user-facing:
- remember volume and repeat settings across app launches
- improved visualizer quality
- reduced download size by a couple megabytes
low-level:
- optimized
Stream
andPlayer
communication. more direct, less locking. no more weird deferred cleanup - removed
fftw3
dependency in favor ofkiss_fft
. much smaller download now - added a proper windowing algorithm to audio samples before they are handed to the fft process
- fixed audio de-interleaving issue which lead to funky fft results in some cases
- included a new spectrum visualizer sample called
GdiVis
(win32 only for now) - overhauled some parts of the audio engine to ensure buffer size is consistent during playback. this makes visualization quality better, and should make it easier to integrate with things like JACK and PulseAudio.