4
$\begingroup$

Section 13.2 of the HD Radio standard for AM defines a pulse shaping function that appears to be some strange variant on a raised cosine filter:

enter image description here

enter image description here

More specifically, H(ξ) is a raised cosine filter with total width (1+α)T, and W(ξ) appears to be a "stretched" version. Why do they do this?

$\endgroup$
3
$\begingroup$

The pulse envelope is convolved with a Gaussian function to soften the transition between the symbols. If the standard pulse envelope were used without any softening, it would be audible to an AM detector (it runs at ~172 Hz).

Source: Chapter 8 of The IBOC Handbook: Understanding HD Radio (TM) Technology by David P. Maxson

$\endgroup$
2
  • $\begingroup$ Any chance of a page reference? :-) $\endgroup$
    – Peter K.
    Sep 18 '13 at 21:24
  • 1
    $\begingroup$ Chapter 8 is the best I can do. I was using the Google Books preview. :-) $\endgroup$ Sep 19 '13 at 4:53
1
$\begingroup$

It may be to constrain a symbol within the period to mitigate ISI. There may be other properties that become apparent in practical application over more traditional tapers. It could be the result of research combined with analysis of real world data.

Given the paper doesn't cite any source, you may have to do some more research on pulse-shaping and OFDM in scholarly publications. Maybe there's a paper out there that has more information or a comparison of windowing functions.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.