I have a video avatar with a custom Generator. When idle with silent audio is yielded, the video playback is at full resolution. But when the Generator yields the audio playback from the STT-LLM-TTS pipeline to the video avatar...the video playback starts off at low resolution. The yielded frames are on schedule. Why?

Thanks for reaching out, Brian. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.