I made a simple sound visualizer using Unreal 4
polycounter lvl 5
It took me some spare time to learn how a sound visualizer works, and implement a custom beat detector in c++.
The beat detector part is fairly technical, so i'll skip (the core idea is to detect the onset signals).
The rest parts are generated in realtime with Niagara. Im using UE4.25, the new Audio Spectrum and Oscilloscope interface helps some parts.
[massive GIFs follow]
Or, to manually type in the music notes (bad compression):
[video, not sure why embedding doesn't work.]
https://www.youtube.com/watch?v=LcFdcs5Q2Vo
I name this scene "Life of Sounds" because I wish to visualize sounds as a live entity. It's both organic and cyber-fi. So when every note reaches the central processing core, it fans out a electric "root" that shaped like wires on a PCB. The root goes out to draw "nutrition"(energy) to the core. The "tree" . in the meatime, absorbs the energy and grows.
When the music reaches the end, the sound fades, and the tree together with its roots will gradually wither and disappear, like the way the nature intended.
The hexagon shaped ground is also inspired by nature's common sight. Like bee-hive or the very famous Icelandic hexagon rock formation:
The rest parts are Niagara related.
The root is implemented using a simple path decision tree. (optimize it by using float mul) . When sound intensifies, the energy it draws also increase, creating flashing roots.
To enhance the ground part's organic feeling, I also added some fur bending effect to the hexagons, like this:
Then give it a water simulation to visualize big beats (yellow ripples).
The tree is just a inward pressure sphere with some elementary physics calculations to make ribbons flow inside its radius.
The core uses a Fibonacci sequence to map the audio spectrum to the sphere, nothing magical here.
The rest parts are just some simple lighting and post-processings, like height fogs, chromatic abbreviation, exposure and DOF. All are blueprints controlled.
Sadly this scene doesn't work well with UE raytracing, otherwise the visual could be prettier.
Hope you guys like this.
The beat detector part is fairly technical, so i'll skip (the core idea is to detect the onset signals).
The rest parts are generated in realtime with Niagara. Im using UE4.25, the new Audio Spectrum and Oscilloscope interface helps some parts.
[massive GIFs follow]
Or, to manually type in the music notes (bad compression):
[video, not sure why embedding doesn't work.]
https://www.youtube.com/watch?v=LcFdcs5Q2Vo
I name this scene "Life of Sounds" because I wish to visualize sounds as a live entity. It's both organic and cyber-fi. So when every note reaches the central processing core, it fans out a electric "root" that shaped like wires on a PCB. The root goes out to draw "nutrition"(energy) to the core. The "tree" . in the meatime, absorbs the energy and grows.
When the music reaches the end, the sound fades, and the tree together with its roots will gradually wither and disappear, like the way the nature intended.
The hexagon shaped ground is also inspired by nature's common sight. Like bee-hive or the very famous Icelandic hexagon rock formation:
The rest parts are Niagara related.
The root is implemented using a simple path decision tree. (optimize it by using float mul) . When sound intensifies, the energy it draws also increase, creating flashing roots.
To enhance the ground part's organic feeling, I also added some fur bending effect to the hexagons, like this:
Then give it a water simulation to visualize big beats (yellow ripples).
The tree is just a inward pressure sphere with some elementary physics calculations to make ribbons flow inside its radius.
The core uses a Fibonacci sequence to map the audio spectrum to the sphere, nothing magical here.
The rest parts are just some simple lighting and post-processings, like height fogs, chromatic abbreviation, exposure and DOF. All are blueprints controlled.
Sadly this scene doesn't work well with UE raytracing, otherwise the visual could be prettier.
Hope you guys like this.