048 conversion only supports marc codes in subfield a, not subfield b or 2.

[kirkhess]

The way the codes are mapped in marc2bibframe2, the codes are extracted from subfield a and subfield b into bf:instrument bf:ensemble, or bf:voice with textual labels not codes. While subfield 2 and ind2=7 is ignored, the source is hardcoded.

When they come through bibframe2marc, only subfield a is supported along with ind2=blank; no subfield b or subfield 2 are created. WC has about 600k v048$b. There's about ~3k v048$2, and in that case ind2 would be 7; so for example, the IAML medium of performance codes wouldn't work with the lookup table.

A tweak to use bf:code instead of strings, and mapping bf:source should fix ind2=7 & $2; a soloists property/class may or may not be needed for $b support, perhaps "not bf:ensemble" will work?

[jodiw01]

The MARC-to-BIBFRAME conversion of the 048 and 382 fields needs further review and reconciliation with the PMO Medium of Performance ontology.