# Mix and add audio

Set clip volume and fades, clean noise, duck music under speech, add free music, make a voice-over from text or record one, and sync cuts to the beat.

Source: https://infinall.ai/docs/ai-editor/audio
Last checked: 2026-10-26

## What it is

The editor has audio tools for each clip and for each timeline row. You can change volume and fades, remove noise, lower music when someone speaks, add free music and sounds, make a voice from text, record your own voice, and line cuts up with the beat.

## Who can use it

Everyone. The app does not check your plan for any audio tool, so Free and Pro see the same controls. On a team, changing a project needs the Editor role or higher. A Viewer can look but not change.

## Cost

Mixed. Almost everything here is free, with no credits: volume, fades, noise removal, ducking, beat sync, recording, and the free music and sounds. Only the **Voice-over** tool in **AI Gen** uses credits. Its button shows an estimate before you press it, written as "about N credits". **Enhance Text** in the same tool also uses credits, and the app does not show a price for it. See [price before you confirm](/docs/credits/price-before-you-confirm) and [what is a credit](/docs/credits/what-is-a-credit).

## Where to find it

- **Clip audio tools:** click a clip on the timeline that has sound (an audio clip or a video with audio). In the inspector on the left, stay on the **Edit** view and scroll down to the audio sections.
- **Mute and Hear only:** in the row header at the left of each timeline row that holds audio.
- **Free music and sounds:** in the strip named **Editor sections**, choose **Audio**.
- **Text to speech:** in **Editor sections**, choose **AI Gen**, then **Voice-over**.

## Buttons and controls

<!-- shot: ai-editor/audio -->

The numbered rows are the **Music & sounds** panel in the screenshot.

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| 1 | Audio | The button in **Editor sections** that opens the music panel. Its hover text is "Free music and sound effects". | Both |
| 2 | Music & sounds | The panel heading. The hint under it reads "Free to use in your videos. Press play to listen, then Add." | Both |
| 3 | Type of sound | A switch with **Music** and **Sound effects**. | Both |
| 4 | Search music, e.g. happy guitar | The search box. For sound effects it reads "Search effects, e.g. door". | Both |
| 5 | Style | A row of chips. For sound effects the row is named **Kind**. See Options explained. | Both |
| 6 | Songs | The list heading. For sound effects it reads **Effects**. | Both |
| 7 | Listen to (name) | The play button on a card. While it plays it reads **Stop (name)**. | Both |
| 8 | Add (name) to your video | The **Add** button on a card. It changes to **Adding**, then **Added**. | Both |
| | Show more | Under the list. Loads more results. | Both |
| | Try again | Shows only when the sounds could not load. | Both |

**Clip audio, in the inspector (Edit view).** These show for a clip that has sound. They are in this order.

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Voice Over | Closed section. Holds **Record Voice Over**. | Both |
| | Record Voice Over | Starts recording from your microphone at the playhead. The card above it reads **Record at playhead**. While recording the button becomes **Stop & Add to Timeline**. | Both |
| | Clip Audio | Open section with **Volume**, **Fade in** and **Fade out**. | Both |
| | Volume | Slider for this clip. It shows a percentage. | Both |
| | Fade in | Slider. How long the sound takes to come up at the start. Shows seconds. | Both |
| | Fade out | Slider. How long the sound takes to go down at the end. Shows seconds. | Both |
| | Auto Cut Silence | Closed section. It removes quiet gaps. See [AI tools](/docs/ai-editor/ai-tools). | Both |
| | Beat Sync | Closed section. Only for an audio clip. Finds the beat and lines video cuts up with it. | Both |
| | Background Noise Removal | Closed section. The title adds **(Active)** or **(Configured)** once you use it. | Both |
| | Audio Effects | Closed section with **Equalizer**, **Compressor**, **Reverb** and **Delay**. | Both |
| | Audio Ducking | Closed section. Lowers music when speech plays. | Both |

**Timeline row header.** For a row that holds audio.

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Mute (row name) | Turns the row's sound off. Its hover text is "Mute track". It changes to **Unmute (row name)**. | Both |
| | Hear only (row name) | Plays only this row. Its hover text is "Hear only this row". It changes to **Hear all rows again**. | Both |

**Voice-over (AI Gen).**

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Voice engine | A menu of four voice engines. | Both |
| | Voice | A menu of voices for the engine you picked. | Both |
| | Text | Box for what the voice should say. It shows a count such as 0/5,000. | Both |
| | Enhance for speech | A switch. When on, an **Enhance Text** button shows. | Both |
| | Enhance Text | Asks the AI to rewrite your text so it reads well aloud. A preview box appears, with **Discard**. | Both |
| | Generate Speech | Makes the voice. The label ends with the estimate, "about N credits". | Both |
| | Play preview | In the result card. It reads **Pause preview** while playing. | Both |
| | Save to Media | Puts the voice file in your **Media** panel. | Both |
| | Add to Timeline | Puts the voice on the timeline. | Both |
| | Download | Saves the voice file to your computer. | Both |

## How to change volume and fades

1. Click a clip on the timeline that has sound.
2. In the inspector, stay on **Edit** and find **Clip Audio**.
3. Drag **Volume**. The number is a percentage.
4. Drag **Fade in** and **Fade out** to set how long the sound eases in and out.

