[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [fluid-dev] [PATCH] Effect level clip

From: Josh Green
Subject: Re: [fluid-dev] [PATCH] Effect level clip
Date: Mon, 17 Sep 2007 17:46:34 -0700

On Mon, 2007-09-17 at 19:29 -0500, David Hilvert wrote:
> On Mon, 17 Sep 2007 15:51:09 -0700
> Josh Green <address@hidden> wrote:
> > I think the wet parameter is related to how much of the "wet" reverb
> > signal is being mixed into the output.  
> Yes.  Renaming it as something else would probably be silly; I was being (at
> least) temporarily stupid.
> > Most of the magic is occurring in the fluid_revmodel_processreplace() and
> > fluid_revmodel_processmix() functions.  I think its correct to use it as the
> > level of the reverb, although I'm not certain if it currently has the 
> > desired
> > effect or not.
> It's probably cleaner, and is arguably more correct, than the effect level
> patch.  It could be extended, by analogy, to the chorus effect module by 
> adding
> a fluid_clip() operation there, and probably adjusting the initial level.  
> Another point is whether the following element of the effect level patch 
> should
> be kept or removed (in fluid_chorus.c:fluid_chorus_processreplace):
>     /* average values based on number of chorus stages */
>     if (chorus->number_blocks)
>       d_out /= chorus->number_blocks;

The idea with that patch was to make the chorus volume more uniform over
changes in the number of chorus stages, correct?  If that factor could
be pre-calculated into the other parameters when the chorus count
changes, then that would have the same effect without introducing an
extra test and division in the synthesis loop.  Looking at the current
code though, I'm also seeing that the chorus->level parameter is never
actually used (seems this parameter is broken too?).  I would think that
could be factored with the number_blocks parameter to come up with the
final scale factor for the chorus signal.  I'll look into that more.
Seems like there is a lot of stuff that was broken at some point or not
fully implemented.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]