TTS/Audio Specialist for YouTube

I run a YouTube channel producing long-form calm reflection videos (60–90+ minutes) based on ancient Christian writings: youtube.com/@MeditationsOfTheFathers I'm looking for an authoritative, consistent voice I can commit to for all future videos. The Problem: The current pacing and intonation read as text-to-speech, and over 60–90 minutes that's the main thing viewers notice. I need a voice that: - Works at 100–120 WPM for long, reflective content - Sounds natural and calm — not obviously AI - Clearly distinguishes narration from quotes — when the video moves from commentary into a Church Father's actual words (e.g., St. Augustine), the listener should feel "ok, now I'm hearing Augustine's words," and when it transitions back to narration, that shift should be obvious and intentional, not jarring My Production Setup (You Need to Understand This): The entire pipeline runs through a custom AI agent called Hermes Agent that I control via Telegram on my phone. I communicate with Hermes, and Hermes executes the production workflow on my Mac. Any solution you recommend must work within this setup — you give me the changes, I relay them to Hermes via Telegram, and Hermes implements them into the toolchain. You don't need direct access to my system. Here's the specific stack: - TTS: Fish Audio S2-Pro API — cloned male voice with zero-shot reference audio - Audio processing: ffmpeg (128k AAC + noise reduction via afftdn) - Script & workflow: Hermes researches sources, drafts the script, and feeds it to the TTS pipeline — all orchestrated through the Hermes agent CLI - Visuals & rendering: fal.ai (Seedance 1.5 Pro), assembled by the Hermes workflow - Delivery: I review everything via Telegram attachments; Hermes uploads the final video to YouTube So your recommendations need to be implementable as configuration, command-line flags, script formatting rules, or provider settings — no GUI-only or manual Audacity workflows. What I Need: 1. Listen and diagnose — review the channel and tell me in plain language what you hear that doesn't sound natural (pacing, intonation, quote transitions, listener fatigue over an hour) 2. Propose fixes — for each issue you hear, describe how you'd adjust it: TTS settings, script formatting, silence timing, or post-processing 3. Quick screen-share — I'll walk you through how Hermes drives the pipeline so your fixes fit the setup 4. Prove it — work with me to apply the changes, then produce 1 video together with audio I'm happy with You Should Be Comfortable With: TTS tools (Fish Audio, ElevenLabs, PlayHT), audio post-processing (ffmpeg, Audacity, iZotope RX), long-form pacing, and ideally some familiarity with API-driven or agent-assisted pipelines. To Apply: Listen to any video on the channel, then reply with: 1. What you hear that doesn't sound good (be specific — pacing, intonation, transitions, fatigue) 2. How you'd adjust it 3. ONE audio clip where you made AI TTS sound natural — ideally showing a narration-to-quote transition Generic proposals without these will be ignored. Budget: $200 fixed price for the audit + collaborative production of 1 video with audio I'm satisfied with. Open to negotiation. Experience Level: Intermediate Skills: - YouTube Marketing - YouTube Development - Audio Post-Processing - TTS / AI Voice