You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During parse the logic about process MRCP Signal of unimrcp, I found the architecture only surpport process it by single-thread? is it right? On the other hand, dose such a architecture will limit the maximum number of occurrences?
The text was updated successfully, but these errors were encountered:
Issue Description:
We have encountered a problem with the following function:
unimrcp/plugins/my_recog_engine.c
/** Callback is called from MPF engine context to write/send new frame */
static apt_bool_t otyun_recog_stream_write(mpf_audio_stream_t *stream, const mpf_frame_t *frame)
When this function experiences blocking under high concurrency, it causes noticeable stuttering in TTS playback. The stuttering issue disappears when we deploy MRCP speech recognition and speech synthesis separately.
Current UniMRCP version: 1.5.0
During parse the logic about process MRCP Signal of unimrcp, I found the architecture only surpport process it by single-thread? is it right? On the other hand, dose such a architecture will limit the maximum number of occurrences?
The text was updated successfully, but these errors were encountered: