrated video for academic sharing.
Voice: ElevenLabs Eleven v3 / George
How to narrate a PowerPoint: record audio slide by slide, redo a single slide, fix a silent mic, export to MP4, see Mac and web differences, or use an AI voice.
Official sources checked:
PowerPoint’s Record feature lets you record narration slide by slide while you move through your slide show. The audio is saved separately for each slide, so you can re-record only the slide where you misspoke. When you’re done, use Create a Video to export an MP4 with your narration.
This guide explains how to record audio on PowerPoint step by step. It focuses on the recording screen in Microsoft 365 for Windows, then covers the differences on Mac and the web, what to do when recording doesn’t work, and how to export a video. If you don’t have the time or a quiet place to record your own voice, it also shows how to narrate a PowerPoint with an AI voice that reads your script.
Menu and button names follow Microsoft Support’s English pages (checked on November 10, 2026). What you see may differ depending on your version of PowerPoint and your organization’s settings.
There are three main ways to add your voice to PowerPoint slides. Choose based on whether you want narration throughout the presentation or only on some slides, and whether you will be the one speaking.
PowerPoint also has a Speak feature that reads selected text aloud. However, Microsoft’s instructions don’t describe any way to save that audio to a slide. The section on Speak below explains this in more detail.
Taking care of these five things before you start will cut down on retakes.
Example script for the notes (a fictional example created by our editorial team, translated from Japanese: slide 3 of a “Statistics Basics” class)
On this slide, we’ll look at the difference between the mean and the median. The graph on the left shows the commute times of 20 students in one class. Most of them take 30 minutes or less, but one student takes two hours, so the mean comes out high, at 32 minutes. The median, on the other hand, is 25 minutes. When your data has extreme values, check the median as well.
(Wait until the next slide appears before you continue.)
The Japanese original of this example is about 140 characters, which takes just under 30 seconds to read at 300 characters per minute (our calculation). The English translation above is about 70 words, or a little over 30 seconds at 130 words per minute (our calculation). Speaking speed varies from person to person, so read your script aloud once, time it, and then adjust the length.
Rather than reading the slide text word for word, say where to look on the screen, as in “the graph on the left.” That makes your talk easier to follow by ear alone. For script lengths by running time and tips on turning written text into spoken language, see our guide to writing narration scripts.
PowerPoint for Microsoft 365 on Windows has been rolling out a new recording experience since early 2022. On work or school computers, you may still see the older screen, depending on when your administrator rolls out new features. If so, check the steps for the classic screen in the table in the next section.
The overall flow is: prepare your script, record slide by slide and check the result, save, and then export to video.
Our editorial team summarized the order of work based on Microsoft Support’s instructions.
Open the recording screen from the Record tab
Open the Record tab on the ribbon, select Record, and choose whether to start from the beginning or from the current slide.
Set up your camera and microphone
Use the icons at the top of the screen to turn your camera and microphone on or off. To record only your voice, turn the camera off. To use a different microphone, choose it from Select more options (···).
Choose how your script is shown
In the Views menu, choose Teleprompter to show your notes at the top of the screen. Use auto-scroll to adjust how fast the script scrolls. You can also switch to Presenter View or Slide View.
Select Start recording and speak
Recording starts after a countdown. When you finish talking about a slide, use the arrow on the screen to move to the next one. If you use the pen, highlighter or laser pointer, those movements are recorded too.
Start speaking after the slide changes
According to Microsoft, narration isn’t recorded during slide transitions. Wait until the next slide appears before you start talking, and finish speaking before you advance.
Stop and play it back
Select Pause to take a break, or Stop when you reach the end. Use Play to check your volume and listen for slips.
Save the file
When you finish recording, save the file. Your recorded audio is stored inside the file, slide by slide.
