Adding custom carrier waveforms

[XenoxYT]

It would be useful if it would be possible to set a custom carrier waveform similiar to the customNN definition like this:

waveformNN: (definition like customNN)

waveformNN would set the waveform of the carrier wave and customNN the envelope at the beat rate.

Example sequence:

waveform00: e=2 0 1 33 40 100 100 100 1 1 1 1 1 1 1 0
custom00: e=2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 40 100 30 0 0 0 0 0

#for a binaural beat
t1: custom00:waveform00:400+2/20
#for a isochronic tone
t2: custom00:waveform00:400@2/20
alloff: -

NOW t1

[lm7137]

Thanks, this idea does have merit.

As I understand it, waveformNN would be a way to define a custom carrier waveform directly, separate from customNN, which would continue to define the beat-rate amplitude envelope.

That split is technically coherent:

• waveformNN = carrier shape / timbre
• customNN = beat envelope

So something like:

t1: custom00:waveform00:400+2/20
t2: custom00:waveform00:400@2/20

does make sense as a concept.

The one important issue is that a true carrier waveform should normally be bipolar, i.e. it should have both negative and positive values and be centered around zero. That is different from customNN, which is intentionally unipolar because it represents an envelope.

So if waveformNN were ever implemented, I would expect its definition to look more like this:

waveform00: e=2 -100 -60 -10 40 100 80 10 -30 -90 -100 -50 0

rather than a list of only non-negative values.

Why this matters:

• non-negative-only values are fine for an envelope
• but for a literal carrier waveform they either create a DC offset
• or force the engine to recentre/renormalize them, which would make the semantics less literal

There is also one practical DSP concern: a custom audio-rate carrier waveform is more demanding than customNN or spinNN, because it operates at the carrier frequency rather than at the beat/motion rate. That means aliasing and band-limiting would need to be considered carefully.

So the short answer is:

• yes, the idea itself has merit
• yes, it would be distinct from customNN
• but a proper waveformNN would probably need signed samples, not just non-negative ones
• and it would need a careful implementation because custom carrier waveforms are more DSP-sensitive than envelopes or motion shapes

[XenoxYT]

Yes it's something that old sbagen had with waveNN but it was binaural only and was just experimental.

[XenoxYT]

Will it get included in the next alpha release?