This adds a sliding window buffer approach to the audio sample data as well as
force the user to set the sampling rate of the input data. This was needed to
stabilize fft output to a usable degree. Also the fft output is now in db, which
makes it a lot better to look at.
Signed-off-by: tyrolyean <tyrolyean@tyrolyean.net>
Texture channel 0 has a 512x2 texture with audio samples. For now,
fill it with dummy data (ascending slope), so we can see some
output.
Signed-off-by: David Oberhollenzer <goliath@infraroot.at>
- Fix uniform name for iFrame
- Move timer initialization to main loop, so we don't measure
the FBO initialization
Signed-off-by: David Oberhollenzer <goliath@infraroot.at>
Unbind the FBO+texture after initialization, bind it before
rendering, unbind it after, re-bind the texture to scrape the
data, unbind the texture.
Signed-off-by: David Oberhollenzer <goliath@infraroot.at>