How you start recording and export to video depends on your version of PowerPoint. Microsoft ended support for PowerPoint 2016 and 2019 on October 14, 2025, so this guide focuses on the Microsoft 365 screens. How to re-record a single slide and how to clear recordings are summarized in the table in the next section.
| Version | How to start recording | Exporting to video |
|---|---|---|
| Windows: Microsoft 365 (new recording screen) | On the Record tab, select Record and choose to start from the beginning or from the current slide. Select Start recording; recording begins after a countdown | Export > Export Video in the recording screen, or File > Export > Create a Video |
| Windows: 2016–2019 (classic screen) | Record Slide Show on the Slide Show tab. The top half of the button starts from the current slide; the bottom half lets you choose the beginning or the current slide. There is a 3-second countdown | File > Export > Create a Video |
| Mac (Recording Studio) | Start with the Record button in the top-right corner. You can start from the beginning or from the current slide. Turn the camera off to record only your voice | Record > Export to Video, or File > Export and choose MP4 or MOV. Narration is included in the video from version 16.19 onward |
| Web (PowerPoint for the web) | Recording a slide show isn’t supported. You can insert audio files recorded elsewhere (MP3, WAV, M4A, AAC or OGA, up to 16 MB) into slides | To record narration, open the file in the desktop version of PowerPoint, record there, and export from the desktop app |
According to Microsoft, animation effects don’t play in videos exported on a Mac. Narration recorded in the slide show plays, but other embedded audio and video doesn’t. If you reveal content step by step with animations, split it across separate slides so the video shows it in the same order.
Recorded narration is saved slide by slide, so you can re-record only the slide where you misspoke. There’s no need to start over from the beginning. The table below shows how to do this in each version, followed by the steps for re-recording in the new recording screen.
| Version | Re-record one slide | Delete recordings |
|---|---|---|
| Windows: Microsoft 365 (new recording screen) | Go to the slide you want to redo, use Delete to remove its recording, then record again with Start recording | Delete removes the recording on the current slide or on all slides. Clear deletes the recorded narration |
| Windows: 2016–2019 (classic screen) | Go to the slide you want to redo and record from the current slide using the top half of Record Slide Show. That slide’s previous narration (audio, ink and laser pointer) is overwritten | Choose Clear from the down arrow on Record Slide Show, then clear the timings or the narration on the current slide or on all slides |
| Mac (Recording Studio) | Delete the current recording with the trash can button to the left of the Record button, then record again | Under Clear Recording, choose Clear Recording on Current Slide or Clear Recording on All Slides |
Go to the slide you want to re-record
In the recording screen, display the slide you want to record again.
Delete the current recording
Use Delete to remove the recording on the current slide. You can also delete the recordings on all slides and start over.
Record it again
Select Start recording to record that slide. When you finish, select Stop, then use Play to check how it flows with the slides before and after it.
If you want to keep the recorded narration but speak live at the venue, clear the Play Narrations and Use Timings check boxes on the Slide Show tab. The narration won’t play, and you can advance the slides by clicking.
Before re-recording, save a copy of the file under a different name so you can go back to the earlier recording if you need to.
If only some slides need narration, or recording the whole slide show in one go is difficult, you can add audio to slides one at a time. In PowerPoint for Windows, you can record directly onto a slide with these steps.
Open Record Audio
Go to the slide where you want the audio, then on the Insert tab, select Audio > Record Audio.
Name it and record
Type a name for the audio, select Record, and speak. When you’re done, select Stop.
Listen, then insert it
Select Play to listen. Select Record to redo it, or OK if you’re happy with it. A speaker icon is added to the slide; drag it to a spot where it won’t cover your content.
Choose how it plays
Select the icon, and on the Playback tab, set Start to Automatically so the audio plays when the slide appears. To hide the icon during the slide show, turn on Hide During Show (use this only when the audio plays automatically).
On Microsoft’s support page, these Record Audio steps are given only for Windows. On a Mac or on the web, insert an audio file recorded in another app via Insert > Audio. Audio files made with an AI voice tool can be inserted the same way.
Prepare your audio files as .m4a (AAC), which Microsoft recommends, or as .mp3. From PowerPoint version 2505 onward, formats such as WMA and AIFF have limited support and are deprecated. Even if such a file can be inserted, it is converted to MPEG-4 when inserted and when played.
With audio added this way, pay attention to how long each slide is shown when you export the video. For files without recorded narration, the export default is 5 seconds per slide. If the audio length differs from slide to slide, enter a display time in seconds that matches each slide’s audio under Advance Slide > After on the Transitions tab, then export using those timings.
Microsoft’s export page doesn’t say how manually entered timings are handled. Play the exported video and check that no slide changes in the middle of its audio.
