A music bed under the narration - #293
Merged
Merged
Conversation
The spot was voice or nothing. A bed now sits under the read, looped to the length of the picture and faded at both ends, because on a five second spot an abrupt bed is most of what you hear. Held at -16dB. Broadcast practice is 15 to 20 dB down: enough to be felt, not enough to compete with the read. Given in dB rather than a linear figure because that is the unit the decision is made in. amix normalises by default, which would have ducked the voice along with the bed; normalize=0 keeps the read where it was. Two sources at these levels cannot clip. The bed is a path from AD_MUSIC_BED rather than a bundled asset: the right music is a brand decision and changing it should not need a deploy. Unset, and with no bed passed, every existing path is byte for byte what it was, which a test asserts. A bed with no voice is ignored, since that is just music. Worth recording: audioMode is already "narrated" by default and has been. A silent spot is not a defaulting problem, it is synthesis failing at render time, which the worker logs as "no key or synthesis failed". ELEVENLABS_API_KEY has to be present wherever the render worker runs. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
ThreatCrush Security Scan48 finding(s) HIGH/CRITICAL: 2 | MEDIUM: 31 | LOW: 15
Snippets are redacted; ThreatCrush never prints matched credential material. |
This was referenced Sep 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The spot was voice or nothing. A bed now sits under the read, looped to the length of the picture and faded at both ends, because on a five second spot an abrupt bed is most of what you hear.
amixnormalises by default, which would have ducked the voice along with the bed.normalize=0keeps the read where it was; two sources at these levels cannot clip.AD_MUSIC_BEDrather than a bundled asset, because the right music is a brand decision and changing it should not need a deploy.114 video tests pass, 7 of them new.
Worth knowing separately
audioModeis already"narrated"by default and has been. A silent spot is not a defaulting problem — it is synthesis failing at render time, which the worker logs asnarration: none (no key or synthesis failed), rendering silent.ELEVENLABS_API_KEYhas to be present wherever the render worker runs, or spots stay silent no matter what this PR does.🤖 Generated with Claude Code