Latency

This topic contains 6 replies, has 6 voices, and was last updated by Profile photo of Pit Lenz Pit Lenz 8 years, 9 months ago.

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #23462
    Profile photo of scottwc
    scottwc
    Participant

    Is there any latency added when routing a channel through a group and then to the mains?

    My very imprecise test (feed signal to a group and LR and reverse the polarity on one) didn’t tell me much. Probably because of it being imprecise [:)]

    If so, how much?

    #32340
    Profile photo of Detonator
    Detonator
    Participant

    Scottwc –

    There will be latency, not sure how much, but you can perform a simple test to see the effects of the latency by the “combing” it produces when combined with its not-latent signal.

    Simply play some source material (music) through a channel directly to LR, then route that signal through a group routed to LR. Play the direct signal, then add in the group, (use the group “mute” to alternately add & remove the group) note the difference, particulary in the higher tones. The muddying of the signal comes from the combing produced by combining the out-of-phase signals. This is generally less noticeable in the lower frequencies.

    Perhaps A-H can give you exact figures of latency.

    Tim Tyler
    Detonator Sound
    Richmond, Va USA
    T112/48 R72/32 laptop iPad Dir-815 Dante

    #32341
    Profile photo of Chris93
    Chris93
    Participant

    Send a click track into an input channel and assign that input to both LR and to a mono group. Assign the group to LR. Pan the input channel hard left and pan the group hard right. Record L and R to a flash drive and take a look at it in a DAW. Audacity would do the job.

    Chris

    #32342
    Profile photo of jcarter
    jcarter
    Participant

    When we were demoing a GLD I asked the (Canadian) distributor’s product rep about this.

    His response, as I recall, is that the extra latency in the longer signal path is automatically compensated with a delay added to the shorter signal path, in order to keep the two signals in phase when they hit the LR bus.

    So as long as all signals stay inside the GLD I wouldn’t expect to hear any comb-filtering between different signal paths (if there’s any external hardware inserted in one of the signal chains all bets are off, of course).

    Not sure if the *total* latency of the console is constant (based on the longest possible internal routing path) or if it depends on the routing.

    #32366
    Profile photo of isaac
    isaac
    Participant

    Or use the click example. Route it to two inputs, Route one direct to L/R, and route the other through a group. pan both to centre. Phase invert one signal. If you hear nothing then the GLD is compensating for the the group outing.

    I would also try it with music.

    Isaac P

    #32408
    Profile photo of Pit Lenz
    Pit Lenz
    Participant

    I did some latency tests and found out a strange behavior of the groups:
    I found out that monoGroups 1/2 and Grps 17/18 (in a 18-monoGrp-setup) and stereoGrps 1 and 9 (in a 9-stereoGrp-Setup) seem to have a delay of 4 samples.

    Here`s the test setup:
    GLD V1.02, Rev. 5638
    I started with a default FOH show and put up 9 stereoGroups in one show, 19 monoGrps in another show.
    Both shows can be downloaded to check for yourselves at http://www.pitlenz.de/public/AllenHeathGLD.zip

    Now I hooked up a white noise generator (from AudioTools) on channel 45/46, because small sample delay differences can be easily detected with white noise. If there is a delay, you`ll clearly hear a coloration in the sound.

    Now let`s take scene #1 of the “9StereoGrps” show. (allGrps straight, no delay).
    WhiteNoise at Input45/46 is sent to all Grps, all Grps are routed to L/R. Everything`s straight, no EQ, nothing delayed. We listen to L/R in the headphones and active speakers.

    Let`s turn up Grp1: WhiteNoise is clear. Grp1 stays up, we add Grp2 : Noise gets fuzzy. Now check all the Grp combinations. Any Grps from 2 to 8 are OK with each other, the sound is just alike to one single open Grp, just louder. Good! But if we combine any of Grp 2-8 with either Grp1 or Grp9, the noise gets fuzzy!

    Now let`s turn up Grps 1 and 2, select Grp2 and go to Processing/Delay. Start turning up the delay time of Grp2 and at 4 samples the noise is clear just as it is between G2-8. Let`s set a delay of 4 samples ro Grps 2-8.
    Now ALL of the groups are time aligned, but 4samples late (This is stored in scene #2).
    That`s OK, as long as we mix ONLY through groups and nothing is routed directly to L/R. (This option is shown in scenes #3 and #4).

    In scene #3 ( Direct to L/R AND to un-delayed Grps) we can combine Grps 2-8 with the direct signal without coloration, bus as we put up Grp1 or 9 WITH the direct signal, the white noise gets fuzzy again.

    In the second show (18monoGrps), the four Grps 1/2, and 17/18 behave just as their stereo combined versions in the StereoGrp-show. Any of these can be combined with each other but not with Grps 3-16, as thery are late by 4samples again. Delay them (scenes #2) and they are aligned with each other, but not with the direct signal (see scene #4).

    Now comes the FUNNY PART:
    PFL signals behave differently, but ONLY on monoGrps 1/2 oe stereoGrp1, NOT on StereoGrp9 or MonoGrps 17/18!
    If nothing`s delayed (scene #1) and the combination of stereoGrps 1 and 2 sound fuzzy, their PFL signal combined sounds OK.
    If we delay Grp2 to timealign to Grp1, our mix sound OK but PFLing both sounds fuzzy now!
    This is not the case with stereoGrp9 (or monoGrps17/18 in the other show). Here PFL gives the same result as the mixed output.

    This is quite confusing, maybe someone from A&H wants to chime ine to shed a little light on this….
    If you want to listen to this for yourself, check out the files, use a WhiteNoise generator hooked to Input 15/16 and report your results.

    Greeitngs from Germany
    Pit

    #33273
    Profile photo of Pit Lenz
    Pit Lenz
    Participant

    Thank you, guys at A&H!
    Without making too much noise about it, in firmware V1.1 you have solved the group latency issue described above!

    that`s great news, now we can freely combine and (more important) add L/R and Groups to an output matrix without running into phasing problems.
    I use it for example to mix to L/R, send that to a MTX for PA and to another MTX for nearfill. Those nearfills get some additional vocals from a vocal GRP to compensate the band volume coming from the stage.
    This was not possible without phase cancellation issues in V1.02.
    Now it works!

    Thanks again…
    Greetings from Germany
    Pit
    [:)]

Viewing 7 posts - 1 through 7 (of 7 total)

The forum ‘Archived GLD Discussions’ is closed to new topics and replies.