Here are the most common sticking points when recording or exporting, and where to check. For the Mac export limitations and the web version, see “Differences in PowerPoint 2019 and Earlier, on Mac, and on the Web” above.
| Problem | Likely cause | What to check / how to fix it |
|---|---|---|
| The microphone doesn’t pick up your voice on Windows | Apps aren’t allowed to use the microphone, or a different microphone is selected | In Settings > Privacy & security > Microphone, turn on Microphone access and Let desktop apps access your microphone. In Settings > System > Sound > Input, choose the microphone to use. Also make sure the microphone icon in the recording screen is on |
| The microphone doesn’t pick up your voice on a Mac | PowerPoint isn’t allowed to use the microphone | Choose Apple menu > System Settings > Privacy & Security > Microphone, and turn on access for PowerPoint |
| The start or end of what you say is cut off | Narration isn’t recorded during slide transitions | Start speaking after the next slide appears, and finish speaking before you advance |
| The exported video has no audio | Don’t Use Recorded Timings and Narrations was selected when exporting, or your version of PowerPoint for Mac is old | Export again with Use Recorded Timings and Narrations selected. On a Mac, recorded narration is included in the video from version 16.19 onward |
| Slides in the video advance every 5 seconds | Without recorded timings, the export default is 5 seconds per slide | Record the timings along with your narration using Record, or change Seconds to spend on each slide in the export settings |
| The export takes a very long time | Depending on the video’s length and content, creating it can take several hours (according to Microsoft) | Leave time to export well before your deadline. You can keep using PowerPoint while the video is being created |
If your voice still isn’t picked up, go to Input under Sound in Windows Settings, open your microphone’s properties, select Start test, and speak. If it responds there, the microphone is working, so review the settings on the PowerPoint side (the microphone selected in the recording screen). If it doesn’t respond, check the microphone connection and your operating system settings.
Recorded narration plays inside the PowerPoint slide show. To share it with people who don’t have PowerPoint, or to submit or publish it, export it as an MP4 video.
To include your recorded narration on Windows, go to File > Export > Create a Video and select Use Recorded Timings and Narrations. If you’ve recorded narration, this option is selected by default.
On a Mac, choose File > Export (Save As has no video options). Choose MP4 or MOV under File Format and a setting under Quality, turn on Use Recorded Timings and Narrations, and then select Export.
After exporting, play the video from start to finish to check that every slide has audio and nothing is cut off. If there’s no audio, or the slides advance every 5 seconds, see the table in “When Audio Won’t Record or Gets Cut Off” for causes and fixes.
For the rest of the export process, such as choosing the video quality and adding captions, see our guide to converting PowerPoint to video.
PowerPoint’s Speak feature reads selected words or passages aloud on the spot. Microsoft’s instructions don’t describe a way to save the spoken audio to a file or to add it to slides as narration.
Japanese-language search terms, translated into English. Darker bars are terms that include 読み上げ (“read aloud”).
Source: Google Ads Keyword Planner (Japan, Japanese; average monthly searches from October 2025 to September 2026; retrieved November 2026)
Google’s Keyword Planner shows that in Japan, people search for having PowerPoint read text aloud about as often as they search for recording audio in PowerPoint. This suggests that some people are looking for a way to have PowerPoint read their text instead of recording it themselves.
To use Speak, open Customize Quick Access Toolbar > More Commands, set Choose commands from to All Commands, and add Speak. The voices available depend on the speech synthesis (text-to-speech) engines installed on your computer and on the language.
If you want a synthetic voice as your slide narration, create audio files with an AI voice tool and insert them, or use a tool that turns your slides into a narrated video. The next section compares the options.
If you want students or an audience to hear your own voice, use PowerPoint’s Record feature. If you have nowhere to record, don’t want to do take after take, or find it a burden to re-record every time you edit the script, an AI voice is also an option. The table compares the methods.
