Filter (signal processing) Puns

I imagine they do exist and have some application, but I have never heard anyone talk about them. I can't even find any examples of what they would look like.

Hi i'm searching the best way to clean up my composite video signal from my ld player and vcr

i saw on internet some people recommending dvd recorder for there good comb filter

also is it worth it to have a tbc (looks already stable) ?

so if you know somme must have hardware for process the signal i can be interested :)

like always thanks

Appreciate any feedback in advance.

I have a calibrated sensor (flow meter) that has quite a bit of noise on it, it is shielded and grounded but the signal is still noisy, and while we are working on a fix for the sensor, I unfortunately have already collected the data so I cannot retest with a different setup.

QUESTION: Is there an ASTM or ASME standard out there that has allowable processes/acceptance criteria for noise filtering in POST PROCESSING? Whether it be averaging, FFT, low/high/band pass filtering?

Some of the noise results in instantaneous values (maybe 1/10 samples per second) below my internal acceptance criteria, but what I am measuring is fairly steady.... I was wondering if there was an "industry accepted" standard/process out there... I am in O&G but I cant find anything in API specifically.

Hi all, I'm under an undergraduated engineering project at the moment, and I'm a bit stuck here. To be more precise, I'm stuck at the difference between theory and practice.

Context: use a laser scanner to evaluate the vibration of a specimen, frequency spectrum of concern: 0 to 500 Hz (or just low frequency, in general), concern about Transfer function (because I want/have to investigate of the transferring process, from the input force to the output vibration), and in dB (logarithmic scale) because I want to see all peaks within the spectrum.

So, when I apply low-pass filter (for example: at 100 Hz), in theory, the signal should be cut off when the frequency is above 100 Hz. However, when I check back with the Transfer functions (with and without the low pass filter), they are both the same.

***

On the one hand, the thing we actually measure (at first) is the vibration of the specimen in time domain. So if we cut off and remove a certain chunk of frequency (>100 Hz in this case), the vibration will not be the same. This is a contradictory, and therefore, after 100 Hz, the spectrum would be the same

On the other hand, in order to use the frequency spectrum, we (or to be more precise, the machine and program) would use FFT (Fast Fourier Transform) to convert time domain to frequency domain. And if we apply the 100 Hz low-pass filter, we simply cut off the higher frequency range. Which means after 100 Hz, the values of the spectrum are simply converted into 0.

I'm just confused here. Is there any comment on this?

Hi everyone, i would like to get your suggestions for a digital filter implementation in the following type of signal.

https://preview.redd.it/0ns0klx8z9w51.jpg?width=1323&format=pjpg&auto=webp&s=95073ded90a7dd5f5b1225606d05550ae7ee035e

This signal corresponds to a diferential pressure sensor installed on a venturi and i think that the noise its due to TURBULENT flow so, not exactly electrical noise.

I was thinking of some High order IIR or FIR filters, im open to any suggestions :)

Thank you

The most programming I've successfully done is math and physics homework in matlab and some leetcode for python basics. I have everything set to go except for a project to work on. Where do I start?

Hi, I am currently working on a Blind Spot Detection System with STM32F103. The type of radar system I am using is a Continuous-Wave IQ (quadrature) Doppler radar(Not FMCW). This is what I have done so far in programming ADC -> FFT -> Speed, obviously this is not complete as I am missing out on signal processing.

Here is a hand-drawn I-signal, sorry for the bad image quality and hand drawing skill.

Region a - when target is stationary

Region b - when target moving

Any advice on how to remove the noise in Region a?

Any other signal processing is needed for CW IQ doppler radar?

For the direction of motion I have implemented based on the method in the link below, if you know any other method to determine the direction of motion please let me know - https://www.infineon.com/cms/en/product/promopages/makeradar/makeradar-school/programming-tutorial/#5

Figure 1: I-signal in time domain

Thank you.

*Note - I am new to Signal Processing and this is my first time doing a radar system.

Edit - Thank you everyone for the wonderful advice and suggestion.

I'm trying to filter out a known frequency in a step response measurement. Currently, I'm using a basic "moving window average" method. This method works for the most part, but it slows down the 10%-90% response time by quite a bit. I've read that moving window averaging is pretty good for random or unknown frequencies, but I know the frequency of my noise, so is there something better I could be using?

Here's a picture of my current method: https://i.imgur.com/0BGltMT.png

Blue is the original measurement, and orange is the averaged measurement.

Also I found this peer reviewed article that has a solution to what I need, but I'm having a hard time understanding it

Also if this is the wrong place to be asking, would you please point me in the right direction. Thanks!

When processing a signal obtained from an experiment and supposing you have a PC with enough computation capabilities to perform digital filtering "on the go" are there any pros & cons that would prioritize one over the other?

Are there any applications where only analog filtering is possible with today's computer capabilities?

Hi everyone,

I have been working with FPGA for some time, but I mainly focus on microprocessor design and communication buses (SPI, I2C, AXI, etc.)

I have been thinking on implementing digital signal processing modules (filters, FFT, etc.) and digital control circuits (like PID) by myself in order to improve my skills (instead of just using the ones given by the FPGA vendors).

I came to the conclusion that my problem is not the RTL but the algorithms. I need to improve my skills and to learn more about digital signal processing (DSP) and digital control.

I have struggled to find good content on these topics on the internet (most of the things I find on youtube just use the IPs already available from the FPGA vendors).

So the question is, does anyone has some suggestions on good books on these topics (preferably with a practical approach, I am an engineer not a mathematician). Any good Youtube channel/Website/Book that targets these type of algorithms and how to implement them on digital hardware.

Any advice on how to properly implement these systems and tips and tricks are also welcome :D

Thanks in advance.

So, our DSP professor decided to create this question that each student will have his own question based on his/her student ID, to prevent cheating or googling.

