melband misbehaving when changing the numbands

[tremblap]

in Max and PureData

expected behaviour: like [fluid.mfcc~ 40] would yield a list of size [numcoeffs] when it is modulated and up to 40

current behaviour: [fluid.melband~ 40] needs a dsp restart in Max to change its list output when modulating [numbands]

see attached patch:

----------begin_max5_patcher----------
554.3ocyVF0aaBCDG+Y3Sgk0djEgMPftuJSUUFvIycfAYftrU09YelyPWWFE
YTbh1Kf7ww82+746re12Cm2bh2gQeA8Ujm2y9ddfoQCdSi8v0rSEUrNvMrbn
NmqvAlOoGIjU7d3ajIihRvyl7G+LIY1yVlhUy64pG3RVdEezkv+Dklg94vPm
rZL0+yVtY9gwAHbNSdDit+sf1W7Mg73CJdQuwKZb1tv.DM5twWIgv.5tPz8i
+xK99iOBrEV9OzTLiPO+DHB9WUnJtbw0.5BqA6wamy0YLxvXHvXT5kw31Rnw
25DJIlZRnI+2jPotNgRhLLFl3fDZMuqicj+OPpmrEM7CG5PehXMpjkQk7gnt
JlY.YDRz3qrvqEl58TkagxTWBIMKFfjB4xLxMr3712rM0.49qWo4gpAQ4t5C
EEuhhCsckHx4UnTfQxdHelEcEQkWA6e2BtNugzzonS3RHWkTaqP98Ws9fmMV
i1INJYUqmSIPk5c.iSGodNhv+iqDxyukDDuQ6+M2cMCph44vTMIJ7sYQIuqW
HY8hF467YrK66bpQUpq4Ggaw0ZaU9rntrzQFetHgn1f3dGHzbPVWoDWnjULE
6.khrQHpCDxlMhoKuOL7hDN0BgcAfD60ws.tg5r0pvMsZXssOwUcS+MnotY5
iMf6YAvPgzLD5uiU7mDy9arvT5Vc859bCJXdhOs2bscbciVX4fXRaMsZIgF0
R8EF5ZYFvf949u3+a3xc73C
-----------end_max5_patcher-----------

[tremblap]

also affects bufmel/chroma/bufchroma/nmfmatch

[tremblap]

nrt versions are affected as they won't grow their output larger than the initial number of outputs

rt versions are affected as described in the original issue above.

[tremblap]

fixed