To silence a whole row, use **Mute (row name)** in its header. To hear one row alone, use **Hear only (row name)**.

## How to clean up noise

1. Select the clip and open **Background Noise Removal**.
2. Click a card such as **Speech Focus**. The card is named **Use Speech Focus noise reduction**.
3. Check the line **Current mode** and the **Applied** tag.
4. To compare, use **Hear Original** and **Hear Cleaned** under **A/B Preview**.
5. To let the editor choose, click **Analyze & Recommend**. When **Recommendation ready** shows, click **Apply Recommended Cleanup**.

For a one click fix, use **Clean audio** in the quick bar at the top of the **Edit** view, or **Quick Dialogue Cleanup** in the **AI** tools. Both turn on the **Speech Focus** cleanup. You see "Noise cleanup applied".

## How to lower music under speech

1. Put the music and the voice on separate rows.
2. Select the music clip and open **Audio Ducking**.
3. Turn on **Enable audio ducking**. The state reads **Ducking Enabled**.
4. Under **Trigger Source (Voice Track)**, click the voice row. Each row is named **Use (row name) as ducking trigger**.
5. Optional: click a preset under **Ducking Presets**.
6. Click **Apply Ducking**. It reads **Analyzing...** and then **Ducking Applied**.
7. To change it, adjust the sliders and click **Update**. To take it off, click **Remove**.

## How to add free music or a sound effect

1. In **Editor sections**, choose **Audio**.
2. Choose **Music** or **Sound effects**.
3. Click a chip, or type in the search box.
4. Click **Listen to (name)** to hear it.
5. Click **Add**. You see "Added to your video". The sound goes on a new row.

If the sound needs a credit, the message also shows "Credit:" and the artist. Add that name to your video description.

## How to make a voice from text

1. In **Editor sections**, choose **AI Gen**, then **Voice-over**.
2. Pick a **Voice engine** and a **Voice**.
3. Type your words in **Text**.
4. Optional: turn on **Enhance for speech**, click **Enhance Text**, read the preview, and edit it or click **Discard**.
5. Click **Generate Speech**. Check the estimate on the button first.
6. Press **Play preview**.
7. Click **Save to Media**, **Add to Timeline** or **Download**. A voice you do not keep is lost when you leave the panel. The app warns you with "Unsaved audio" and the next line tells you the three ways to keep it.

## How to record a voice-over

1. Move the playhead to where the voice should start.
2. Select a clip that has sound and open **Voice Over** in the **Edit** view.
3. Click **Record Voice Over**. Allow the microphone if your browser asks.
4. Speak. The timer counts up.
5. Click **Stop & Add to Timeline**. The recording is added on a new row named **Voice Over**. It stays in the project and exports with the rest of the sound.

## How to cut video to the beat

1. Put your video clips on the timeline and your music on its own row.
2. Select the music clip and open **Beat Sync**.
3. Click **Detect Beats**. You see **BPM Detected** and the number of **Beats**.
4. Under **Select tracks to sync to beats:**, click the video rows to include.
5. In **Cut every**, choose **1 beat**, **2 beats**, **4 beats** or **8 beats**.
6. Check the **Preview**. Click **Apply Beat Cuts to N Clips**.
7. To change the settings first, click **Sync Settings**.

## Options explained

**Music chips (Style).** For you, Upbeat, Calm, Cinematic, Lo-fi, Business, Acoustic, Electronic and Piano. **For you** picks a fresh starter word each time. Typing in the search box overrides the chips.

**Sound effect chips (Kind).** For you, Whoosh, Click, Pop, Hit, Applause, Ding, Nature and Typing.

**Where the sounds come from.** The panel says the sounds are openly licensed, either CC0 or CC BY, so you can use them in any video. For CC BY sounds you should credit the artist.

**Noise cards.** Balanced is general cleanup for moderate room noise. Speech Focus keeps speech clear while it reduces hiss and the room. White Noise is strong hiss removal for fans, air and camera noise. Music Bed pushes background music down. Heavy Noise is stronger cleanup for loud air, fans and street sound. Wind & Rumble targets low rumble, handling noise and wind. Hum & HVAC targets steady hum and air conditioning drone.

**Noise sliders.** **Threshold** goes down to -80 dB. **Reduction** is a percentage. **Attack** is in ms. **Release** goes up to 500 ms. **A/B Preview** is for listening only. The note says export still uses the cleaned sound.

**Audio effects.** Each effect has a switch named like **Equalizer enabled**, and a button named like **Expand Equalizer**.

| Effect | Controls |
| --- | --- |
| Equalizer | Five bands: 60Hz, 250Hz, 1kHz, 4kHz and 16kHz. Each goes from minus 12 to plus 12. |
| Compressor | Threshold, Ratio, Attack and Release. |
| Reverb | Room Size, Damping and Wet/Dry. |
| Delay | Time (up to 2,000 ms), Feedback (up to 95 percent) and Wet Level. |

**Ducking presets.** Subtle is a gentle lowering. Moderate is the middle. Aggressive lowers the music a lot. Podcast is tuned for a talking voice over music.

