[CLAM] Re. New SMSBatch

xamat at iua.upf.es xamat at iua.upf.es
Wed Jul 28 14:46:45 PDT 2004


> 1) About interpolation -  in the example transformation score, I only
> see "Step" and "Linear" interpolation for the various time-varying
> functions.  But I also found "Round," "Spline," and a number of
> "Polynomial" interpolations in a src/Standard/GlobalEnums.cxx ...would
> those be available to use in SMS functions (particularly for the command
> line programs)?  If so, what would they do?

They are indeed available. These interpolation types specify how the
interpolation between two points in the break-point function is performed.
I.e. if set to linear, a line is drawn from one point to the other and
intermediate values are obtained from this line; in step interpolation the
value of the left point is used in any intermediate point; in "round" the
average value from the two points is used; in "spline" a spline
approximation function is drawn and in the polynomial an nth order
polynomial is used. Note that both these last interpolations need a number
of points, in the case of the polynomial related to its order, in order to
give coherent results... We have never used them in this context and they
are interesting for approximating spectral shapes, tell us what happens if
you try.

> Second question.  For SMSMorph, in the sample xml transformation score,
> there are listed only two or three things: HybBFP, etc.  I assume the
> other ones would follow the same pattern

Yes, they do indeed follow the same pattern. In the SMSTools graphical
interface you will find a brief explanation of their meaning. I must warn
you that some of these parameters have not been thoroughly tested and your
feedback will be most valuable.

> What controls the "Interpolate intermediate frames" switch I find in
> SMSTools, if I want to implement that in the text xml transformation
> score?

In SMSMorphConfig
you will find the InterpolateFrame boolean. Just add to the score:

<InterpolateFrame> 0 </InterpolateFrame> (for deactivating or 1 for
activating frame interpolation)

> Thanks,
> Matt
> _______________________________________________
> CLAM mailing list
> CLAM at iua.upf.es
> http://www.iua.upf.es/mtg/clam

   Xavier Amatriain
 Music Technology Group
Universitat Pompeu Fabra

More information about the clam-users mailing list