Gaffaweb > Love & Anger > 1991-38 > [ Date Index | Thread Index ]
[Date Prev] [Date Next] [Thread Prev] [Thread Next]


Yet more about the Fairlight

From: katefans@chinet.chi.il.us (Chris n Vickie)
Date: Sat, 19 Oct 1991 02:41:00 -0700
Subject: Yet more about the Fairlight
To: love-hounds@wiretap.spies.com

Chris here,

Jeffrey C. Burka opines:
>> But what Chris pointed out is that the Fairlight is not a sampler in the
>> conventional sense.  From Chris' description, it does much the same thing
>> that Alchemy for the Mac, which you mentioned, does--analyzes the input
>> sample, turning it into a synthesis algorhythm, allowing it to be played
>> back along a keyboard without the pitch being dependant upon playback
>> speed.

and the Drukman replies:
> Chris, would you care to document this astounding claim?  As far as I am
> aware, only the Synclavier does this particular trick.  I just talked to a
> guy who used to work for Sequential and he confirms that the Fairlight was
> just a big fancy sampler with some nice anti-aliasing circuitry.  I'm
> certain that the I, II and IIx didn't do this.  I don't know about the III,
> but I don't see any reason why it would, unless it was an add-on software
> package.  Nothing I know of is fast enough to resynthesize in real time 
> anyway, unless it's a Cray or something.

   I didn't say it did it in real time, Mr. Person. The Fairlight captured
one sample in ram at 8-bit quality (allowing for a quick playback to check)
and then used a Fast Fourier Transform (which took no little time). FFT
breaks a complex waveform into its simple component waveforms. The freqs.
of the component waveforms, and their relationships was the only information
stored.

     [ASCII graphic example of FFT process on a sample waveform 
      ommitted due to dire threat to sanity....]

  The synth in the Fairlight then had the fairly simple task of re-creating
(somewhat inaccurately) the original waveform from this "blueprint".

  I'm writing from memory of a three-part series on the Fairlight in
Keyboard, so I might be a tiny bit shakey on some details, but I'm quite
sure of the basic operation. One piece of proof is the fact that the
Series II stored multiple sounds on 8" floppy disks, one of the most
inefficient data storage mediums known to human-kind. CP/M formatted
8-inchers were 128k, for Kate's sake!

   BTW, the "re-synthesis" option was introduced for the Synclaiver only
a few years ago. The late, lamented GDS system (see Wendy Carlos) also
re-synthesied sound. You young whipper-snappers have grown up in a world
of cheap ram and even cheaper storage. The stock Series II didn't even
_have_ a hard-disk.
   I remember when engineers had to really work hard at their jobs and
....blab...blab...drool....

   Besides, what does "some guy from Sequential" know, anyway?

                            Chris (the old fart) Williams of
                                Chris'n'(another old fart)Vickie of Chicago
                                    katefans@chinet.chi.il.us