Sync wrapper with Veskslop: streaming/crunch modes, vox/hev voice IDs

The Veskslop integration copy had grown word streaming
(VESKSLOP_TTS_STREAM, stream gain), 1990s crunch filters
(VESKSLOP_TTS_CRUNCH[_BITS|_RATE]), echo timing/level, and first-class
vox/hev voice IDs with preset-aware default metadata. Merge those back
into the standalone repo, keeping the tuned VOX preset (pitch 0, rate 15,
head size 34, lowering 50) in the preset and in the published defaults.

Update the wrapper regression test to the tuned registry values, add
__pycache__ to .gitignore, and document the new variables, voice IDs,
placeholder set and the word-mode-over-VOX precedence in README, NOTES
and TTS-DESIGN.

docs/NOTES.md

@@ -127,8 +127,8 @@ without them.
   the Python port; outputs are identical on the 31-line corpus. Re-check after any
   parser change while the clone is available.
 - VOX mode is Flextalk-only. The wrapper enables it with `VESKSLOP_TTS_VOX=1`
-  (`VESKSLOP_TTS_VOX_CLEAN=1` drops the stutter wrapper) and it takes precedence
-  over `VESKSLOP_TTS_WORDS`.
+  (`VESKSLOP_TTS_VOX_CLEAN=1` drops the stutter wrapper); word mode
+  (`VESKSLOP_TTS_WORDS=1`) wins and skips VOX prosody.
 - Half-Life presets (`VESKSLOP_TTS_PRESET=vox|hev`) target Keith Bell only. VOX starts
   from NetTTS `nettts_flextalk_vox_profile.reg` (pitch 3, rate 17, head 52, lowering 40)
   and is tuned against the game's word assets from `suchipi/half-life-vox-console`: