1.8.0 — group QoM returns a different number, and the old one was confounded (#364)

`group_qom`, `pose_qom` and `normalized_qom` are re-exported from `micromotion`,
which released 1.0.0 today. They averaged over every marker at every frame while
the underlying `band_limited_qom` interpolates gaps, so an occluded marker
contributed a near-zero speed and still counted in the divisor. The result
tracked how much the cameras saw rather than how much the body moved: 16 to 17
per cent low on a realistic dropout pattern, with the speed series correlating
up to +0.70 with the per-frame count of visible markers.

The new default, `normalize="visible"`, excludes each marker at the frames where
that marker was absent. Pass `normalize="worn"` to reproduce a figure published
with an earlier release, and say which you used. Clean pose data with no
dropouts is unaffected; occluded mocap is affected most.

The `micromotion` floor is now `>=1.0.0`. That is a correctness floor rather
than a documentation one: below it the re-exported function returns the
confounded number.

All 545 tests pass against micromotion 1.0.0 unchanged, which also means none of
them covered the behaviour that changed.


Claude-Session: https://claude.ai/code/session_01Qx7SWwc4XGYgXNNKrcNAgP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

Author: alexarje

CHANGELOG.md

@@ -7,6 +7,25 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.8.0] — 2026-08-05
+
+### Changed
+- **`group_qom`, `pose_qom` and `normalized_qom` return different numbers.** They are re-exported
+  from `micromotion`, which released 1.0.0 today, and the change is a correctness fix rather than a
+  refinement. They averaged over every marker at every frame while the underlying
+  `band_limited_qom` interpolates gaps, so an occluded marker contributed a near-zero speed and
+  still counted in the divisor. The result tracked how much the cameras saw rather than how much
+  the body moved: 16 to 17 per cent low on a realistic dropout pattern, with the speed series
+  correlating up to +0.70 with the per-frame count of visible markers.
+
+  The new default, `normalize="visible"`, excludes each marker at the frames where that marker was
+  absent, and lands within 0.8 per cent of the unoccluded value. Pass `normalize="worn"` to
+  reproduce a figure published with an earlier release, and say which you used. Pose data from a
+  clean, well-lit recording with no dropouts is unaffected; occluded mocap is affected most.
+
+- The `micromotion` floor is `>=1.0.0`, which is a correctness floor rather than a documentation
+  one: below it the re-exported function returns the confounded number.
+
 ## [1.7.1] — 2026-08-05
 
 ### Fixed

docs/releases.md

@@ -1,6 +1,6 @@
 # Release Notes
 
-The current stable release is **MGT-python 1.7.1**.
+The current stable release is **MGT-python 1.8.0**.
 
 Install or upgrade from PyPI:
 
@@ -16,6 +16,15 @@ which is the single source of truth for release notes.
 
 ## Recent highlights
 
+### 1.8.0
+
+- **Quantity of motion over a group of markers returns a different number, and the old one was
+  confounded.** `group_qom`, `pose_qom` and `normalized_qom` come from `micromotion`, which
+  released 1.0.0 today. They used to average over every marker at every frame while gaps were
+  interpolated, so an occluded marker contributed almost no speed and still counted in the divisor,
+  and the result tracked camera coverage rather than movement. The new default excludes a marker at
+  the frames where it was absent. Pass `normalize="worn"` to reproduce an older figure.
+
 ### 1.7.1
 
 - **The published API pages described a band the package no longer uses.** They showed

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "musicalgestures"
-version = "1.7.1"
+version = "1.8.0"
 description = "Musical Gestures Toolbox for Python"
 readme = "README.md"
 license = {text = "GPL-3.0-or-later"}
@@ -25,9 +25,10 @@ classifiers = [
 ]
 dependencies = [
     # >=0.3 was wrong twice over: no such release exists on PyPI below 0.6, and the functions
-    # this package re-exports arrived much later. 0.15.2 is the floor that makes the
-    # committed API pages true, since they are generated from that package's docstrings.
-    "micromotion>=0.15.2",
+    # this package re-exports arrived much later. The floor is now 1.0.0, where `group_qom`
+    # stopped counting markers it could not see. Below that the re-exported function returns a
+    # different, confounded number, so this is a correctness floor and not only a docs one.
+    "micromotion>=1.0.0",
     "numpy",
     "pandas",
     "matplotlib",