Music and audio
Valmera mixes your footage’s sound, music, voiceover and sound effects as separate layers. Upload a track or name a song and it’s laid under the video with smooth ducking under speech; add narration, find real recorded sound effects, lower the music already in your footage, and master the final mix to −14 LUFS. Valmera doesn’t generate music.
Add background music
- Bring a track
Upload an MP3, WAV, M4A, AAC or OGG file (up to 50 MB), paste a link, use the audio from a video clip you have, or name a specific song.
- Ask for it under the video
Type “Put my track under the whole video.” Start and end times are optional; the default is the whole video.
- Let the defaults do the mixing
Under speech the music sits low as a bed and dips smoothly while people talk; where nobody speaks it plays as the lead. Short tracks loop, and the music fades in and out.
- Adjust by ear
Ask to start later, skip the song’s intro, lower or raise it, stop the ducking, or swap the track. Each change keeps the other settings.
- Put my track under the whole video.
- Start the music at 0:30 and skip the first 20 seconds of the song.
- Lower the music a little and let it run to the end.
- Try a different song: use upbeat.mp3 instead, same position and level.
Type these in Studio chat, or ask Claude, ChatGPT, Cursor or another assistant connected to Valmera’s MCP server. More in the prompt library.
Use a specific song
Name it, for example “add Blinding Lights by The Weeknd”, and Valmera searches the web for the official recording, downloads one version and tells you which. Finding a song doesn’t give you the right to publish it, so use music you’re licensed to use.
Have the song only as a video, such as a TikTok or Reel download? Upload the clip and say “use the audio from this clip as the music.” Valmera extracts the sound; the picture isn’t used.
Ducking under speech
By default, music ducks smoothly: a sidechain dip that follows the voice in your footage and swells back in the gaps. Prefer a classic fixed dip? Ask for “step ducking” (a hard 12 dB drop under speech). Don’t want any? Ask “don’t duck the music.”
Ducking listens to your footage’s own audio. Over a separate voiceover, set the music level directly: “lower the music under the narration.”
Balance the layers
| Layer | What you can change | Example request |
|---|---|---|
| Your footage’s audio | Volume over any time range, or mute | “Lower the original audio by 6 dB from 1:00 to 1:20” |
| Music | Level, start, end, offset into the song, loop, fades, ducking | “Drop the music 4 dB and fade it out over 3 seconds” |
| Voiceover | Level, start, offset, duration | “Make the narration a little louder” |
| Sound effects | Level and timing of each effect | “The whoosh is too loud; move it 0.2 s earlier” |
| Inserted clips | Mute each clip’s own sound | “Mute the b-roll clip at 0:30” |
Add a voiceover
Upload your narration as an audio file and say where it starts. While it plays, your footage’s own audio drops by 12 dB (ask to turn that off). You can start partway into the file and play only part of it.
You can also reuse a line from your own footage over another scene, the classic J-cut or L-cut: “play the sentence ‘we doubled revenue’ over the b-roll at 0:40.” Valmera doesn’t generate AI voices; record the narration yourself.
- Lay narration.mp3 over the video starting at 0:05.
- Start the voiceover 12 seconds into the file and play 20 seconds of it.
Sound effects
Describe the sound by what it is: “a whoosh on the cut at 0:12”, “a camera shutter when the photo appears.” Valmera searches real recordings on Openverse and Freesound, compares candidates by their actual waveform, places the best one on the exact moment and reports its source and license. It may decide a moment works better without a sound and tell you. You can also upload your own effects.
- Effects are placed at −6 dB by default and run up to 15 seconds unless you ask for longer.
- Some recordings require credit or are for non-commercial use only; Valmera passes on those terms.
- Move or remove any effect later: “move the whoosh to the next cut”, “remove the pop at 0:40.”
- Add a soft whoosh on each scene change in the intro.
- Add a camera shutter sound when the photo appears at 1:12.
Remove or lower music in your footage
If your recording already has music under the speech, ask “remove the background music but keep my voice” or “lower the song under his voice.” Valmera separates the soundtrack into voice and everything else, then sets each level (anywhere from −60 dB, effectively muted, to +6 dB). Dense mixes can leave traces, so listen to the preview. Ask “restore the original soundtrack” to undo it.
Loudness mastering
Ask “master the audio for social” and the final mix is normalized to −14 LUFS with a −2 dBTP ceiling, on the preview and the export. It fixes a quiet export; it doesn’t change the balance between voice and music, so set that first.
Music-driven edits
Valmera measures a track’s tempo, beats, loudest sections and biggest energy rise. Use that to “start the music on the drop” or to cut to the beat.
Troubleshooting
| Symptom | Fix |
|---|---|
| Music is too loud under the narration | Ask “lower the music under the narration.” Ducking follows your footage’s audio, not a separate voiceover. |
| Music drowns out speech | Lower the music, or switch to step ducking for a deeper dip under speech. |
| The ducking is too noticeable | Lower the music overall and turn ducking off, or try step ducking. |
| The song starts with a slow intro | Ask to start 20 seconds (or wherever the energy starts) into the song. |
| Music stops before the video ends | Ask to loop it or let it run to the end. |
| The wrong version of a named song was used | Give the artist and exact title, or paste a link to the version you want. |
| Background hum or noise | Valmera has no audio denoise. Clean the audio in another tool, or record again. |
| The export sounds quiet | Ask for loudness mastering to −14 LUFS. |
Limits
- No AI music generation, and no stock music search by mood or genre.
- No audio denoise or “studio sound”.
- No AI voice generation; voiceovers come from audio you upload or from your footage.
- No automatic per-speaker leveling inside one track; adjust sections by time instead.
- Clips used as overlays are silent.
Tools behind this page: add_music swap_music set_music_fit remove_music find_song fetch_url extract_audio set_audio_gain set_volume add_voiceover remove_voiceover search_sfx audition_sfx_candidates fetch_sfx add_sfx add_web_sfx move_sfx remove_sfx separate_music remove_stem_mix set_master_loudness get_audio_analysis review_audio audit_audio_mix · Full toolset reference
Frequently asked questions
Where does the music in Valmera come from?
From you: upload a track, paste a link to a song or video, use the audio from a clip you uploaded, or name a specific song and Valmera finds the recording on the web. It doesn't search a music library by mood, and it doesn't generate music. Use music you have the rights to publish.
Can Valmera generate music with AI?
No. Valmera doesn't generate music or songs. It mixes tracks you provide or a named song it finds on the web.
Can Valmera lower the music when someone talks?
Yes. Music ducks automatically under the speech in your footage with a smooth sidechain dip that swells back in the gaps. You can switch to a fixed 12 dB dip or turn ducking off.
Can Valmera remove background noise?
No. There is no audio denoise or “studio sound”. Record as cleanly as you can, or clean the audio in another tool before uploading.
Can Valmera remove the music from my recording?
Yes. Ask “remove the background music but keep my voice.” Valmera separates the soundtrack into voice and everything else and sets each level. Dense mixes can leave traces, so listen to the preview.
What loudness does Valmera master to?
Ask for mastering and the final mix is normalized to −14 LUFS with a −2 dBTP ceiling, on both the preview and the export. It changes overall loudness, not the balance between voice and music.
Can I add sound effects?
Yes. Describe the sound (“a whoosh on the cut at 0:12”) and Valmera finds real recorded effects from Openverse and Freesound, compares them, places the best one and tells you its source and license. You can also upload your own.