Experiment X: Audio and Image Filtering

Transcription

1 Experiment X: Audio and Image Filtering Table of Contents Purpose... 2 Background... 3 Laboratory:... 7 Task 1: Hardware Component... 7 Equipment:... 7 Run Down... 7 Task 2: Software Componen... 8 Equipment:... 8 Run Down... 8 Procedure... 8 Hardware Experiment... 8 Software Experimen Suggestions for further reading Revised February 3 rd, P13361 MSD Group

2 Purpose This experiment involves hardware and a softwar to demonstrate the influence of filters in audio and imaging. The audio filtering board (Figure 1) is used in the hardware part of the experiment to examine how audio filtering is performed. The board has an option for High-Pass, a Band-Pass and a Low-Pass filter. The filters are used to demonstrate audible change of an audio signal. To demonstrate image filtering, a MATLAB GUI was design to easily filter images (Figure 2). Figure 1: Audio Filtering Kit Figure 2: Interactive Image Filtering Matlab Tool Bo Revised February 3 rd, P13361 MSD Group

3 Background Filters are used in a variety of configurations to perform a multitude of tasks within th engineering world. Whether used to clean up the biomedical signals from an EKG monitor or to select specific frequencies bands on a military grade radio, they provide an invaluable tool to today s design engineer in implementing frequenc-based solutions. Manipulation of differen frequency bands, including the electromagnetic spectum for radio signals as well as the human range of audible sound frequencies, allows the designer to narrow their range of inputs or selectively determine the desired output In general, electrical filters can be defined by in terms of their response to alternating current electrical signals in terms of a band of frequencies denoted by either radians per second (r/s) or cycles per second (Hertz, or Hz). The major types of filters used include low-pass, highpass, band-pass, and band-reject filters. Low-pass filters will pass lower frequencies (defined by a cutoff frequency) and attenuate higher frequencies. Hig-pass filters do the opposite, passing the highs and attenuatin g th e lows. Ba-pass filters select a band of frequencies to pass through, and atteuate all others. Band-reject filters will severely atrouanr a smrcrfic baoa o certain frequencies and pass all others. The primary components used to specify any filter type include a pass band, a stop band, and a cutoff frequency. The pass band defines the range of frequencies that are ideally not attenuated at all, as these frequencies pass through. The stop band includes those frequencies that are heavily attenuated and are not desired in the output. The cutoff frequency is the frequency where the magnitude response drops by 3 decibels from its pass band value. A typical low-pass filter magnitude response is shown below: Figure 3: Low-pass filter Bandwidth Revised February 3 rd, P13361 MSD Group

4 Source- Of the types of realized filters available, a common example is the lowly RC analog filter. These simple filters use the variable frequency response of passive linear circuit elements to shape and alter the AC signals that pass through them. Different configurations can yield lopass, high-pass, band-pass, and band-reject filters depending on the position of tr filnra elements within the circuit. These filters can be cascaded to chain multipl desired frequency responses together. A typical first order low-pass RC circuit is shown below: Figure 4: Low-Pass and High-Pass filter Source- For more advanced designer capabilities, tr omraatooal aomlrfira cao br usra no produced more complex and intricate circuits with more desirable response characteristics. The fascinating behavior tat op-amps provide allows for incredibly interesting combination s o otherwise very abstract components. Analysis is done through nodal equations written at th input nodes which describe the currents flowing, and determinants and Cramer s Rule can be used to solve for the resulting equations. A mor e advance d ban d pas s circui t utilizi ng t powerful op-amp is shown below: Figure 5: Filter using an Operational IpiFfier (O-Amp) Source- Revised February 3 rd, P13361 MSD Group

5 Depending on the number and configuration of the components used, the designed filters have a specified order. This order defines the response of the circuit and the transfer function used to describe t. In general, the more capacitors or inductors that are used in a filter, the higher the order of the filter. In additio, trgtra oaara filnras cao br oaade by stringing multiple st or 2 nd order filters together as such: Figure 6: Cascading Filters Source- The magnitude response of the filter with respect to the frequency axis is the main area of focus for most filter designs. However, often times the phase response ntr filnra trll tavr a significant effect on the operation of the circuit. Some circuits are designed purely to affect the phase response of a circuit and leave its magnitude component unchanged. An all-pass circuit produces constant gain at all frequencies, but due to the nature of the op-amp circuitry, will introduce phase into the circuit where desired. An example is provided here: Figure 7: Active Filter Circuit using O-Amp Source- Revised February 3 rd, P13361 MSD Group

