[BUG]: Speech recognition in DebateRoom fails silently on unsupported browsers and lacks feedback on permission denial

### Bug Description

In `DebateRoom.tsx`, the Speech-to-Text feature relies on the Web Speech API (`SpeechRecognition` / `webkitSpeechRecognition`). Currently:
1. **Unsupported Browsers (e.g., Firefox, Brave):** `recognitionRef.current` remains `null`. Clicking the microphone button does nothing silently without explaining why to the user.
2. **Permission Denied (e.g., Chrome/Edge):** When a user denies or dismisses the browser's microphone permission prompt, `onerror` fires with `"not-allowed"`. The app handles this internally by turning off the recording state, but displays **no message or notification to the user**.
---

### Steps to Reproduce

**Test 1 (Unsupported Browser):**
1. Open the app in **Firefox**.
2. Go to `/startDebate` -> **Practice with Bot** -> enter `DebateRoom`.
3. Click the microphone button.
4. *Observed:* Nothing happens; no error or message is shown.

**Test 2 (Permission Denied in Chrome):**
1. Open the app in **Google Chrome**.
2. Enter `DebateRoom` and click the microphone button.
3. Click **Block** or close the browser permission prompt.
4. Click the microphone button again.
5. *Observed:* The mic toggles briefly and resets, but no feedback explains that permission was denied.
---

### Logs and Screenshots

_No response_

### Environment Details

OS: linux (ubuntu)
browser: firefox, chrome
node: v20.20.2

### Impact

Medium - Feature works but has issues

### Code of Conduct

- [x] I have joined the [Discord server](https://discord.gg/hjUhu33uAn) and will post updates there
- [x] I have searched existing issues to avoid duplicates

Start in DebateRoom.tsx by tracing the microphone button, recognitionRef.current setup, and the onerror handler. Reproduce the unsupported-browser case in Firefox and the permission-denied case in Chrome, then verify that each condition gives the user feedback and that recording still resets correctly.

由索引模型根据 Issue 内容生成。