| Method | Voice | Script | When you change the content | What you need / conditions |
|---|---|---|---|---|
| PowerPoint Record | Your own voice | You prepare it. Text in the notes appears in the teleprompter | Re-record that slide | Desktop PowerPoint, a microphone and a quiet place |
| Record Audio or inserting audio files | Your own voice or audio made separately | You prepare it | Replace that slide’s audio | Recording is Windows only; inserting also works on Mac and the web. Match each slide’s display time to its audio |
| Creating audio in an AI voice tool and inserting it (e.g., Clipchamp, VOICEVOX) | AI voice | You prepare it | Regenerate and reinsert the audio for the slides you changed | Create and insert one audio file per slide. Check each tool’s terms (such as credit requirements) |
| Video creation in Microsoft 365 Copilot | AI narration | Requires PowerPoint speaker notes | Open and edit it in Clipchamp | A work account with a Copilot license (per Clipchamp’s description as of August 2025), with Clipchamp enabled in the organization |
| SpeechSlide AI | AI voice (choose from 4 voice engines) | AI writes it from your slides, and you can edit it | Edit that slide’s script, regenerate its audio, then recreate the video | PDF or PowerPoint (.pptx or .ppt, up to 20 MB). You can’t record your own voice |
With an AI voice tool, you write the script, create one audio file per slide, insert them one by one, and match the display times. Clipchamp lets you use AI voices even on its free plan and export the audio on its own (video exports go up to 1080p without a watermark). VOICEVOX, a free Japanese text-to-speech app, can be used for commercial and non-commercial purposes at no cost, but its terms of use require a credit.
How an AI voice sounds depends on the engine and the voice. Listening to real samples before you start makes it easier to choose a voice that suits your material.
SpeechSlide AI turns a PDF or PowerPoint file into an MP4 video: after you upload the file, AI writes a spoken-style script for each slide and adds AI voice narration. Writing the script, generating the audio and exporting the video all happen one after another within a single project.
It has no feature for recording your own voice. Use it instead of recording when you don’t have the time or a place to record, or when you don’t want to re-record every time you change the content.
Create a new project and upload your file
Upload a PDF or PowerPoint file (.pptx or .ppt). Files uploaded through the app can be up to 20 MB each. For Google Slides, export to PDF or PPTX first, then upload.
Generate and edit the script
Generate scripts for all slides at once, then edit them slide by slide. You can set the script language and instructions on how it should sound (for example, “keep it concise and explain technical terms”). PowerPoint speaker notes aren’t imported as the script, so if your notes already contain a script, edit the AI-generated script to match them.
Choose a voice and generate audio
Choose a voice from four engines: ElevenLabs, Google, Google (Gemini) and OpenAI. You can adjust settings such as speaking speed, and use a different voice for each slide. According to the How to Use page, generating audio takes about 10–30 seconds per slide.
Set pauses and display times
Slides without narration can be set to “no audio” with a display time in seconds. You can also set the pause after each slide (1 second by default).
Create and download the video
Videos are exported as Full HD (1920×1080) MP4 files. According to the How to Use page, exporting takes about 1–5 minutes. If a word is misread on a slide, fix that slide’s script, regenerate only its audio, and create the video again.

The free plan lets you create 2 projects a month with up to 8 slides per project and up to 4 videos a month, and you can download the videos with a watermark. No credit card is required. The Standard plan ($15/month, or ¥2,200/month including tax in Japan) removes the watermark and allows up to 20 slides per project.
Captions, background music and avatars (on-screen presenters) aren’t included. If you need them, finish the exported MP4 in video editing software.
You can also use the AI-generated script as a reading script when you record yourself. Copy the text from the script field and paste it into your PowerPoint notes, and it will appear in the recording screen’s teleprompter.
Decide based on whose voice you want people to hear and how likely you are to change the content later.
Even if you record in your own voice, polishing the script first reduces retakes. If you’re unsure about using an AI voice, listen to voice samples first to see whether there’s a voice that suits your material.
Just upload your slides to get a narrated presentation video like this one.
A university research presentation turned into a narrated video for academic sharing.
Voice: ElevenLabs Eleven v3 / George
Explore slide-to-video workflows across education, healthcare, training, sales, and creator use cases.
View Use CasesYes. The free plan lets you create up to 2 projects and 4 videos per month. No credit card required.
PDF and PowerPoint (PPT/PPTX) files are supported. Upload the slides you already have.
AI narration is available in Japanese, English, Chinese, Korean, German, Spanish, and more.
Yes. Videos can be used for training, sales, lectures, and marketing. Paid plans allow watermark-free MP4 downloads.
A three-step guide to creating AI narration: script techniques that make it sound natural, cautions on commercial use and credit requirements, and four criteria for choosing a tool — with real audio samples embedded.