Part Number: TMS320C5535
Other Parts Discussed in Thread: TIDEP-0077
Hi,
I have 2 C5535EZDSP boards.
- Board 1, I am using for TIESR Demo Application for detection of words- open door, lock , unlock, vent windows, etc, running at 16Khz.
- Board 2, I am using for Audio Pre-processing library for 16Khz.
The problem I am facing is when I interface these boards, I am expecting better noise reduction. Instead the speech recognition accuracy reduces drastically.
Without Pre-Processing board interfaced the recognition accuracy is very good.
I have interfaced these boards by connecting wires between L-R channels of Stereo OUT of Board 2 and Stereo IN of Board 1.
Please indicate what I can do to improve accuracy even with pre-processing module attached.
