? diff morph/fade

[jh]

>A good simple example of the differences can be seen
>when you consider morphing between a ramp and reverse 
>ramp. The exact middle position would be a triangle
>wave

This is *one* way to look at it.
I think there is no exact definition of "Morphing". 
Sure, morphing means a smooth transition from one appearance
to another one, but there is all freedom of art *how* this
transition is going to happen. If you morph the picture of a
human face into a lion's face, you can say "it was a good morph",
or "a bad morph", but you can't say "this is the only right way to
do the morph". 
Though it's the *shape* or appearance that matters in the end,
changing parameters is an important part of it. Like moving
poles in a filter continuously. But there are many "right" ways
to do this. Your ramp example has at least two valid solutions
in the synth history, one defined by EMS, and one defined
by Korg. (And others, most probably.) Interesting enough,
both have a triangle in the middle indeed, but everything 
else than the 0%, 50% and 100% reults are different.
And both versions are highly useful for modulation, btw.

>One could argue that audio morphing is all about finding
>new and interesting ways to get from A to B.

This would be my argument - yes.

> Crossfading
>being the one we all know well enough to find uninteresting.
>Or it may be uninteresting because no new components are
>created in the middle regions.

Think about this example: Multimode filter of the Oberheim SEM.
Apart from the click stop BP position, there is a pot that creates
a variable mix of LP and HP output of the same filter. This is
Crossfading between HP and LP. Is it also a Morphing between
HP and LP ? I would say no, because halfway in between there 
is the Notch filter response.
Now forget about the implemetation of the SEM filter for a moment,
and consider you should create a 3-stage Morphing from
LP to Notch Filter to HP. You *can* find a set of parameters
(e. g. poles and zeros, or unity delays if you're designing it
in a sampled system, and find a clever algorithm to change
this set of parameters to get one nice transition between these
filter modes. If you're doing it bad (e.g. amplitude increasing 
by 25dB at 25% of your morphing), people won't like it, and
you will find it hard to argue that it is actually "Morphing". If 
you do a good job, it's just fine, and people will say it's a good 
Morphing.
Now if another one comes and says "I have a SEM filter, and it
does Morphing from LP to Notch to HP" - would his solution
be "wrong" ??

JH.