newIns.md documents SID3 filter scaling center notes as 1 byte; the fork writes 2

[axionbuster]

Problem

papers/newIns.md describes the SID3 per-filter scaling center-note fields as single bytes holding a 12-EDO note value:

  1  | cutoff scaling center note: `0` is `c_5`, `1` is `c+5`, ..., `179` is `B-9`
  1  | resonance scaling center note: `0` is `c_5`, `1` is `c+5`, ..., `179` is `B-9`

Both the field width and the value range are wrong for this fork's current instrument format.

Actual behavior

src/engine/instrument.cpp writes both fields as 16-bit values:

w->writeS(sid3.filt[i].bindCutoffToNoteCenter);
w->writeS(sid3.filt[i].bindResonanceToNoteCenter);

The read path branches on format version, taking readS() for current files and readC() for older ones. Legacy values are migrated with DIV_EDO31_LEGACY_OFFSET and clamped to DIV_EDO31_MAX_SLOT, so the stored value is a slot in the 465-slot 31-EDO domain, not a 12-EDO note number. FREQ_FOR_NOTE in src/engine/platform/sid3.cpp interprets it against DIV_EDO31_A4 and DIV_EDO31_STEPS, confirming the 31-EDO reading.

179 is therefore not B-9 here. In this fork's domain, slot 179 is Bbb0, and B-9 is slot 462.

Impact

This is the fork's own format specification, so anyone writing a parser from it will misread every byte after the first scaling field in an SID3 instrument — a size error, not just a mislabeled range.

Suggested fix

Document both fields as 2 bytes for the current format version, note the 1-byte form for earlier versions the way other version-dependent fields in this file already are, and give the value range as slots in the 465-slot domain.