Audio breaking/spikes at lower voiceRate values (0.8 / 0.9) on all video quality levels

When testing the voice.rate parameter at 0.8 and 0.9 with low, mid, and high video quality in the Streaming Avatar SDK, I’m hearing consistent audio issues.

The problem sounds like short, sharp breaks or missing audio packets, visible as spikes in the waveform. These spikes can occur multiple times during a single sentence. They are not limited to one per sentence — rather, the longer the avatar is speaking, the more likely it is that at least one or more spikes will be heard.

When using voice.rate 1.0, the audio remains clean with no spikes/audio breaking.

Is this an expected side effect of slower voice rates, or could it be related to encoding or streaming?
Has anyone found a way to maintain a slower speaking pace without introducing these packet-like breaks?