As of: 2026-10-14 00:22 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/v/fish-audio-drama-3-is-here-launch-video/

# Fish Audio Drama 3 is Here

Fish Audio, 13 October 2026, YouTube. 3,685,033 views across 2 copies, as of 13 October 2026.
Kind: Official.
Watch: https://www.youtube.com/watch?v=L5IjgBBuCpg

Copies counted in the views:
- YouTube, this upload: 194 views (2026-10-13)
- X: https://x.com/FishAudio/status/2110038429199450126, 3,684,839 views (2026-10-13); Fish Audio's launch post on X with the 163 s video (same length; assumed to be the same video)

## Why it is here

Fish Audio's official release video for Drama 3 (2:43). The YouTube upload had under 200 views; the same-length video in the launch post on X had about 3.7M views a few hours after release (the post offered a free month for commenting 'DRAMA', and its links carry influencer tracking, so the reach is not purely organic). Listed as one work on the basis of identical length (163 s). Whether this is the video with the licensed Uncle Roger voice is to be confirmed when it is described.

## Description (written by Gemini from the video)

**Summary**  
This is an official launch video for Fish Audio’s speech synthesis model, Drama 3, presented by comedian Nigel Ng in character as Uncle Roger alongside Fish Audio’s founders. While cooking egg fried rice, Uncle Roger tests a cloned AI version of his voice that dynamically roasts his cooking, responds to emotion and delivery instructions, switches languages, and demonstrates in-line word replacement.

**What is shown**  
* **Voice Cloning & Live Audio Dialogue:** Uncle Roger clones his voice on Fish Audio Drama 3 with granted permission, and the AI interacts with him in real time while tracking roasts on a "Roast Counter" [00:11–00:30].  
* **Emotion & Tone Control via Natural Language:** After the AI delivers a line flatly ("You hear sizzling? I hear my ancestors crying"), Uncle Roger prompts it to sound like it has given up on life and is crying; the model regenerates the phrase complete with crying and sobbing inflections [00:32–00:49].  
* **Cross-Lingual Voice Synthesis:** Uncle Roger asks the model to insult him in another language, and the AI reproduces his distinctive voice and accent speaking Japanese to criticize clumpy rice ("なぜご飯が手をつないでいるの？離して。") [00:56–01:07].  
* **Smoothness & Emotion Tuning:** Uncle Roger prompts the AI to speak with more emotion and smoothness regarding MSG, and the model adjusts its cadence accordingly [01:24–01:36].  
* **Word-Level Editing (In-Line Regeneration):** After the AI calls the fried rice "a disaster," the UI demonstrates selecting only the word "disaster," editing it to "masterpiece," and regenerating just that segment while preserving the remainder of the generated audio [01:54–02:06].  
* **Founders & Platform Details:** Fish Audio co-founders Jiahua Liu (CTO), Shijia Liao (Chief Scientist), and Rissa Cao (CEO) eat the fried rice and introduce Drama 3's availability on Fish Audio and via API [01:48–02:37].

**Claims & numbers**  
* The presenter claims Drama 3 is Fish Audio's "most controllable voice model yet" [00:03] and "most emotive model yet" [02:12].  
* Rissa Cao claims users can "change or emphasize a single word or sentence without regenerating the rest" [02:14].  
* Creators and developers who comment "Drama" within 48 hours get 30 days of free access [02:24].  
* For enterprises switching from another provider, the company offers the first three months free [02:30].

**Notable quotes**  
* "You hear sizzling? [crying/sobbing] I hear my ancestors crying." — AI Uncle Roger [00:43]  
* "Change or emphasize a single word or sentence without regenerating the rest." — Rissa Cao [02:14]  
* "Try out Uncle Roger voice on Fish Audio today." — Uncle Roger [02:34]

**Assessment**  
This is a polished official promotional launch video that blends sketch comedy with product demonstrations. While the conversational pacing and UI graphics are staged for comedic and narrative flow, the underlying audio clips showcase actual Drama 3 synthesis features, including emotive conditioning, cross-lingual voice cloning, and single-word audio regeneration.

_Described by gemini-3.8-flash on 2026-10-13 from the video's audio and frames._

## Related

- 2026-10-13: [Fish Audio releases Drama 3](https://postcutoff.com/e/2026-10-13-fish-audio-drama-3/)