Jconvolver is a Convolution Engine for JACK, based on FFT convolution and using
non-uniform partition sizes: small ones at the start of the IR and building up
to the most efficient size further on. It can perform zero-delay processing with
moderate CPU load. Jconvolver uses the convolution engine designed for Aella, a
convolution application for reverberation processing (to be announced later).
This distributes the calculation over up to five threads, one for each partition
size, running at priorities just below the the one of JACK's processing thread.
This engine is a separate library that will be documented as soon as I can find
the time.

Main features:
* Any matrix of convolutions between up to up 64 inputs and 64 outputs, as long
  as your CPU(s) can handle it.
* Allows trading off CPU load to processing delay, and remains efficient even
  when configured for zero delay.
* Sparse and diagonal matrices are handled as efficiently as dense ones. No CPU
  cycles or memory resources are wasted on empty cells in the matrix, nor on
  empty partitions if IRs are of different length.
`