EECE 520 Project DFT Processing & Filtering for Noisy Guitar Note REPORT DETAILS & SUBMISSION Electronically submit a written report by the due date/time that discusses the theoretical analyses you did to develop your approach and describes the processing approaches you created and discusses the results of your tests as described in the Procedure below. There will be a TurnItIn assignment link created in Blackboard for the purpose of submitting your report. (Do not include your code in this submission – see below) Turnitin will check your submission for plagiarism against others in the class as well as online and book sources. You should read the following link to make sure you understand what is considered to be plagiarism: Link to Academic Honesty Policy. Your MATLAB code must also be free of plagiarism – you are expected to write your own code! In this project you may not use MATLAB code obtained elsewhere; you may use any standard MATLAB function but you may not use code obtained elswhere. The format of the report should be as follows: • Include an Introduction section that provides an overview of the report • Include a separate section for the body of the report: o Use class theory to support and/or explain what you observed. o Discuss HOW your processing works, WHY it works, and discuss HOW your test results support the claim that it works. • Include a Conclusions section that summarizes your performance testing and the results. o Use concepts and methods from the course to summarize your results and describe how/why they match (or do not match) expectations. o Use class theory to support and/or explain what you observed. o A common error is for students to write a very minimal summary section that does not really say much. Avoid that! o Do NOT just give an overview of WHAT you did • Write your report NOT as if you are writing to a professor but rather like you are writing to your co-workers and manager to explain the essence of what you did, what your results were, why they seem reasonable, and what they mean. o Don’t give details on exactly what was done (in other words, avoid a step-bystep “travelogue” discussion such as “First we did this and then we did that”… Discuss what the results show and what they mean… discuss them in terms of the theory we studied in class. • Make your report looks professional!! o Use the MS word equation editor to make professional looking equations It is pretty easy to use – ask me for help if you need it! o Use equation numbers for any equation set off from the text and refer to the equations by number rather than saying something like “… in the equation above…”. o Use figure numbers and captions and refer to figures by number rather than using something like “… the figure below…” Embed the figures into the report where they are discussed – not at the end! See the MATLAB tutorial I posted for a discussion of how to paste a MATLAB figure into a document. o Yes… spelling and grammar count! Clarity of writing also counts! Write clearly so that your reader can understand your writing. Be careful when using pronouns such as “it”, “this”, “that”, etc. so you can be sure your reader knows precisely to what you are referring! • Do NOT include your MATLAB code in the report – see below for how to submit that. CODE DETAILS & SUBMISSION Make sure your code is well-commented so I can easily determine what you are doing and why. Email your code with the subject exactly like this: EECE520 Project Code Send your code to
[email protected] Attach your m-files to the email Do NOT put them into a zip folder or any other equivalent form!! ASSIGNMENT I have posted a wav file of a guitar playing the note G. Each note on a guitar can be thought of as being a periodic signal (at least over a short duration). The standard fundamental frequencies for this note is 196 Hz but the guitar I was playing was slightly out of tune so these fundamental frequencies are not quite what you’ll see in this wav file. Since the note is viewed here as being a periodic signal it will have many harmonic components – the fundamental is called the first harmonic, the component at twice the fundamental frequency is called the second harmonic, …. the component at times the fundamental frequency is called the th harmonic. The specific tasks you must address are 1. Determine the lowest SNR at which you can use a DFT to reliably detect all of the harmonics up to the 10th harmonic. o This will require both theoretical analysis and simulation, the later will verify the former. 2. For the SNR determined in the first task, design and test a comb filter that will pass the harmonics of the guitar signal but suppress the noise at the other frequencies such that the output SNR is 60dB lower than the input SNR. o To design the comb filter you should first design a linear phase FIR filter as your prototype and then “combify” it to get the desired filter o You should use simulation to assess the output signal’s SNR o You should also compare the filter’s output to the non-noisy guitar signal in both the time domain and frequency domain. Note that you can use the randn command to create a Gaussian white noise and you can use the MATLAB command “var” to compute the power of a signal, which then allows you to compute the SNR in dB of the signal plus noise: >> v = a*randn(1,length(x)); % choose value of a to adjust the power of the noise >> Pv = var(v); >> Px = var(x); >> SNR = 10*log10(Px/Pv); >> x_n = x + v; % this will be a signal with the computed SNR You should provide a thorough analytical assessment/development as well as MATLAB code of the processing you used and the results of that processing.
欢迎咨询51作业君