Talk:Pulse (signal processing)

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Graphic[edit]

A graphic illustrating what a pulse is would be nice... — Preceding unsigned comment added by 84.208.80.123 (talk) 13:34, 2 September 2012 (UTC)[reply]

Done SpinningSpark 17:10, 19 September 2012 (UTC)[reply]

Sinc pulse[edit]

I have rolled back the following insertion to its previous version;

The sinc pulse is of some significance in signal-processing theory and was until recently not possible to produced by a real generator. However, at May 27:th 2013, A scientific report from researchers Brès and Thévenaz at the Swiss Federal Institute of Technology in Lausanne (EPFL) proved a way to acomplish Nyquist pulse signals and increase data throughput capacity by ten times, compared to earlier methods, yet compatible with common optic infrastructure.

The paper being referenced is here. First of all, it is well known that to create a complete sinc pulse causality must necessarily be violated because a true sinc pulse goes all the way back to minus infinity (as well as all the way forward to plus infinity). Are these researchers claiming to have violated causality? I don't think they are, but if that is what the claim is it would be an astounding discovery. Extraordinary claims require extraordinary evidence and I think this is so extraordinary a claim that we would require other researchers to verify it before reporting it as fact in Wikipedia.

However, I don't think the paper is making any such claim and the addition made to our article is in error. The paper is very careful to describe what they have generated as "sinc-shaped pulses" rather than actual sinc pulses. They generate these using a rectangular band-limited frequency comb. It is a mathematical result from Fourier analysis that a bounded function in the frequency domain is of necessity unbounded in the time domain (and vice versa). Certainly, they have shown that their generator when generating a steady stream of pulses meets the Nyquist criterion. But this stream of pulses is implied to go back all the way to minus infinity to get that result. In other words, such a pulse chain violates causality just as much as a single sinc pulse. In reality the pulse chain had to have a beginning, it cannot have existed before the generator was switched on. At the very most, it could not exist before the system was invented. For sure they have produced a pulse chain that is very close to ideal sinc pulses, but then it is theoretically possible to produce a single sinc pulse to any arbitrary accuracy provided an arbitrary delay can be introduced. SpinningSpark 21:47, 6 December 2013 (UTC)[reply]

Nyquist ISI criterion[edit]

I really don't think that it is helpful to litter such a top-level article with detailed mathematics. To the vast majority of general readers it will be meaningless, especially as none of the terms were explained. It's really not necessary when we can just link to Nyquist ISI criterion. The edit was especially unhelpful by removing the sentence that said Nyquist pulses are important for data transmission, the main piece of context a general reader could relate to. SpinningSpark 20:43, 28 September 2018 (UTC)[reply]

About "The edit was especially unhelpful by removing the sentence that said Nyquist pulses are important for data transmission, the main piece of context a general reader could relate to.": Notice that the sencence was not removed. It was shifted. Fvultier (talk) 10:12, 29 September 2018 (UTC)[reply]