| I want a VAX (or similar data processing system) as part of my music
synthesizer not to do FM synthesis but to do data processing work on the
music. For example, storing and retrieving songs in named files,
backing up said files and exchanging them with other people. Obviously,
a new song can take parts, such as instrument definitions and even
phrases, from old songs. I would also like to do analysis of songs,
such as determining their running time and the range of each part.
Since my VAX can communicate with others, if I don't have a printer
myself I can print my music as a score on someone else's printer. A VAX
with a MIDI interface can do all of these things for any synthesizer
that supports MIDI, given proper software.
I don't want to ask about your product, since you gave us an address for
that, and it is unannounced, but could you say a few more words about
your dislike of MIDI. My biggest problem with it is its speed--you can
only specify new notes at the rate of 1000 per second, which is likely
not fast enough for complex passages involving lots of instruments. Your
machine is clearly 'high-end', like the Fairlight. Assuming you and the
other 'high-end' manufacturers got together with the 'high-end'
customers, such as Industrial Light and Magic, to produce a better
interface, how would it differ from MIDI? Would it be a better
interface for 'low-end' products, such as the Yamaha DX7?
John Sauter
|
| My complaint with MIDI is general - architecture. Given the organization
of the system they have done a good job, but I would much prefer a better
organization. I feel that a true parallel data arrangement that could
handle the throughput of any sample driven system would be superior.
Such tecnhology is available now. I prefer to use the IEEE-488 standard
interface, mostly because of the already existing add-ons to such
systems like Spectrum analyzers, etc. CompSync intends to introuduce
an IEEE-488 linkable control voltage generator/sequencer, and also a
'dynamic patcher', which is a group of analog tri-states that have
programmable combinations of in-outs for use in modular systems, such
as Roland's. In general you will find that companies, such as
CompSync that participate in the high-end market are not all that fond
of MIDI. It is a very LOW end solution to a very complex problem.
You might note that none of the MIDI desingers were ever network
architects, and what after all is such a system ???
Brian Markey
|
| I was just re-reading the early notes in the file when I stumbled
across ``this''. Read the base note (20.0) and tell me what you
think.
For all the frequency modulation and phase angles and wave tables
and harmonics and light pens and sampling and envelopes and hard
disk and slave keyboards, I haven't the foggiest where the sound
comes out. Is he implying that there is some form of FM which is
used as an index into a wave table which after a multitude of
mutilations comes out through a DAC?
Sounds like the sort of useless complexity you get when you stack
6 or 8 operators in a single FM stack.
Anybody heard anything from this company since? (Gee boss, I ain't
never heard of a CompSync... Sting uses one all over his latest
album.)
Steph
|
| The basis of what I get out of .0 is that FM is used to derive waveforms,
which are then converted into wavetables (basically sampling the FM waveform
within the digital domain). A large number of these wavetables can be stored
on hard disk, and swapped in on demand in pseudo real time (probably a function
of key velocity, modulation, or just plain multitimbrality via different
MIDI [or IEEE-488] channels) giving the illusion of n voice multitimbrality,
where n is very large. Kind of like on-demand virtual voice paging. I
don't understand how a hard disk is gonna do a seek and read and data
transfer (16x256=4k bits) fast enough to not be noticeable though, which
may explain why we havn't seen the beast.
Todd.
|
| Well, upon rereading, I see that I'm off the mark. The wavetable waveforms
are sampled or entered via light pen, and fed into the FM engine to modulate
each other. The demand wavetable still stands.
Sounds like the volume envelope is stolen from one of the wavetables and
fed to the FM unit so that the FM waveform envelope emulates the one in
the wavetable.
Todd.
|