Skip to main content
GilloSugar
Participant
Nov 13, 2026 6:00 PM UTC
Open for Voting

Firefly Video Editor inserts unrequested background music ignoring negative prompts

Hello Adobe Firefly Team,

Following my previous bug report, I would like to submit a critical feedback request regarding the audio management in the Firefly Video Editor (Beta).

The Issue & The Evidence: Despite inserting strict negative commands in my generation prompt to prevent any musical elements, the system automatically inserts background music tracks anyway, permanently mixing them with the dialogue.

Here is the exact prompt I used for the generation:

"A photorealistic corporate video based on the reference image, featuring a professional woman with glasses sitting at an office desk, wearing a beige blazer, with her hands resting on the desk. She maintains a warm, welcoming expression and speaks directly to the camera in English with a mature, soothing, and trustworthy female voice, saying: 'Hi everyone, Alma here! I want to recommend some truly wonderful selections for you.' While speaking, she makes a gentle, elegant, inviting gesture, keeping her posture fluid and natural. ISOLATED DIALOGUE TRACK ONLY. MONOPHONIC SPEECH. ABSOLUTELY NO INSTRUMENTAL TRACK, NO BACKGROUND MUSIC, NO MUSICAL SCORE, NO AMBIENT SOUNDTRACK, NO MELODY OF ANY KIND. High-end professional office background with a white orchid and a framed city photo, cinematic warm lighting, high quality."

Why this matters for professionals: As a professional visual creator, complete control over the audio mix is mandatory. Ignoring explicit prompt instructions regarding audio, forcing unrequested music, and offering no split tracks for voice and sound bed ruins the professional quality of the asset and requires tedious third-party audio editing workarounds.

Proposed Solution:

  1. Ensure the generator strictly obeys negative audio prompts (respecting NO BACKGROUND MUSIC commands).

  2. Make background music entirely optional and provide separated audio tracks on the timeline (Dialogue/Voice distinct from Music/SFX) so editors can control or export clean stems.

Thank you for listening to the creator community.

Best regards,

Gianni Lo Magno (GilloAds)

    1 reply

    Gabriel Grundler Y.
    Community Manager
    Gabriel Grundler
    Nov 13, 2026 6:29 PM UTC

    Hello, GilloSugar!
    Thanks for sharing this feedback! Separated audio tracks on the timeline and better control over background music are great suggestions as we continue developing the Firefly Video Editor Beta.

    As a quick workaround for your current projects, try removing negative commands like 'NO BACKGROUND MUSIC' or 'NO MELODY' from your prompt. AI models often misread negative phrasing and end up adding the exact words you mention. Replacing those lines with positive descriptors like 'clean acapella speech, quiet room tone, voice-over only' usually gives much cleaner vocal results!

    We appreciate you sharing your thoughts with the community, so keep the suggestions coming!

    Best,

    Gabriel

    Translate