WebSphere Voice Toolkit does not support synthesis and playing of audio with 22 KHz voices.
Voice Toolkit commands for generating and playing audio, e.g., play pronunciation do not work with 22KHz voices. The generated audio sounds like noise.
Resolving the problem
Use the VoiceTest tool to synthesize 22 KHz audio. Note that you should use ".pcm" output files, as VoiceTest does not support .au or .wav files for 22 KHz output.