Short answer : Yes, a delay of 40-50ms is actually on purpose.

Long Answer:

Wallpaper Engine only supplies audio data at a rate of 20-25 updates per second. This means that if you want to create a 'real-time' audio visualizer, you are capped at 20-25 FPS.

To enable smooth animation, and be able to render the wallpaper at much higher framerates ( like 60fps ). I have to 'remember' the last piece of audio data received and 'morph' it to the current audio state. ( so I can generate the 'frames' to fill the gaps )

This results in a delay of 1/20th - 1/25th of a second ( 40-50ms )

But trust me, it's much better compared to having only 20-25 FPS :P

Any other delay you might experience is due to delays in Wallpaper Engine capturing the audio and 'posting' it to the Wallpaper running inside a different process. ( which I have never measured, but should in theory be negligible )