6 Figure 8: PCB Block Diagram Figure 9: ISD Button TnfiauratTn Revised February 3 rd, P13361 MSD Group

7 Laboratory: Figure 10: Instek Supply Front Panel Task 1: Hardware Component Equipment: 1. Audio Filter Kit (Recording circuit and pre-built filters) 2. Instek Power Supply 3. Power Cables (Banana Male-Male power supply cables) 4. Data Acquisition cabl (Male-BNC DAQ cable) 5. MP3 player, Computer or Smartphone. Run Down Record 10 seconds from microphone and play unfiltered. Playback audio recorded using High-Pass, Band-Pass and Low-Pass filter option Play audio of choice from audio jack using FT (feed-through) feature. Filter feed-through signal using High-Pass, Band-Pass and Low-Pass filter option Build a High-Pass, Band-Pass and Low-pass filter on breadboard and play feed-through/mic audio through the filters built. Acquired unfiltered audio signal using Labview and compare to filtered audio signal Revised February 3 rd, P13361 MSD Group

8 Task 2: Software Componen Equipment: 1. Computer running MATLAB with Image Processing Toolbox. 2. Copy of Filters_Laboratory.zip 3. Cellphone camera or.jpg.png.tif equivalent. Run Down Capture an image from your cellphone or find one off the internet. Filter the image using IMGPRO_TOOL using a HighPass Filter, Low Pass filter, Gaussian Blur, Edge detection Use predefined images to compare different filters. Deconstruct the Image in terms of its magnitude and phase. Procedure Hardware Experiment Part A: Record 10 seconds from MIC and play unfiltered 1. Verify switch on the Filtering board is OFF. 2. Turn ON the DC power supply and set to output 9V DC from the power supply to connect to (+9V) yellow input on the front panel. Connect ground (GND) to white input on the front panel. 3. Turn switch on the Filtering board ON. 4. Verify the RED LEDs on the Filtering and ISD boards are ON. 5. Connect a Speaker or Headphones to the 3.5mm Audio Jack output. If you are using the headphone provided by the professor or teacher assistant, set the switch on the headphones to MONO. 6. Press and Hold the REC button on the ISD board while speaking into the microphone for 10 seconds. 7. Place Jumpers on pins 3-4 of JP1 and JP2 (labeled Unfiltered). 8. Press PLAY on ISD board to hear the recorded sound unfiltered. Question: What is the bandwidth of the human voice Part B: Playback audio recorded using High-Pass, Band-Pass and Low-Pass filter optio 1. Repeat steps 1 through 6 from part A. 2. Place Jumpers on pins 5-6 of JP1 and JP2 (labeled HPF). 3. Press PLAY on ISD board to hear the recorded sound filtered through a High-Pass Filter. 4. Place Jumpers on pins 7-8 of JP1 and JP2 (labeled BPF). 5. Press PLAY on ISD board to hear the recorded sound filtered through a Band-Pass Filter. 6. Place Jumpers on pins 9-10 of JP1 and JP2 (labeled LPF). 7. Press PLAY on ISD board to hear the recorded sound filtered through a Low-Pass Filter. Revised February 3 rd, P13361 MSD Group

9 Question: Calculate the cut off frequency of a Hig-pass with R=1.6kΩ and C = 1uF. Part C: Play audio of choice from audio jack using FT (feed-through) feature 1. Repeat steps 1 through 5 from part A. 2. Place a Jumper on pins labeled FT on ISD board. 3. Connect a Male-Male Audio Jack cable from the front panel to MP3 player, computer or Smartphone audio output. 4. Place Jumpers on pins 3-4 of JP1 and JP2 (labeled Unfiltered). 5. Play audio of choice to hear the sound unfiltered Questio: Wtrct filnra staarpens the audio Part D: Filter feed-through signal using High-Pass, Band-Pass and Low-Pass filter optio 1. Repeat Steps 1 through 3 from Part C. 2. Place Jumpers on pins 5-6 of JP1 and JP2 (labeled HPF). 3. Play audio of choice to hear the sound filtered through a High-Pass Filter. 4. Place Jumpers on pins 7-8 of JP1 and JP2 (labeled BPF). 5. Play audio of choice to hear the sound filtered through a Band-Pass Filter. 6. Place Jumpers on pins 9-10 of JP1 and JP2 (labeled LPF). 7. Play audio of choice to hear the sound filtered through a Low-Pass Filter. Questio: Wtrct filnra alllows the bass to be heard Revised February 3 rd, P13361 MSD Group

