a pipeline with appsrc doesn't sound right

[bb]

> So i have an audio pipeline appsrc -> rawaudioparse -> autoaudiosink.
> The appsrc element is being fed buffers with raw audio samples (S16LE)
> from a text-to-speech application.  The 'rawaudioparse' element is
> configured for PCM, 22050 rate, mono audio.  I'm getting some audio on
> the speakers but it sounds "sped up", as if the timing is not right. On
> the other hand, the WAV file that text-to-speech application can save
> sounds normal, so it appears the pipeline is misconfigured.

some progress here;
turns out that if a *single* GstBuffer with all audio samples in it is
pushed to appsrc - all is well;

initially had a loop that would send multiple GstBuffers containing 16k
audio data chunks to appsrc (its "block" property is set to true)

not sure what the difference is but this sounds better than before