A comparison of sinusoidal model variants for speech and audio representation | IEEE Conference Publication | IEEE Xplore