Confused about group delay...

Discussion in 'Archived Threads 2001-2004' started by Dave Milne, Nov 14, 2002.

  1. Dave Milne

    Dave Milne Supporting Actor

    Jul 2, 2001
    Likes Received:
    There seems to be a consensus that the improved transient response/lower group delay of sealed subwoofers "sounds better", particularly for music. My own emperical data tends to support this.

    However, since most of these alignments need equalization (LT, BFD, etc) to achieve flat response down deep, doesn't that add group delay? And don't the high slope (4th order typical) crossovers add more group delay?

    I'm currently trying to integrate a 6th-order sub (vented box with an active 2nd-order filter/boost network) with the sealed woofers in my satellites. It's not going well. Part of the problem is that I didn't include a continuously variable 0-180 phase control into the active filter/crossover (does anyone have a schematic for such a circuit?). On the other hand, it this pointless since the phase response slopes are so different between the sat and sub? Intuition tells me that the best one could hope for is to match them up at only one frequency.

    I'm ready to throw out the sub and go for a "minimialist approach": sealed sub with 55Hz 2nd order active crossover to match the natural 2nd order rolloff of the satellites. I especially like the idea of eliminating electronics in the main signal chain. And the satellites can take the power - dual Dynaudio 24W100s with D76s and 330 Esotar in WMTMW config.

    Advice? A sealed Tumult would be nice, but it really needs an LT and my sub amp is only 800W.

    Incidentally, BassBox 6 models 6th-order vented-equalized subs, but doesn't seem to take the group delay of the filter network into account in the graphs...aargh!
  2. Janne Ahonen

    Janne Ahonen Extra

    May 4, 2002
    Likes Received:
    Kuopio, Finland
    Real Name:
    Janne Ahonen
    Hi Dave,

    LT is best equalizer for closed box when it comes to group delay. It has "interesting" phase property, it produces negative group delay, so that resulting transformed system has no more group delay than is absolutely necessary (same amount of group delay would have been resulted, if driver+box combination would have same transformed parameters Qp=Qtc and fp=fsc).

    Parametric equalizer and peaky second order highpass are very much inferior in this respect.

    Key here is how fast amplitude changes respect to frequency. Faster changes in amplitude means faster changes in phase (at least when it comes to minimum-phase systems) which in turn add higher group delay. It should be noted that group delay, even high, is not bad, if it is relatively constant. Varying group delay is bad thing.


Share This Page