Re: [tuning] Digest Number 938

Paul,

>Robert, I've been thinking that the .wav files would be needed in
uncompressed form by the end user. This would make them occupy a
prohibitively large amount of memory. So I guess the MIDI approach is the
only reasonable one.

Well, the hexany with .wav clips came to about 1.63 Mb uncompressed, for
26 audio clips. That's for 1 sec each, 16 bit, 22050 Hz, = format I chose to
reduce the size while keeping quality fairly reasonable.

Would be getting on for three times that for each model if you did them all,
or just a bit over that if you did just the tetrads and triads.

So based on my exp.with the hexany, I'd have thought doing just the
tetads and triads would prob. be okay. Even with them all, would be
less than 6 Mb, which isn't so much nowadays. I don't know whether
they would all be kept in memory or played from disk as needed.

>As for rotating the cylinder, I wrote a program in Matlab that finds the
simplest rotation that gets an object from one configuration to another, but
it uses Matlab's optimization toolbox . . . I guess I could just run this
for all the configurations in question . . . write me offline if you'd like
my help on this.

I think the 3)5 dekany is the problem here. For the 2)5, I think one would
soon get used to how the spheres work, with the colour coding, and the
feedback one gets from hearing the sounds, plus the very clear 3D
tetrahedra that make up this model.

I've tried a fair number of variations on the 3)5 dekany this evening.

Tried tetrahedra for the triads. That gave a really nice stationary model,
but as soon as you try to turn it, you get rather confused with all the
extra straight lines.

So I think it is better to keep to spheres for the tetrads for this model.

See what you think - here are the two models:

http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_faces_for_triads_0p25.wrl

and
http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_faces_for_triads_tetrahedra_for_tetrads_0p25.wrl

What I feel is that the tetrahedra distract one from the basic shape of
the dekany as one turns the model. While the spheres, not being angular
at all, let one focus more clearly on the dekany.

Notice I've left out a lot of the spheres in these models. That's because
now you can click on the faces themselves for the triads.

Also made the diads smaller.

I think it would work, given the extra feedback from hearing the sounds.

>it uses Matlab's optimization toolbox . . . I guess I could just run this
for all the configurations in question . . . write me offline if you'd like
my help on this.

Right, been doing essentially the same thing then!

Yes, one can always reduce any number of rotations in 3D to a single rotation,
or a translation.

If you move an object from one place to another in 3D, you can always do it
with a single translation, or a single rotation about one axis, whatever it's
start and end orientation. Kind of remarkable result, that one can do that!

It's okay thanks, I can do it in c / java if it needs to be done, and have almost
all the code there, prob. only need to change a few lines in my VRML viewpoints
applet / windows prog., in order to change the start vector from 0 0 -1 to 0 1 0.

The template files are
http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_faces_for_triads.txt
and
http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_faces_for_triads_tetrahedra_for_tetrads.txt

and you will need the latest version of

http://members.nbci.com/_XMCM/tune_smithy/dekany/srvrml.exe

or

http://members.nbci.com/_XMCM/tune_smithy/dekany/srvrml.c

to replace some of the equations.

Would indeed be nice if you were to do them now. Be sure to let me know if anything needs clarifying,
off-line if that seems best.

However if you wait a little while, I'll be doing the MIDI clips. Still have to finish FTS 1.09 first, and don't want to
get involved in adding new features to it right now. Hopefully by next week it will be released, based on how
it is going at present, but can never really say about such things...

I'll prob. call the code in svrml.c as a subroutine from FTS to do the S&R to make the vrml dekany.

Robert

>If you move an object from one place to another in 3D, you can always do it
>with a single translation, or a single rotation about one axis, whatever it's
>start and end orientation. Kind of remarkable result, that one can do that!

Sorry, you can always do it with a single translation plus a single rotation about
one axis.

In 2D, single rotation is enough.

In 3D, it isn't - imagine spinning a plate about it's axis, then lifting it up vertically.

Robert

--- In tuning@egroups.com, "Robert Walker" <robert_walker@r...> wrote:

http://www.egroups.com/message/tuning/15524

>
> See what you think - here are the two models:
>
>
http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_fa
ces_for_triads_0p25.wrl
>
> and
>
http://members.nbci.com/_XMCM/tune_smithy/dekany/dekany3_with_diads_fa
ces_for_triads_tetrahedra_for_tetrads_0p25.wrl
>
> What I feel is that the tetrahedra distract one from the basic
shape of the dekany as one turns the model. While the spheres, not
being angular at all, let one focus more clearly on the dekany.
>

Hello Robert!

Although I'm somewhat a novice at this, it seems as though the first
of these two figures is clearer and, actually, more interesting to
look at... Is there any way, though, to make the semi-opaque faces a
bit more opaque. One can see through them, but not too well. I
think a more opaque face would make the figures even more exciting...

Joseph