**Ducking sliders.** **Detection Threshold** is the voice level that triggers ducking, from -50 to -10 dB. **Volume Reduction** is how much the music is lowered, as a percentage. Click **Timing Controls** for three more: **Attack** (how fast the volume drops when the voice starts, up to 0.5 seconds), **Release** (how fast it returns, up to 1 second) and **Hold Time** (the shortest time to stay lowered between words, up to 0.5 seconds).

**Beat Sync modes.** Open **Sync Settings** to pick one.

| Mode | What it does |
| --- | --- |
| Cut Video to Beats | Arranges video clips with a cut every 1, 2, 4 or 8 beats. |
| Smart | Adjusts each clip's length to fit the nearest beat count. |
| One per Beat | Each clip gets exactly one beat. |
| Preserve Duration | Keeps each clip's length and snaps its start to a beat. |

For the three modes other than **Cut Video to Beats**, you can also set **Beat Subdivision** (1/1, 1/2 or 1/4), **All Beats** or **Downbeats**, and **Offset**, which moves the beats by up to 500 ms either way.

**Voice engines.** There are four: one marked "(expressive)" that is the default, one marked "(fast)", one called HD, and one HD marked "(multilingual)". The list of voices changes with the engine. The expressive and fast engines offer Ethan, Olivia, Harper, Grant, Iris and Jasper (English, US), and Kavya and Dhruv (Hindi). The HD engine offers Ava, Andrew, Emma and Brian. The multilingual engine offers Ava Omni and Andrew Omni.

## Limits

- Clip **Volume** goes up to 400 percent.
- **Fade in** and **Fade out** can each be up to half the clip length.
- Text to speech takes up to 5,000 characters.
- A single media file can be up to 500 MB when you bring it in. See [import media](/docs/ai-editor/import-media).
- A request to the AI that has already used 100 credits stops. This applies to chat requests, see [AI tools](/docs/ai-editor/ai-tools).

## When something goes wrong

| Message in the app | What it means | What to do |
| --- | --- | --- |
| Sounds could not load | The free music list could not be fetched. The line under it ends "Check your internet connection." | Check your connection and click **Try again**. |
| Nothing found | The search has no results. | Try a different word, like happy, rain or applause. |
| Could not play this sound | The preview file would not play. | Try another sound. |
| Could not add this sound | The file could not be downloaded or imported. | Try again, or pick another sound. |
| Allow microphone access to record a voice over. | Your browser blocked the microphone. | Allow the microphone for this site and click **Record Voice Over** again. |
| Voice-over recording is not supported by this browser. | The browser cannot record. | Use an up to date desktop browser. |
| This browser does not provide a supported microphone recording format. | The browser has no usable recording format. | Use another browser. |
| The microphone recording stopped unexpectedly. | The microphone was lost during recording. | Record again. |
| Could not start microphone recording. | The recording could not begin. | Check that no other app is using the microphone. |
| Insufficient credits | You do not have enough credits to make the voice. | Add credits in Settings > Credits. Editing, saving and export stay free. |
| Please enter some text | **Enhance Text** or **Generate Speech** was used with no text. | Type your words first. |
| Reset music speed and freeze frames before detecting beats. | Beat detection does not work on a clip with changed speed or freeze frames. | Reset the speed on the music clip, then click **Detect Beats**. |
| No unlocked tracks with eligible clips found. Add video clips to the timeline first. | **Beat Sync** found no unlocked video row to sync. | Add video clips, or unlock the row. |
| No speech crossed the trigger threshold. Lower the threshold or choose a louder source track. | The voice row is too quiet for the **Detection Threshold**. | Move the threshold toward the left end of its slider, or pick a louder row. |
| Add another audio or video track with speech to use as trigger | There is no other row to use as the voice. | Put the voice on its own row. |
| Failed to apply ducking. | Ducking could not be saved. | Try again. |

## Questions

### Is there an audio mixer with faders and a pan control?

No. The editor has no mixer panel and no pan control. Use **Volume**, **Fade in** and **Fade out** on each clip, and **Mute** or **Hear only** on each row.

### Do I see beat lines on the timeline after Detect Beats?

No. **Detect Beats** shows the BPM and the number of beats in the **Beat Sync** section. The timeline does not draw beat lines today.

### Can I use the free music in an ad?

The panel says every sound is CC0 or CC BY, and both allow commercial use. For CC BY sounds you must name the artist, and the app shows the credit line when you add the sound.

### Does recording a voice-over cost credits?

No. Recording uses your microphone and your browser. Only **Generate Speech** and **Enhance Text** use credits.

### What happens if I leave the Voice-over panel with a voice I did not save?

The voice is lost. The panel warns you with "Unsaved audio", and if you click **All tools** the editor says "Your voice-over was not kept". Use **Save to Media**, **Add to Timeline** or **Download** first.

### Does text to speech work in other languages?

The multilingual engine is for that, and the first two engines include Hindi voices. Pick the engine first, because the **Voice** list depends on it.