Start .NET rewrite with wire and media crypto conformance

docs/architecture.md

@@ -1,5 +1,10 @@
 # Architecture
 
+The parallel .NET rewrite under `dotnet/` currently implements shared protocol framing,
+voice headers, and media crypto. Existing server/client/audio behavior remains in C++.
+See `docs/api-dotnet.md` for the initial managed contract and
+`docs/porting-to-dotnet.md` for subsequent migration phases.
+
 ## 1. The shared-core model
 
 All non-UI logic lives in one C++ library, **`libvoicecat`**. The same library is linked
@@ -205,8 +210,10 @@ callback: no allocations, no blocking calls.
 ```
 
 - **Voice router is a relay, not a mixer.** For each incoming voice frame it looks up the
-  sender's channel and forwards the *unmodified Opus payload* (restamped with the sender's
-  user id) to every other subscribed member. No server-side decode/transcode → low CPU,
+  sender's channel and forwards the *unmodified encoded Opus bytes* to other members.
+  It authenticates/decrypts incoming media, then reseals with each recipient's directional
+  key and counter. SSRC/timestamp/flags/codec pass through; sequence and ciphertext/tag change.
+  No server-side decode/transcode → low CPU,
   low latency, and end-to-content is just Opus. Per-channel Opus params are enforced so all
   members are mutually decodable.
 - **Subscriptions.** Clients implicitly subscribe to their current channel's voice; text