Jump to content

Talk:Guard interval

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

This paragraph should only appear on the de:Guard Interval pages. Also, should it be in first person perspective. "This is the case in Belgium because we have the very large GI (224 µsec) 25% time out for the transmission data but , in SFN in 16 or 64 QAM in SD mode the transmission around 100 km is very good (largely to compare whith the analog transmission in UHF.

The flow of data in 16 QAM increases the pixellisation a bit but this is acceptable." Emperor Dalek (talk) 15:05, 15 June 2009 (UTC)[reply]

Explanation of how guard interval actually works.

[edit]

Absolutely no explanation here that appear to show just how the guard interval is meant to combat multipath. No proper explanation (mathematics, equations, diagrams etc.) to how it deals with multipath signals. If a continuous wave signal arrives, and then another 'echo' of that signal arrives, then both signals sum together to give an overall signal. We all know this. If both signal are 'continuous' and both remain present at the receiver, then what students and people want to know is - what does the guard interval do to deal with the multi-path effect? If multipath signals are being received, then what's the theory that allows the OFDM signal to be decoded? For example, if two replica signals (each one containing the same 'symbol' are incoming, and are delayed by delta T (in time), then what's the all-important piece of theory that allows us to recover the symbol? Let's assume that both symbols are basically still over-lapping each other at the receiving side. How does the guard interval help to recover the message when FFT is applied to the OFDM signal? Show pictures. Show diagrams. Show equations. Show theory. For example - from math theory, a direct signal sinusoid added to a delayed version will still sum to a sinusoid at the same frequency. Another orthogonal sinusoid will also sum with their counterpart multipath delayed versions will also sum to be a sinusoid of the same frequency. This means - the resultant multipath-summed sinusoids are still going to be orthogonal sinusoids even in the presence of multipath. The guard interval is a pre-defined of information tacked on (as extra) to the front (beginning) of an OFDM symbol, which just gives enough time for transient activities due to the "front ends" of all multipath echos to arrive before taking samples for the FFT. When all multipath signals have arrived (and still present), the FFT will still be able to extract the symbol, because carriers remain sinusoidal and orthogonal - as long as all the multipath components have arrived and present for the time we carry out the FFT procedure. KorgBoy (talk) 13:04, 26 February 2018 (UTC)[reply]