# Guitar — articulation diagnostic 01

**Experimental and unapproved. Not a completed guitar or sax model.**

Same first solo chorus of Giant Steps, 19.2 seconds, unchanged written melody,
onsets, strings, backing, physical string and amp chain. A retains the frozen
calibrated model's connected pitch paths. B picks each scored note at its actual
pitch and velocity; there are no connected synthetic bends. This also changes
note gates and excitation, so it is an articulation comparison, not a
single-parameter timbre experiment.

The current model has no physical hammer-on or pull-off mechanism. This page
does not pretend to add one. It tests a picked baseline before committing to
more complex finger/string interactions. The string and pickup tone remain
provisional, and it is not demonstrated to sound better.

`a-band` and `b-band` contain identical backing, within 24-bit rounding.
`a-guitar` and `b-guitar` contain the same passage without backing. A constant
post-amp gain matches guitar loudness; any shared safety gain is reported.
There is no new melody, bus processing, recorded sample or paid API.
All earlier recordings, pages and sources remain unchanged.

The complete source recipe is local in `music4/studies/guitar-articulation-04/`:

```sh
node render.mjs
OPENBLAS_NUM_THREADS=1 python3 export.py
```

`render-report.json` contains each actual performed event and source hashes;
`report.json` records levels, output hashes and preservation checks. Dependencies
are documented in the local study. This page has no autoplay or audio preload.