Basic definitions for causal signal processors which are controlled by another signal. Additionally to Synthesizer.Dimensional.ControlledProcess you can convert those processes to plain causal processes in the case of equal audio and control rates (synchronous control).

It is sensible to bundle the functions computation of internal parameters and running the main process, since computation of the internal parameters depends on the sample rate of the main process in case of frequency control values even though the computation of internal parameters happens at a different sample rate.

ToDo: - Is it better to provide the conversion method not by a record but by a type class? The difficulty with this is, how to handle global parameters like the filter order? - Note, that parameters might be computed by different ways. Thus a type class with functional dependencies for automatic selection of input types and conversion will not always be flexible enough. - Is it possible and reasonable to hide the type parameter for the internal control parameter since the user does not need to know it? - The internal parameters that the converter generate usually depend on the sample rate of the (target) audio signal. However, it does not depend on the sample rate of control signal where it is applied to. How can we ensure that it is not used somewhere else? We could discourage access to it at all. But it might be sensible to define new external parameters in terms of existing ones. We could add a phantom s type parameter to internal control parameters. Would this do the trick? Is this convenient?

This buffers internal control parameters before interpolation. This should be faster, since interpolation needs frequent look-ahead, and this is faster on a buffered signal than on a plain stateful signal generator.

Since the look-ahead is constant, it is interesting whether interpolation can be made more efficient without the inefficient intermediate list structure.