Talk:Linear phase

Linear phase $$\nRightarrow$$ symmetrical impulse response
Contrary to what the article current implies, a symmetrical impulse response is not a necessary condition for a discrete-time FIR to be phase-linear (it is merely a sufficient condition). It isn't hard to come up with an example which is asymmetrical, yet linear-phase nonetheless.

However, in the case of continuous-time FIRs (which don't exist in the real world), symmetry is a necessary condition for linear-phase.

Oli Filth 09:24, 13 December 2006 (UTC) Comment by another user

Is it traditional to exhibit the diagrams with no labels on the axes? If so, there should be a link to an article about the diagrams. No labels makes it hard for the non-expert to interpret. . . —Preceding unsigned comment added by 65.217.188.20 (talk) 16:51, 16 June 2008 (UTC)

It would be helpful to label the axes on the plots here? Are you showing the change in phase as a function of frequency for different filters? A proper description of what was done (code?) to generate the figures here would be helpful. —Preceding unsigned comment added by 147.143.81.22 (talk) 15:23, 24 June 2008 (UTC)

Not just filters but any linear time invariant device
Doesn't this article apply to any linear device? There is already wiki page "Linear time invariant". So word it like "a linear time invariant (LTI)" device such as a filter. Ohgddfp (talk) 16:14, 24 May 2022 (UTC)
 * Filter itself is a general enough term to include virtually any input/output device, even non-linear and not time-invariant. So, using the word filter does not exclude anything.  However, it is common in the sources to talk about linear-phase in the context of filters, so, I think the article should stay as it is.  Constant314 (talk) 16:44, 24 May 2022 (UTC)


 * I also think it is fine as is.--Bob K (talk) 19:14, 26 May 2022 (UTC)
 * About "Linear phase is a property of a filter where the phase response of the filter is a linear function of frequency.", at the beginning of the article. Now a phase response can be mathematically correctly derived from a transfer function, which shows solutions to linear ordinary differential equations.  A "linear system" in that context means the system does not create frequencies at the output that are not present at the input. For that reason, talking about phase does not make sense if the system is a non-linear system.  So if "filter" includes non-linear systems, then that kind of "filter" shouldn't be talked about in the article.  Now the word "linear" in "Linear Phase" has a completely different meaning than its use in "Linear System", even though talking about "Linear phase" makes sense only in a "linear system".
 * Now I see inside that sentence there is a link to "linear function". Now, what if the phase vs frequency graph is straight over the frequency range of interest, but the straight line portion is extended to zero frequency to give a linear function like that in the example graph in the wiki "Linear function" article? Looking only at positive values on the example graph x axis, would that be "linear phase"? Ohgddfp (talk) 06:43, 3 June 2022 (UTC)

But to get to the central issue of "Linear phase", I think the problem is that "linear function" is not sufficient to get "linear phase". What's required is "linear function with no phase constant". "linear function" alone yields only a "generalized linear phase", as explained later in this article. Ohgddfp (talk) 07:40, 3 June 2022 (UTC)