The volume level increases when switching to a nested-scene source in OBS

This article discusses a problem that can occur when switching from a single source scene to a multi-scene source in OBS. If configured incorrectly, there will be a volume increase from a single source scene > a multi-scene source and a decrease in volume going from a multi-scene source to a single. The presence of the issue is only noticeable in the stream/recording and is not reflected in the monitor levels in OBS.

More Information

Take a hypothetical scenario where you are using a Skype NDI source in your OBS stream. Scene 1 holds your camera feed. Scene 2, the speakers camera feed. Scene 3 a PowerPoint presentation originating with the speaker. Scene 4 contains a ‘head-to-head’ camera shot with both your feed and that of your speaker present.

Switching between Scene 1 and Scene 2 will produce steady and consistent audio to your audience from the Desktop audio feed or a custom Down-mix channel. Switching to Scene 3, which includes the speaker and their PowerPoint presentation will result in a perceptible gain increase of the speaker’s mic from the point of view of your audience. Unless you are monitoring the live stream, there will be no noticeable increase of the levels in OBS. The increase may be uncomfortable to headphone users.

Similarly, while switching from Scene 3 to Scene 4 will result in consistent audio at the higher level. Switching back to either Scene 1 or 2 will result in an appreciable reduction in audio gain from the speaker to its original level.

There will be no gain level difference between the different scenes for the (global) audio source level setting sliders in OBS, however and OBS volume increase will be perceived by viewers.

The Fix

The issue is caused by an limitation in the way nested scenes work in OBS. When you created Scene 1 and 2, both were provisioned with an audio down-mix feed (likely Desktop Audio). When you created Scene 3, you will added your Skype NDI audio down-mix (Desktop Audio) feed to the new Scene. The issue arises because when you add the Speaker’s video feed as a nested scene into the new one, it will also introduce its own version of the Down-mix audio source into the new Scene. If you already have a default/existing Desktop Audio source on the audio channels list, OBS will hide the duplicate.

When you switch from a single source Scene to a aggregate one, the audio from BOTH the nested scene and the current foreground scene are merged together. An unintended behaviour of this presents itself as an unmonitored gain increase in the output.

The solution is to remove the Down-mix/Desktop Audio audio source from the parent (aggregate) scene, leaving it present in your single source scenes. After making this adjustment, the audio should remain consistent across scene changes without a perceptible OBS volume increase.