Click with AET

Discussion in 'KONTAKT' started by stephen23, Jun 14, 2019 at 10:48 PM.

  1. stephen23

    stephen23 NI Product Owner

    Messages:
    379
    I have just started noticing a soft click sometimes at note transitions, in an instrument which uses AET. It's a French horn which uses an "articulation" morph with just 2 layers, and otherwise works extremely well. I've been unable to identify the circumstances necessary to produce the click - it's reproduceable if I send exactly the same sequence of messages, but on other occasions I send a sequence which is very slightly different, and there is no click, again reproduceably. Morphing is controlled by the same cc as volume and is constantly changing (I use a breath controller). There is no click when I bypass the AET.

    It's not related to the sound from the samples - giving a long attack time to the second note of the transition makes no difference to the click. It happens on or offline. I'm on Kontakt 5.

    I wonder if anyone has had the same problem and perhaps even found a workaround? I suspect it's a bug in the AET software.

    It's worth mentioning incidentally that this marvellous AET filter is in a dire state: the programmers couldn't have made it more user-unfriendly if they tried. It's got silly bugs - for example some of the fields for data entry have to be clicked twice before they will work, and the only way to delete a morph layer is to pretend to use it to make a new map. I get the impression it's been abandoned - is this because people don't use it much? What a shame.
     
  2. EvilDragon

    EvilDragon Moderator Moderator

    Messages:
    14,564
    What's the sample rate of the samples vs sample rate of the project you're using?
     
  3. stephen23

    stephen23 NI Product Owner

    Messages:
    379
    Good thinking. The horn samples are from EWQLSO, all at 44.1, my host is Bidule, also running at 44.1. There are Spitfire samples (48kHz) in other instruments in the same multi but they don't have to be playing.
     
  4. EvilDragon

    EvilDragon Moderator Moderator

    Messages:
    14,564
    AET should be fine at 44.1k. But sometimes it really doesn't work well with certain material.

    Yeah, it's pretty much neglected unfortunately. But also it's not exactly CPU efficient either... It was a great idea, but with a bit tricky implementation and not enough explanation, and ultimately extremely low adoption rate amongst 3rd party developers.
     
  5. stephen23

    stephen23 NI Product Owner

    Messages:
    379
    Does require a bit of effort, certainly. But if you know what you're doing, and use the right samples, it certainly repays that effort. It's never given me any CPU issues. Maybe more people would use it if it were better presented and documented. There have been a few queries on this forum about how to fade from one sample to another without a velocity change. This is how you do it.