9
The collection loaded on the SD Card included with the EPSi has impulses from
many different high quality sources (acoustic, electronic, electromechanical, and
synthesized).
The maximum Impulse response duration is about 6 seconds (262144 samples
at 44.1 KHz). Only the first 6 seconds of data get loaded for longer duration files.
This can result in a very abrupt sounding cutoff if the wave does not include
some sort of fade out within that time span.
The processing latency of the wet path is about 10.0 ms. As pre-delays in reverb
sounds are often longer than this, the latency is not noticeable on reverb sounds.
For zero latency processing, see firmware version C.
4.2.2.
Stereo image generator
In order to reach our 6 second time window the DSP can only process one
channel. The EPSi only uses the first channel of a multichannel wave file.
EPSi uses an original adaptive algorithm to synthesize a new stereo image. The
result of this process sounds as if the impulse was re-captured with an M/S
microphone pair and is convincing for most space simulation applications.
However any hard panning built in the original impulse is lost through this
process.
The stereo image generator is disabled when the right output jack is left
unplugged (mono output mode). In this situation the processed signal is an exact
rendering of the convolution.
4.2.3.
Adding some life to linearity
Basic convolution is a perfectly linear process. It can not simulate any sort of
distortion or time varying effect. This can result in the output sounding somewhat
stale. To remedy this, the stereo image generator also includes a little bit of
detuning. The detuning amount is very slight (1.5 cent). The result is some subtle
added sense of motion and chorusing, widening and warming up the sound with
no audible pitch shifting.