Did anyone face similar techniques? during online classes. lol

https://preview.redd.it/45jez5nji2a81.jpg?width=563&format=pjpg&auto=webp&s=9957489485b4c4a83fe1af3e162c61be243525e0

Pressing mute is not an option :)

Hello everyone!

Im working on a project where I use a SEN0161 V1 pH sensor. The kit comes with the electrode and a small PCB described as a signal processing module.

I was wondering what is actually meant with "signal processing" there and took a look at the schematic. There are two op-amps visible there, do I understand correctly that the first one (left) is acting as a active low-pass filter and the second one as a inverting amplifier? If this is the case then what does exactly the potetntiometer R5 regulate? The manual says its used to calibrate the reading in reference to mesurement of a known-pH-value solution, but I would like to know how does it exactly affect the signal.

Just to be sure I also embed the schematic of the module below.

Thank you all in advance!

https://preview.redd.it/5adrn2rjyc181.png?width=1316&format=png&auto=webp&s=f7790cdc198480a3d20cf8aa86e6cba167bcf0f0

I am new to ADSP course,and i have not covered dsp course.Guys how can i cover this gap.Because i am willing to do this course.

I work in flight control and nav avionics, and pretty much all our designs are implemented in C, C++, Ada, or (most often) in Simulink and auto-coded to C for deployment. All the code runs at a fixed Hz, but different sections of code run at different rates.

I see the same kind of patterns over and over.... like (re) initializing state variables, recording and processing data history over time, synchronizing different data sources, matrix math, timing of subsystems, simulating motion, etc etc. Simulink is capable at accomplishing the dynamic component of these tasks, but I find the graphical language very cumbersome for implementing any non-trivial algorithms. On the opposite end, general languages like C are great for implementing algorithms, but their generality makes the dynamic part harder than in Simulink.

Essentially, I'm wondering if there are any code-focused languages that are tailored to the kind of domain that Simulink is often used in (continuous / discrete time systems).

Would anyone happen to have a solutions manual or even just the final answers for problems? The course I'm taking doesn't assign homework from the textbook. It would just be really helpful if I could work through lots of problems and have some answers to compare my work to.

I see posts once in awhile where the theory of digital signal processing (DSP) could be used to solve issues with measurement/hardware requirements. A good example of this would be time averaging a noisy signal (as long as the application permits), or filtering it in DSP terms.

PLCs are not my area of specialty though I'm required to assist my team with troubleshooting occasionally, so I'm by no means an expert. I'm aware of DSP though RF engineering (specifically modulation and demodulation). Mod/demod of analog RF signals is exactly the same as measuring electrical inputs, just a different way of getting the information to/from the processor.

Has anyone implemented more simple or advanced DSP, and just wondering if it's on anyone's radar. I get that everything is application specific but wanting to put it out there, DSP can be a useful tool in problem solving.

Hello everyone,

I am finishing up my undergrad degree in mathematics and physics with a huge interest in DSP.

All my summer internships (that did not require specific degrees) were related to DSP:

1. Statistically analyzing ultrasound data for diagnosis
2. Developed neural networks for audio data that can classify musical chords
3. Applied machine learning on low and high-frequency energy data

Everyone around me who does signal processing seems to have a background in electrical or computer engineering - all of my industry advisors and professors had engineering degrees.

I am looking for permanent jobs in DSP / audio AI engineering, but they all seem to request computer science or engineering degrees. I recently applied to one and they rejected me as they were "impressed by experiences", but were suspicious of my potential without an "engineering degree".

I do believe that my math and physics courses got me really prepared for DSP (e.g. advanced calculus, real analysis, waves, acoustics), but I am worried that my "degree" will hold me back.

Am I a good fit for DSP? Should I worry about my degree?

Thanks for any advice.

Hi all! Im trying to learn DSP in C++, can anyone recommend a solid textbook I can buy?

1. https://www.researchgate.net/publication/357698587_PIEEG_Turn_a_Raspberry_Pi_into_a_Brain-Computer-Interface_to_measure_biosignals_arXiv220102228

For example: I want to use parallel compression on my drums and I could open up two channels for that routing my drums to them, one is the unprocessed (dry) signal and the other one where the compressor sits (wet). Then I can blend them together. But I just could use one bus channel with all my drums and put a compressor on it with strong settings and using the mix level knob on my effect adding slightly the compression effect maybe 10% or something. I mean you're just adding a bit of the effect but also a part of the dry signal is there.

is there any difference? or is it better to make separate channels for that. I guess you have a lot of benefits when you use two channels because you have them separated and full control over both signals, but I'm wondering if this is just the same technically.

I have just finalized a draft for my SOP, it would be great if any experts in here would guide me in making any changes or corrections or a general feedback. :) Thanking y'all in advance.

Hi,
I am a student looking to implement simple DSP algorithms like convolution, fft, ... for learning purposes for myself. I was looking for existing libraries to later compare my (noob rust and noob dsp) implementations vs more mature libraries. However I have had trouble finding any existing libraries that have these things and are also maintained. I would like to look at existing implementations and maybe even contribute to one such library?

I found Scirust which seems unmaintained and has "signal processing" as a focus listed, but not much at all implemented, and rust-dsp which has a different focus (and the last commit is from 2014).

There is Dasp which seems healthy (focusses on audio) and Basic_dsp which looks more like what I am looking for and which also seems maintained, but I am not sure if it is currently actually developed because the last updates seem to focus on dependency updates and documentation.

So my question is: Did I miss a crate? Are you using basic_dsp for example, if yes, what for? Is there generally an interest in Rust dsp libraries?

Cheers

Need to know more about Signal Processing, can anyone help me find any articles or handbooks that can help?

Also is Signal Processing Analog or Digital or is it Mixed?