Add Camb AI as a TTS provider

[neilruaro-camb]

Summary

Hi! We're the team at Camb AI, the localization engine trusted by brands like the Premier League, the NBA, NASCAR, the Australian Open, and more. We'd love to be included as a TTS option in abogen.

This PR adds Camb AI as a third TTS provider alongside Kokoro and Supertonic, using our streaming MARS models (mars-flash, mars-pro, mars-instruct) with support for 140+ languages.

What's included

• New file abogen/tts_camb.py - CambPipeline class following the same callable protocol as Kokoro/Supertonic
• Full backend integration - provider factory, voice resolution, synthesis dispatch, job config, settings defaults, form handling, voice profiles
• Speaker Studio UI - Camb AI appears as a provider option with voice ID, model, and language controls
• Settings page - API key, model, voice ID, and language defaults
• Preview support - voice preview works through the existing preview API

How it works

• Uses the camb-sdk Python package (streaming TTS endpoint)
• Requests PCM float32 audio at 24kHz to match abogen's internal format
• Voice selection via integer voice IDs (e.g. 147320)
• API key configurable via CAMB_API_KEY env var or the Settings page

Testing

• All 177 existing tests pass with no regressions
• Verified live synthesis produces valid 24kHz float32 audio
• Preview API returns correct WAV output

Thank you for building such a great project. Happy to address any feedback!