Sampling, Aliasing and Nyquist

[@laurenirenebiddle3057]

This is by far the best explanation of signal sampling and fft output analysis that I've come across thus far! Thank you so much, and well done, sir.

[@TechTins_Projects]

Great video. This stuff has always been (I once thought) way beyond me. Now I am not so sure after watching this. So clear. Well done.

[@nurahmedomar]

Great video!

[@joegodby3750]

Great. This helps me a lot for using GNU Radio and understanding the vector to steams, 1 to n and fft blocks, because I am building a circuit the measure signal phase. Have you ever used GNU radio to do this?

[@chriswilson1853]

Thanks, this is a very informative video, which makes a lot of sense. I have a question though, if I have understood the first part of this video correctly, what the Nyquist-Shannon sampling theorem really tells us is that any signal of a frequency that is less than half the sampling frequency can be captured perfectly, but only if there are sufficient cycles of the waveform, the number of which are needed depends on the Ft/Fs ratio you talk about. What are the implications of this for digital audio reproduction? How does 44.1KHz CD audio accurately reproduce high frequencies since the audio waveform is ever changing and doesn't repeat?

[@dyonisiusdony]

Interesting. But have you read the paper by Piya Pal and Palghat Vaidyanathan. They were talking about coprime sampling but it is for power spectral density reconstruction instead of signal reconstruction

[@LanceFromYangon]

Thank you very much for this video, sir! It has been very informative for me. (:

[@ayazdani88]

Good informative expositions; thank you. One question though: what if the ratio of the signal frequency to the sampling frequencies is not a rational number? How does the analysis work?

[@DansFlix]

It doesn't need to be a rational number in any sense. The only potential problem is if you don't have synchronization between the source and the capture, then you end up smearing the energy across multiple bins. That is a problem, but you can Window it using a Blackman-Harris or other window to coalesce the energy into just a few bins and sum them together.

[@ayazdani88]

DansFlix Thanks Dan. My interest and applications are digital control of PWM converters in which, by means of the hardware, the sampling frequency is indeed synchronized to and a submultiple, say 1/10, of the PWM fundamental frequency. So, based on your explanations, I will have 10/2+1 = 6, bins. Am I correct? Also, in your example, where is the number 32 coming from? Forgive my ignorance.

[@PesquisasMultiversos]

You are a genius!

[@DanBullard]

Thank you very much. I try.

[@maziarghorbani]

Great video, THANK YOU

[@MrLennon70]

Thanks, you explain it well!

[@BrentCartwright-DocsLabAudio]

this applies to video/audio as well as electronic circuitry ? peace hope love doc

[@DanBullard]

brent cartwright Yes it does.

[@BrentCartwright-DocsLabAudio]

in my daw (digital audio workstation), ive been told I can timestretch, or tune (up/down) the original (recorded) signal, +/- 10% or so, before it starts to distort the wave, and start causing audio disfiguring (sounding like crap), worbling, Donald duck sound Ect. do you know of a way to compensate for that ? also want to use the advice you may give in the video realm. any advice you could give would be greatly appreciated.! peace hope love doc

[@BrentCartwright-DocsLabAudio]

sir, im on a tight budget, but need a simple program to correct these problems in my DAW, do you have any suggestions. peace hope love Doc

[@DanBullard]

brent cartwright I'm asking around the audio community to see if I can find an answer. With some poking around I found lots of info on cubic spline interpolation. Look for an app that does cubic spline interpolation.

[@BrentCartwright-DocsLabAudio]

That is so cool dude let me know if you get any more videos up peace hope love I might call every now and then and ask you a couple questions thanks dude let me know if you need anything peace

[@BrentCartwright-DocsLabAudio]

I found a program called Serato but they want like $800 for it so I'm trying to find a free open source 1 then I can use and adapt to how I want it

[@BrentCartwright-DocsLabAudio]

What kind of interests do you use your studies, and knowledge for ? peace

[@BrentCartwright-DocsLabAudio]

That Prog I have in a lot of my software it's a connect A to B location my program uses it a lot for setting paths when doing automation of different envelopes in the audio interface kind of a connect the dots type of thing but what I was trying to find as when I x stretch my audio or change the pitch of it the program seems to go plus or minus 10 percent of the original waveform then it starts to distort or starts to sound like Donald Duck I was hoping to be able to find an answer to that, without having to buy an 800 dollar Prog the cubic splines are used in the DAW for different things, but this isn't one of them

[@codered5867]

sir what will happen if m=8 ie if fs=2ft. we will only get samples of 0's.

[@DanBullard]

code RED It depends on phase. With an M of 8 you get 4 samples per cycle, so I think I will explain it with an M of 16, because in this example N = 32. If Fs=2*Ft then the two samples will fall 180 degrees from each other. Will that be zero? It depends on what the initial phase is. If the samples happen at 90 degrees and 270 degrees, then you get the peaks. If the phase is at 45 and 315, you get + and - 0.707V. If the phase happens to be zero, then you get zero. And that is the problem with Nyquist, the amplitude in the Nyquist bin (N/2) depends entirely on the phase at which the wave was captured. I never use the Nyquist Bin, although Teradyne (aka Nextest) uses it on a product that I hold a patent on, which bothers me a lot. Using Nyquist is extremely dangerous.

[@BrentCartwright-DocsLabAudio]

vid, im tring to solve some distortion/aliasing of audio, in my studio, when adding processors, time stretching, and tuning recorded parts thanks. peace hope love doc

[@BrentCartwright-DocsLabAudio]

another one you might like, "harmonic visualizer" peace

[@DanBullard]

brent cartwright That is really cool! Right up my alley, as I am about to release my new book on Harmonics. Thanks!

[@BrentCartwright-DocsLabAudio]

Cool, am glad I could help, did you check out the program called spear and the project called sonic visualizer also check out smcnetwork.org/resourses/centers. And check out the Queen Mary universities -sound software project Let me know when I can get a copy of your book. And send me your email address peace hope love Doc

[@BrentCartwright-DocsLabAudio]

thanks dude, i found a few universities that have some experimental programs that im investigating, but will follow up on your link, pro tools is expensive, and i dont think the plugins will work in my studio, but ill ask. how have you been doing ? what you been researching lately ? interesting ? one you may be interested in is a program called "spear", its a free research tool, for manipulating music/waveforms and thier components. peace hope love, doc

[@BrentCartwright-DocsLabAudio]

somethimg that may interest you, scilab, gnu octive, canopy, and ipython peace hope love doc

[@BrentCartwright-DocsLabAudio]

also, anaconda. peace

[@LawrenceDuffy477]

Can't hear it, not going to waste my time.