Bottleneck and Stream Delay in get_frames_as_ndarray() Method

I am experiencing stream delay due to a bottleneck in the get_frames_as_ndarray() method. This method is currently fetching only one frame at a time, which is causing the delay. Below is the line of code identified as the bottleneck:

ndarray_frames = self.kvs_fragment_processor.get_frames_as_ndarray(fragment_bytes, one_in_frames_ratio)

I need assistance in optimizing this method to enhance performance and reduce the delay.