Clarify benchmark documentation

docs/EVALUATION.md

@@ -7,10 +7,9 @@ cards show the headline results; this document defines the datasets, systems,
 normalization, uncertainty, exclusions, runtime boundaries, and raw artifacts
 behind those results.
 
-> **Interpret results by axis, not as one universal score.** Intelligibility,
-> predicted naturalness, human preference, model footprint, and runtime measure
-> different properties. Inflect does not combine them into a proprietary
-> aggregate.
+Read each metric separately: WER measures intelligibility, UTMOS22 predicts
+naturalness, listening tests record preference, and runtime measures deployment
+cost. None of them is an overall quality score.
 
 ## Evidence at a glance
 
@@ -25,13 +24,13 @@ behind those results.
 
 ![Two-ASR semantic WER](../assets/evidence/asr-consensus.svg)
 
-## Comparator policy
-
-The comparison set intentionally uses serious compact or on-device systems:
-KittenTTS Nano, Piper Low, and Supertonic 3. It does not use deliberately weak
-baselines. Every comparator in this package-level study has a larger deployable
-weight footprint than both Inflect releases. That makes the comparison
-demanding, but it does not mean one benchmark proves universal superiority.
+## Compared systems
+
+The benchmark includes KittenTTS Nano, Piper Low, and Supertonic 3 because they
+are established compact or local TTS systems. In this package-level comparison,
+each has a larger deployable weight footprint than both Inflect releases. That
+provides useful size context, but no single result establishes overall
+superiority.
 
 Voice variants are separate rows when voice identity can affect quality or
 intelligibility. Kitten and Piper may be shown as equal-weight two-voice means

release_manifest.json

@@ -84,8 +84,8 @@
     },
     {
       "path": "docs/EVALUATION.md",
-      "bytes": 12599,
-      "sha256": "b350f2990dc6a192211bd065f5f28920cb513507981f42d76e0ad0bc1e6e724b"
+      "bytes": 12488,
+      "sha256": "9e0d4c1811ca9959529db4e7b6907ca931f094dfd910dbe7f61b550fbc732f17"
     },
     {
       "path": "docs/EXPORTS.md",