Our goal is to recover an input signal from its corresponding output for a system knowing the impulse response. First, we measure the step response of the low-pass RC filter. Knowing that the derivative of the step response is the impulse response, we obtain the impulse response for the filter. Using the function generator, we generate an input square wave and pass it through the filter.
The Keithley A/D card is used to capture the output of the filter into MATLAB. The next task involves deconvolving in MATLAB the output with the RC impulse response. To verify the deconvolution, we compare the deconvolution result to the actual input.
We repeat this process for an input triangle wave. Since the impulse response stays the same, the process can easily be repeated for other input waveforms.
The project will consist of three stages. The first one of the system is a graphical equalizer consisting of five bands: 100Hz, 300Hz, 1KHz, 3KHz and 10KHz. These will provide adequate coverage of the audible range of hearing. The equalizer will allow about +/- 10 dB of gain per frequency. The controls for this will likely a microcontroller in combination with digital potentiometers. High quality operational amplifiers will be used to ensure low distortion.
The next stage of the project is the volume control and preamplifier for the following stage. This will amplifier the low-amplitude signals from the equalizer to the required levels for the power amplifier.
The final stage is the power amplifier. This will be able to drive a speaker with voltages up to +/- 38 volts with a maximum current of five amps. This enables driving 4 ohm or 8 ohm speakers. The entire system will be a mono one, with the possibility of stereo depending on time.
Extras (if we have enough time)
Our project will explore some of the simple facets of audio signal processing. We will produce two or three effects encompassing features such as delays, echoes and reverberations as well as convolution. The majority of the project will be designed and demonstrated with Matlab because of its easy implementation style.
The first part of this project will involve the generation of the dual tone, multi-frequency tones necessary to dial a specific phone number. For input, the user will enter the telephone number to be dialed as input into the MATLAB program. MATLAB will generate the tones for each digit. The program outputs the tones using the computer's speakers. To test the functionality, the microphone of a landline phone will be placed next to the compute's speaker as the program generates the tones. In this way, the MATLAB program should be able to dial the correct phone number without every pushing the buttons on the phone.
The projec's second part involves the decoding of a DTMF signal. Given the DTMF signal from a sample phone number, the MATLAB program will detect the number dialed in the sample. As a final test, a landline phon's speaker will be placed against the compute's microphone. The MATLAB program will read the tones created as someone pushes the buttons on the phone and determine the number dialed.
There are a few ways to transfer the DTMF tones into a phone number. The first way is to use a system of band pass filters to determine the two strongest frequencies present in the tone. The other, more efficient and precise method, is to examine the Fast Fourier Transform (FFT).
The project will consist of 2 parts. It is designed to portray the mood of the music by changing the intensity of the light. Being that low frequency music portrays solemnest and high frequency portrays joyousness, I propose that as the frequency increases, the intensity of the light becomes stronger. Also I decided to make the number of lights that light up increase as the amplitude of the sound increases. I will need a microphone to detect the sound and a light bulb. I will use a high pass filter in order to pass more voltage as the frequency increases. I will need to turn the analog signal of voice into a digital signal so I will need to use the Keithley.
We will research different imagine processing techniques and implement these techniques including de-blurring, noise removal, and imagine transformation. We will be using Matlab to implement these image processing techniques. We plan to use techniques such as convolutions, ideas in linear algebra like matrix transforms, and other mathematical models we will accumulate through research.
Purpose: The purpose of this project is to create a harmonics generator in MATLAB that will be able to synthesize the sound of any musical instrument by altering the amplitudes of a set number of harmonic frequencies.
Method: The program will start with a harmonic frequency given by the user and will add several multiples of the harmonic frequency to produce different sounds. The amplitudes of the multiple frequencies will be controlled with a GUI series of sliders. The program will also play the tone when prompted. Other possible additions to the program include a few preset tones to mimic certain instruments and the ability to play a series of tones to produce a song.
Supplies: MATLAB, Computer with compatible soundcard, Speaker.
For our term project, we will be creating a musical synthesizer. This will be done by utilizing our knowledge of Fourier Series, musical instruments, and MATLAB.
To begin, will start by analyzing sound waves of different musical instruments, namely a violin and trumpet. Wave files for each instrument will be referenced from the Internet. Next, the FFT function of MATLAB will be used to determine amplitude of the higher order harmonics for the sound waves produced by each instrument. This will allow us to determine the Fourier coefficients of the instruments. Once the coefficients have been determined, the instruments tones will be recreated as a sum of sine waves.
The second portion of our project will be creating a MATLAB program that allows a user to write music by manipulating the Fourier representations of the instruments. The user will first select the instrument they wish to play. Next they will input the frequency they wish to here that instrument at. This frequency will be used as the fundamental frequency for the Fourier Series representation. If the user wishes to play the instrument at more then frequency they will simply enter the frequencies as a vector. The program will then play each tone at the frequency they are entered into the vector.
As time allots, we will refine the program to allow the user to choose the length of time they would like each note to play for. Also, we hope to be able to allow the playing of both instruments simultaneously. Finally, we will increase the number of instruments that our synthesizer can simulate.
Our proposed project is to build an audio crossover for a set of speakers. The set would most likely placed in a car or used in a home theater setup and consists of a woofer, midrange, and tweeter. For example, someone replacing the stock sound system in their car might put a set of these in each front door. This type of speaker setup is referred to as a 3-way component system in car audio. Oftentimes, these component systems have all three speakers physically separate from one another, where a typical 3-way coaxial system (all three speakers in one unit) doesn't. Note: Speakers are sometimes referred to as drivers.
The crossover comes into play in these component systems by defining cutoff points so that certain frequencies of the incoming signal get sent to the correct drivers. The crossover sets up cutoff frequencies so that low frequency signals get sent to the woofer, medium frequencies get sent to the midrange, and high frequencies get sent to the tweeter. There are two types of crossovers: Passive Crossovers and Active Electronic Crossovers. Both types utilize basic circuit components like capacitors, inductors, and resistors to do highpass and lowpass filtering. However, the former splits the signal after it has been amplified and has already reached the drivers while the latter splits the signal and then amplifies each driver's input.
Building a quality crossover is very important, because unless very high dB roll-off and roll-on slopes can be achieved, the signals will get mixed. That is, sometimes a certain frequency will have both the woofer and midrange handling it. If the designer is not careful, situations in which the signal handled by the woofer might have a 3dB gain and midrange might have a -3dB gain could occur. At this frequency, the signals would cancel, leaving a holes or gaps in the music. Other times, both the woofer and midrange might have a gain of 3dB for that frequency. In that case, the signals would add and sound very loud or harsh. Since perfect, cut-and-dry cutoffs in which the frequency/Bode plots send a signal to only one of the three drivers does not exist, the mixing of two (or even three!) drivers playing the same signal has to be carefully considered.
Our group plans on building a crossover for a three way system using two, three, and possibly even fourth-order filtering systems to achieve cutoff points that permit the most music transmission and allow the least mixing between drivers. Basically, we're striving to design that perfect circuit that passes a frequency to only one driver. In the end, we should be able to run a 3-way component system in which a user places his/her head near each driver and hears the frequencies coming out of that driver that should belong to that driver.