10 Part E: Build a High-Pass, Band-Pass and Low-pass filter on breadboard and play feed-through/mic audio through the filters built. 1. Repeat steps 1 through 5 from part A. 2. Build the Schematics shown onfigure X using the breadboard next to the Filtering Board. Figure 10: (a) High-Pass Filter (b) Low-Pass Filter (c) Band-Pass Filter 3. Connect the Audio signal from the Filtering board to the High-Pass filter built on the Breadboard. 4. Connect the output of the High-Pass filter built to the Signal IN on the Filtering board. 5. Play audio of choice to hear the sound filtered through the High-Pass filter built on the Breadboard. 6. Connect the Audio signal from the Filtering board to the Low-Pass filter built on the Breadboard. 7. Connect the output of the Low-Pass filter built to the Signal IN on the Filtering board. 8. Play audio of choice to hear the sound filtered through the Low-Pass filter built on the Breadboard. 9. Connect the Audio signal from the Filtering board to the High-Pass filter built on the Breadboard. 10. Connect the output of the Band-Pass filter built to the Signal IN on the Filtering board. Revised February 3 rd, P13361 MSD Group

11 11. Play audio of choice to hear the sound filtered through the Band-Pass filter built on the Breadboard. Softwar Experiment Question: Why does the Ban-Pass attenuate the signal more than the Hig-Pass or Low- Pass filter? Part A: 1. Extract the contents of Filters_Laboratory.zip on the desktop of your machine. 2. In MATLAB change the directory to Filters_Laboratory on the desktop. 3. In MATLAB command prompt type in MSD_GUI 4. Push the Cameraman Button in the GUI. 5. Push the HighPass Filter button, vary the slider, Observe and record any visual changes. 6. Push the LowPass Filter button, vary the slider, Observe and record any visual changes. 7. Push the Blur Filter button, vary the slider, Observe and record any visual changes. 8. Push the Edge Detection Filter. Observe and record any visual changes. Note: Slider is deactivated in this section. 9. Repeat for Rice Image. Question: Which filters can be used to reveal the edges and corners of an image? Part B: 1. Upload from your cell phone an image of your choice, or find an image from the net of your choice. Note the Software GUI automatically resizes the largest dimensions of the image to 640 pixels. 2. Repeats Procedure A. Note you can use the black and white feature and zoom in and out for your convenience. Question: Which filters can be used to smooth or unsharpen an image? Part C: HighPass and Edge Detection. 1. Click on Open Image and open the image titled HighPass_Example_1.JPG 2. Click on the HighPass filter button, vary the intensity. Toggle between Black and White as you are varying the filter. Record your observations. 3. Run the Edge detection filter, toggle between black and white and record your observations. 4. Repeat For HighPass examples 2 and Compare the HighPass filter with the Edge Detection filter. Questions: Compare each images after each filter. What information is left after each filtered image? What do you think was done to each example image? How can you tell? Revised February 3 rd, P13361 MSD Group

12 Part D: Low Pass and Gaussian Blur 1. Click on Open Image and open the image titled LowPass_Example_1.JPG 2. Click on the LowPass filter button, vary the intensity. Toggle between Black and White as you are varying the filter. Record your observations. 3. R Click on the Blur filter button, vary the intensity. Toggle between Black and White as you are varying the filter. Record your observations. 4. Repeat For LowPass examples 2 and Compare the LowPass filter with the Gaussian Blur Filter. Questions: How does the Low-Pass filter compare to the Gaussian blur? The first picture is an example of an image that has been sharpened. How would you go about unsmoothing the image and making it more like examples 2 and 3? Part E: Magnitude and Phase. 1. Reopen the Cameraman Image and click on Phase. Which Image more closely represents the original cameraman image? 2. Repeat with rice image, and any other image of your choice. Question: Which component contains more information of an image signal? Thmagnitude or the phase? Revised February 3 rd, P13361 MSD Group

13 Suggestions for further reading. Revised February 3 rd, P13361 MSD Group