Digital Audio Signal Processing

The project we are proposing is a high-fidelity audio system. When designing the system we came across the early decision of if we should use and FPGA or DSP chip. While the DPS chips are more specialized for signal processing, the FPGA has more processing power overall. Since we wanted to incorporate room correction into the design, the extra power and flexibility was desired. We then decided on the Arty A7 35 as our base FPGA development board. It met all our requirements and came with all the extra peripherals we could need for the project. 

The system itself will comprise of a Bluetooth signal inputting a data stream to an FPGA. Using a cellular phone, we plan on transmitting a signal to and ESP 32 development board. The ESP32 will then convert the signal to and signal and send the data to the FPGA. The signal will be 24-bit and sampled at 192kHz. The board programmed to be operating separately from the Arty board in a standalone mode. We will attempt stereo sound with dual channel output using word select. 

On a breadboard, we will have a UAD1334 DAC and 1 maybe 2 SPH0645  microphones. The microphones are 8-bit, and the software will extend that to 24-bit to match the system. The audio signal after running through the Arty and software will come out to the DAC and then out to a waiting pair of speakers. The DAC has a built-in mute as well. Upon exiting the speaker, the audio and noise will be picked up by the microphones and sent back to the system. The system will then process the audio smoothing the curve and eliminating noise for optimal listening. 

The room correction software and FIR filter will do all the work in this design. The FIR filter is designed to have a frequency response of 10Hz-23kHz. It will sample the audio at 192kHz and read at 24-bits. The correction software will measure each speaker's output and relay that back per channel. The software will create filter coefficients based on the reading and correct the frequency and timing of the signal. The new signal is output, and the system continues to monitor the room for acoustic changes.