Looking for someone to help program something in udk for me. ( or externally by loading external C++ as DLL and calling it as a Native Function) https://udn.epicgames.com/Three/DLLBind.html
suggestion: Export the songs to a greyscale image using Photosounder or similar.
make sure its a power of 2 high by wide so you can do the math apppropriately.
The height is the frequency, width is length, pixel brightness is loudness.
Have you attempted to use real music converted to greyscale instead of random noise yet?
Also in the first video's description it says your doing a tutorial, hows that coming along so far?
Had it started, but realized my scope was too large. Still in the works once I can get time to record it.
And no I haven't experimented with a converted piece yet. It would take a little bit of work to sync the music start to the panning texture, but should be doable with blueprint.
The way I've gone about doing the music controlled paramiters some months back is as follows:
Broke the music track into separate wavelengths (8 total using Adobe Audition), than in 3D studio max I used the 8 separated wavelength files to move some models up and down, than I baked the animation down to boned models. Imported into UDK, than attached some cubes to the boned animation sockets with kismet and retrieved the cube positions, and used them to control material parameters,
I than set up a music track in UDK and again used kismet to load both the animation and music track at the same time so they synched up at level load.
It's a real pain to set up, I was hoping we could just do everything in UDK automatically instead of using some outside program to set everything up with manually. I Bet it can be done using the dllbind feature in UDK but the thing is I cant program worth a shit.
Replies
make sure its a power of 2 high by wide so you can do the math apppropriately.
The height is the frequency, width is length, pixel brightness is loudness.
Use the image data to drive your visualization.
I'm doing exactly this, just with a random noise based input, in these videos
[ame]http://www.youtube.com/watch?v=y55-AalGsLs[/ame]
[ame]http://www.youtube.com/watch?v=eErMi1DyOTg[/ame]
Also in the first video's description it says your doing a tutorial, hows that coming along so far?
And no I haven't experimented with a converted piece yet. It would take a little bit of work to sync the music start to the panning texture, but should be doable with blueprint.
Broke the music track into separate wavelengths (8 total using Adobe Audition), than in 3D studio max I used the 8 separated wavelength files to move some models up and down, than I baked the animation down to boned models. Imported into UDK, than attached some cubes to the boned animation sockets with kismet and retrieved the cube positions, and used them to control material parameters,
I than set up a music track in UDK and again used kismet to load both the animation and music track at the same time so they synched up at level load.
It's a real pain to set up, I was hoping we could just do everything in UDK automatically instead of using some outside program to set everything up with manually. I Bet it can be done using the dllbind feature in UDK but the thing is I cant program worth a shit.