- #1
- 4,430
- 327
I want to work on a bandpass filter program, but I'd like some realistic data to try it on. There's got to be something on the web to write/test programs with.
Any suggestions?
Any suggestions?
Finding Realistic Data for Bandpass Filter Program Testing
- Context: MATLAB
- Thread starter Pythagorean
- Start date
Click For Summary
Discussion Overview
The discussion revolves around finding realistic wave data for testing a bandpass filter program. Participants explore various sources and methods for generating or obtaining suitable data, particularly focusing on infrasound signals and amplitude vs. time data.
Discussion Character
- Exploratory
- Technical explanation
- Homework-related
Main Points Raised
- One participant seeks realistic wave data for testing a bandpass filter program, specifically infrasound data or any amplitude vs. time data.
- Another participant suggests creating custom test signals using random and sinusoidal functions, expressing concern about ensuring the data is transformable.
- There is a clarification about what "transformable" means, indicating a need to convert data from amplitude vs. time to amplitude vs. frequency and back.
- A suggestion is made to compose a signal as a sum of sinusoids, some of which should be attenuated by the filter, or to use MATLAB's built-in datasets like chirp or handel.
- A participant expresses newfound understanding of how to generate a signal using a combination of sine and cosine functions.
- There is a mention of the participant's preference for creating their own data rather than using built-in datasets.
Areas of Agreement / Disagreement
The discussion includes multiple viewpoints on how to generate or obtain suitable data for testing. While some participants suggest creating custom signals, others point to existing datasets. No consensus is reached on the best approach.
Contextual Notes
Participants express varying levels of familiarity with MATLAB and its functionalities, which may affect their ability to utilize built-in datasets. There is also uncertainty regarding the availability of specific datasets in the student version of MATLAB.
- #2
- 4,430
- 327
to be more specific, it's wave data I'm looking to transform, not image data.
Infrasound data is most relevant, but any amplitude vs. time data will do.
It doesn't even really have to be mock as long as it's downloadable and in txt or m format.
Infrasound data is most relevant, but any amplitude vs. time data will do.
It doesn't even really have to be mock as long as it's downloadable and in txt or m format.
- #3
chroot
Staff Emeritus
Science Advisor
Gold Member
- 10,270
- 45
Why don't you just create your own test signals?
- Warren
- Warren
- #4
- 4,430
- 327
chroot said:
How would I generate data? I imageine using rand and sin functions, but I want to make sure that it's actually transformable, so any mistakes are in the code and not in the data I generated.
- #5
chroot
Staff Emeritus
Science Advisor
Gold Member
- 10,270
- 45
It's just a bandpass filter, right? So any continuous (or sampled) data of any kind would be fine. I don't know what you mean by "transformable."
- Warren
- Warren
- #6
- 4,430
- 327
chroot said:
By transformable, I mean transformed from amplitude (pressure) vs. time, into amplitude vs. frequency, where you cut out the frequencies you don't want, then transformed back into amplitude vs. time.
I guess I should know that any continuous function can be transformed like that, I just want to make sure it's somewhat relevant to what I do at work (which is analyzing infrasound signals)
edit: I'll try throwing the rand function into a sin function unless you have a better idea that's relevant.
- #7
chroot
Staff Emeritus
Science Advisor
Gold Member
- 10,270
- 45
Just compose a signal that's a sum of some number of sinusoids, some of which should be attenuated by your filter.
Or, use some of MATLAB's built-in data sets like chirp or handel.
- Warren
Or, use some of MATLAB's built-in data sets like chirp or handel.
- Warren
- #8
- 4,430
- 327
Wow, didn't know about those. Thanks for your help, Warren!
- #9
chroot
Staff Emeritus
Science Advisor
Gold Member
- 10,270
- 45
Haaaaallelujah! 
- Warren
- Warren
- #10
- 4,430
- 327
I feel kind of stupid after giving this a little more thought, it's as easy as:
Amp = sin(w1*t) + sin(w2*t) + sin(w3*t) + cos(w4*t) + cos(w5*t)
where t is your equally spaced time array and the Wn are simply constants.
whew, that grain of sand was tough to climb...
I have the student version of Matlab at home, don't know if chirp or handel come with it, but I can't seem to find them. I like making my own data better though!
Thanks once again!
edit: HA! I found the chirp/handel and played them, now I see where the Hallelujah! came from..
This rocks, I could totally write music in matlab...
Amp = sin(w1*t) + sin(w2*t) + sin(w3*t) + cos(w4*t) + cos(w5*t)
where t is your equally spaced time array and the Wn are simply constants.
whew, that grain of sand was tough to climb...
I have the student version of Matlab at home, don't know if chirp or handel come with it, but I can't seem to find them. I like making my own data better though!
Thanks once again!
edit: HA! I found the chirp/handel and played them, now I see where the Hallelujah! came from..
This rocks, I could totally write music in matlab...
Last edited:
Similar threads
- · Replies 5 ·
- · Replies 6 ·
- · Replies 2 ·
- · Replies 1 ·
- · Replies 17 ·
- · Replies 4 ·
- · Replies 1 ·
- · Replies 5 ·
- · Replies 10 ·
- · Replies 1 ·