DCO's, Anti-Aliasing, and Filters

[inman at interpath.com]

Roman Sowa wrote:
> > Aliasing comes from reading sample faster than it was recorded
> > (i.e. for playing it at higher pitch)

Larry Hendry wrote: 
> Doesn't that come from playing the sample slower for a lower pitch?
> Larry Hendry

I ask:

In Hal Chamberlin's "Musical Applications of Microprocessors,"
he talks about alias distortion as a result of direct synthesis 
of waveforms (p. 423) before discussing aliasing as a result of
table lookup methods in which samples are read at different 
speeds.  Is it possible that the kind of direct synthesis 
"aliasing" that Chamberlin was talking about is no longer a 
common usage of the word "aliasing" because his book was last 
updated in 1985 and microprocessors are so much faster now and 
new DAC's for converting signals are so much better that such 
aliasing is generally not a problem?  Except perhaps to us
re-inventers trying to do it with a 2K PIC...
  
The idea that aliasing could be produced by direct computation
of waveforms (and not just a DAT machine reading in and
replaying signals that we didn't know were there or want) was
new to me, until a couple of recent posts on digital waveforms.
I had also never considered that one could produce aliasing 
just by outputting a digital square wave with positive and 
negative output intervals that are not consistent across musical 
notes -- until some recent posts on DCO's drew that to my attention.  
If you cannot time the positive and negative pulses to parallel 
perfection, you will create additional (possibly random or
unpleasant) harmonics, which I believe the writer was calling
"aliasing."  Is it is or is it ain't?

Any corrections to these thoughts are more than welcome.

Elliot "choosing embarassment over ignorance any day of the week"