Connexion
Ce forum permet à des personnes du monde entier de communiquer, c′est pourquoi les messages échangés sont en anglais.

High-resonance filter vs. volume envelope

  • DE 9 0
    Message de Dan Eble le
    I've come across a difference between the synthesizer in Polyphone (1.7 on OS X) and Fluidsynth (1.1.6, installed with homebrew). At first, I thought it was a bug in Fluidsynth, but the gist of the response I received on the Fluidsynth developer list was that the problem is mine. I am a newbie to audio processing, so it's hard for me to judge. If there is no problem in Fluidsynth, then is there a problem in Polyphone? (But Polyphone's behavior makes more sense as I read the SoundFont spec.) Is the spec vague enough that both implementations can be considered compliant? That would be a shame, since their output is so different.

    I posted tiny SF2 and MIDI files in the thread on the Fluidsynth developer list.
    • first post: http://lists.nongnu.org/archive/html/fluid-dev/2016-09/msg00000.html
    • post with SF2 and MIDI: http://lists.nongnu.org/archive/html/fluid-dev/2016-09/msg00005.html
    Thanks.
  • 410 0
    Message de Davy le
    From what I learn from the fluid-dev discussion, you are in the case where you quickly stop a highly amplified sample because of an exclusive class applied on a sample filtered with a high Q. Even if Polyphone and Fluidsynth are both compliant with the sf2 specifications regarding these points, the click that can be heard in Fluidsynth may be due to an optimization concern.

    For example, if Fluidsynth considers that a sound continues until its envelop reaches 0.0001, Polyphone may go further and wait that the envelop reaches 0.000001 (values are arbitrary) to ensure that stopping the sound is not audible anymore. Your case is maybe too particular to ask them to modify this behavior. Fluidsynth is used since long time, is very nicely optimized for 99% of the use cases.

    I hope I understood the problem you are facing
  • DE 9 0
    Message de Dan Eble le
    The click might be of secondary importance. These differences seem more fundamental:

    Case 1: Play and release a single note
    Polyphone: after release, the sound diminishes to silence by the end of the release phase
    Fluidsynth: after release, the sound continues without diminishing and stops abruptly at the end of the release phase
    The difference does not seem to be a matter of time spent, but of the rate of change in loudness during the release phase.

    Case 2: Play and release a single note, then play another in the same exclusive class
    Polyphone: after the second note begins, the sound of the first stops immediately
    Fluidsynth: after the second note begins, the first continues to sound for about 0.75 s
    The difference seems to be a significantly different interpretation of "rapidly" in the SoundFont spec.

    I hope that is more clear.

    And I don't want to sound like I'm complaining, but I'm sure you can imagine the feeling of working for a while in Polyphone, believing I've created something very nice, and then learning that it works badly when I go to render my music in Fluidsynth. But I have found Polyphone useful; thanks for creating it.

Connectez-vous ou inscrivez-vous pour participer à la discussion.

Polyphone a besoin de vous !

Polyphone est gratuit mais il y a des coûts associés à son site web et à son développement. Un petit coup de pouce aidera beaucoup.

Faire un don
Apprenez les bases Voir le tutoriel
Haut de
page