Okay, thank you.
One example he might be thinking of is if one wants a 20kHz bandwidth FFT with .1Hz bins, then it takes a lot of points.
That's another point you don't need to resolve frequencies in their own bins. The derivative of phase is frequency (i.e. Hilbert transform) and you can resolve frequency without the long FFT's.
Why are people trying to prove Nyquist doesn't work when it's been shown by measurements that it does, even in ways that are difficult to understand? Perhaps that is the real issue, people can't believe it 🙂
Who is trying to do that? I have seen others try to restate my statements in an attempt to strawman me incorrect. (Yes, I used strawman as a verb).😀
I said at exactly 2x, the limit, the sampled stream of a sine cannot be reconstructed as a single amplitude and phase sine.
As sampling increases above 2x, the required sample size for a specific accuracy falls.
Jn
Benb: a 2khz signal riding a cone running 20 hz, the time spent at the extremes in excursion...the distance to ear exceeds ITD thresholds.
Jitter as a location variation, not Doppler as a velocity variation.
Last edited:
On the other hand if one knows there is only a single fixed sine and wants to determine what it is, that seems like the simplest case since the equation for a sine is defined in terms of three variables (A, omega, theta) and with perfect data three samples could suffice.
That's where the fun (and confusion) starts 😀. Two samples per period (slightly less than a period apart) are enough for the infinite sine, since the concept of phase doesn't exist (what would be the reference for phase?).
What I mean is, my CD player does not take 10 ms worth of the data stream and calculate what the output should be.
Are you one of those NOS guys? Even 2x oversampling with an analog anti-imaging filter effectively uses a lot of samples in each output instant.
A misunderstanding between sampling at zero crossings and all sample at zero crossings.Think it over, please...
I said at exactly 2x, the limit, the sampled stream of a sine cannot be reconstructed as a single amplitude and phase sine.
Don Quichotte, meet the windmill

Per #33062 that would "kinda" violate Nyquist. Like in half pregnant, mind you.
Last edited:
All samples at zero crossing are possible only if Fsignal = Fs/2. If Fsignal is properly limited to < Fs/2, your scenario of all samples exactly at zero crossing is impossible.A misunderstanding between sampling at zero crossings and all sample at zero crossings.
Back to DAC’s again I see.
I am sure this is what hell must be like. Caught in an interminable time warp recycling the same questions and the same answers.
I used to feel the same way during our monthly, quarterly and annual business review meetings when I worked for a large MNC.
A bunch of SVP’s grilling us unsuspecting grunts about minutia that was of no consequence and then announcing with a flourish that clearly we weren’t up to the job, before flying off to repeat the whole charade with the next business in the corporation.
This is the beauty about being an atheist. You know you will face none of this.
😀
I am sure this is what hell must be like. Caught in an interminable time warp recycling the same questions and the same answers.
I used to feel the same way during our monthly, quarterly and annual business review meetings when I worked for a large MNC.
A bunch of SVP’s grilling us unsuspecting grunts about minutia that was of no consequence and then announcing with a flourish that clearly we weren’t up to the job, before flying off to repeat the whole charade with the next business in the corporation.
This is the beauty about being an atheist. You know you will face none of this.
😀
Last edited:
A bunch of SVP’s grilling us unsuspecting grunts about minutia that was of no consequence and then announcing with a flourish that clearly we weren’t up to the job, before flying off to repeat the whole charade with the next business in the corporation.
You clearly had an expectation management issue. Were you expecting praises from a gang of SVP fat cats? Being always unhappy with the business results is on their job description, I have never seen one otherwise *)
Rest assured they had the same feeling when the CEO did their reviews.
*) Well, in the last 25 years. Good ol' times when we were all in the same boat are long over.
Last edited:
Not my scenario, at Fs integer >2 multiply of f signal, it is possible to have zero crossing samples interspersed in the samples.... your scenario of all samples exactly at zero crossing is impossible.
Last edited:
Are you one of those NOS guys? Even 2x oversampling with an analog anti-imaging filter effectively uses a lot of samples in each output instant.
I do not know what an NOS guy is.
Jn
Benb: a 2khz signal riding a cone running 20 hz, the time spent at the extremes in excursion...the distance to ear exceeds ITD thresholds.
Jitter as a location variation, not Doppler as a velocity variation.
That's one heck of a driver to cover that range. Now humour me. speed of sound is approx 340m/s so 5us is 1.7cm. That's a lorra xmax and I wouldn't want to use those (although the purifi drivers are getting to that range)
That's one heck of a driver to cover that range. Now humour me. speed of sound is approx 340m/s so 5us is 1.7cm. That's a lorra xmax and I wouldn't want to use those (although the purifi drivers are getting to that range)
I just googled speed of sound in inches per microsecond.
Came up .0135 inches per microsecond.
5 uSec is about .065 inches.
Jn
That's where the fun (and confusion) starts 😀. Two samples per period (slightly less than a period apart) are enough for the infinite sine, since the concept of phase doesn't exist (what would be the reference for phase?).
Right. It could possibly matter relative to something in a physical problem being investigated. Or not.
How do you mean Mark, you/we know how signals sum but the DAC doesn't and it doesn't need to?
- Status
- Not open for further replies.
- Home
- Member Areas
- The Lounge
- John Curl's Blowtorch preamplifier part III