Regarding Robbie's comments about the limitations of MIDI sent through
a wire and question as to whether these were noticeable in the
performance of the PowerRoll:

When we contemplated the design of the PowerRoll I was concerned about
the possibility of corruption of data if MIDI were used, because Wayne
Stahnke had told me for many years that he disliked the MIDI format due
to frequent inaccuracies and lack of standardization.

Laurent Coray conferred with Wayne about the project, and afterward I
asked Wayne whether we needed to be concerned about the use of MIDI.
Since both Laurent and I were perfectionists, any degradation in the
performance would have been completely unacceptable.  Had we believed
that to achieve accuracy we had to use a format other than MIDI, we
would have done so without hesitation.  This project was to allow no
compromise of any kind.

Wayne said that if the job were done right, no corruption of data would
occur, and that he assumed that Laurent, as a competent engineer, knew
how to handle MIDI so that the resulting reproductions would be
accurate.

We find that when playing QRS audio/MIDI discs on the PowerRoll, the
data sent to the PowerRoll has to be slowed down to match the audio
track, which has been delayed to match the slow response of the QRS
Pianomation device.  Laurent informed me that the PowerRoll plays in
real time and has no delay.

It is also interesting that there seems to be no difference in
performance whether using a 1 foot cable or 150 foot cable; the
PowerRoll plays perfectly with either.  I don't know at what length
it would not work, if any.

All of our most particular customers have expressed the belief that the
PowerRoll plays as well as or better than paper rolls, although we have
never attempted to quantify this.  Repetition and speed of response are
generally breathtakingly accurate.

There was one further grade of MIDI data optimization for which Laurent
had provided, which he and Richard Brandle had planned when time
permitted; unfortunately he died before this could be done.  Eventually
I hope that Laurent's successor, Ron Perry, will complete this work,
although I don't understand what it consisted of, and although no one
has ever reported any deficiency in performance to me which I am unable
to easily correct.  Richard can undoubtedly explain this optimization
matter in technical terms, for anyone with the background to understand
it.

Larry Broadmoore, Broadmoore Research Corporation
1709 First Street, Unit "C", San Fernando, CA 91340
tel: (800) 497-4266 ; (818) 365-6231 ; fax: (818) 361-0911