vtube-studio/components/Studio.tsx
James Twose 3eff403fb4 feat: Generate VTuber character sheet with expression assets
The Gemini service has been updated to generate a character sheet rather than a single avatar image. This sheet includes the main character and separate assets for closed eyes and an open mouth.

The `AvatarConfig` type and `RiggingEditor` component have been extended to handle these new expression assets (`textureClosedEye`, `textureOpenMouth`). A new `Sprite` component has been added to `Studio.tsx` to correctly render these specific regions from the generated character sheet. The UI has been updated to reflect the new generation process.
2025-11-20 20:55:47 +01:00

299 lines
12 KiB
TypeScript

import React, { useEffect, useRef, useState } from 'react';
import { useFaceTracking } from '../hooks/useFaceTracking';
import { AvatarConfig, Rect } from '../types';
interface StudioProps {
avatar: AvatarConfig;
onBack: () => void;
}
/**
* Sprite Component
* Renders a specific crop of the source image into a target container.
*/
const Sprite: React.FC<{
imageSrc: string;
sourceRect: Rect;
style?: React.CSSProperties;
className?: string;
}> = ({ imageSrc, sourceRect, style, className }) => {
// To display a cropped region (sourceRect) of the image, we use an inner <img>
// positioned negatively and scaled up.
// Example: If sourceRect.w is 0.1 (10%), the image must be scaled to 10x (1000%) size.
const widthScale = 100 / (sourceRect.w * 100);
const heightScale = 100 / (sourceRect.h * 100);
return (
<div
className={`overflow-hidden relative ${className}`}
style={style}
>
<img
src={imageSrc}
alt=""
style={{
position: 'absolute',
top: `-${sourceRect.y * 100 * heightScale}%`,
left: `-${sourceRect.x * 100 * widthScale}%`,
width: `${widthScale * 100}%`,
height: `${heightScale * 100}%`,
maxWidth: 'none',
maxHeight: 'none',
pointerEvents: 'none'
}}
/>
</div>
);
};
const Studio: React.FC<StudioProps> = ({ avatar, onBack }) => {
const videoRef = useRef<HTMLVideoElement>(null);
const [cameraReady, setCameraReady] = useState(false);
// We use the custom hook to get tracking data
const { trackingData, isLoading: isModelLoading, startTracking } = useFaceTracking(videoRef.current);
// Initialize Camera
useEffect(() => {
const startCamera = async () => {
try {
const stream = await navigator.mediaDevices.getUserMedia({
video: { width: 640, height: 480 }, // Lower res is fine for tracking
audio: false
});
if (videoRef.current) {
videoRef.current.srcObject = stream;
videoRef.current.onloadeddata = () => {
setCameraReady(true);
};
}
} catch (err) {
console.error("Error accessing camera:", err);
alert("Could not access camera. Please ensure permissions are granted.");
}
};
startCamera();
return () => {
// Cleanup stream
if (videoRef.current && videoRef.current.srcObject) {
const stream = videoRef.current.srcObject as MediaStream;
stream.getTracks().forEach(track => track.stop());
}
};
}, []);
// Start tracking when both camera and model are ready
useEffect(() => {
if (cameraReady && !isModelLoading) {
startTracking();
}
}, [cameraReady, isModelLoading, startTracking]);
// Calculate styles based on tracking data
const getAvatarStyle = () => {
// Deadzone for jitter reduction
const smooth = (val: number) => Math.abs(val) < 0.02 ? 0 : val;
const rX = smooth(trackingData.rotationX); // Pitch
const rY = smooth(trackingData.rotationY); // Yaw
const rZ = smooth(trackingData.rotationZ); // Roll
const tX = smooth(trackingData.translationX);
const tY = smooth(trackingData.translationY);
// Bounce effect on mouth open (Speaking emulation)
const bounce = trackingData.mouthOpen > 0.1 ? -5 * trackingData.mouthOpen : 0;
return {
transform: `
translate(${tX * 150}px, ${tY * 100 + bounce}px)
rotate(${rZ * 1}rad)
perspective(500px)
rotateX(${rX * 15}deg)
rotateY(${rY * -25}deg)
scale(${1 + trackingData.mouthOpen * 0.02})
`,
filter: `brightness(${1 + trackingData.mouthOpen * 0.05})`, // Slight flash when speaking
transition: 'transform 0.1s ease-out, filter 0.1s ease'
};
};
return (
<div className="h-screen w-full flex flex-col bg-slate-900 overflow-hidden relative">
{/* Hidden Video Element for Tracking */}
<video
ref={videoRef}
autoPlay
playsInline
muted
className="absolute opacity-0 pointer-events-none w-1 h-1"
/>
{/* Top Bar */}
<div className="absolute top-0 left-0 right-0 z-20 p-4 flex justify-between items-center bg-gradient-to-b from-slate-900 to-transparent">
<button
onClick={onBack}
className="px-4 py-2 bg-slate-800/80 hover:bg-slate-700 backdrop-blur rounded-lg text-white font-medium transition-colors border border-slate-600"
>
Exit Studio
</button>
<div className="flex gap-2">
<div className={`px-3 py-1 rounded-full text-xs font-bold flex items-center gap-2 ${isModelLoading ? 'bg-yellow-500/20 text-yellow-400' : 'bg-green-500/20 text-green-400'}`}>
<span className={`w-2 h-2 rounded-full ${isModelLoading ? 'bg-yellow-400 animate-pulse' : 'bg-green-400'}`}></span>
{isModelLoading ? 'Loading Vision Model...' : 'Tracking Active'}
</div>
<div className="px-3 py-1 rounded-full text-xs font-bold bg-purple-500/20 text-purple-400 border border-purple-500/30">
{avatar.name}
</div>
</div>
</div>
{/* Main Stage */}
<div className="flex-1 relative flex items-center justify-center overflow-hidden">
{/* Background Grid/Effect */}
<div className="absolute inset-0 opacity-20"
style={{
backgroundImage: 'radial-gradient(#4f46e5 1px, transparent 1px)',
backgroundSize: '30px 30px'
}}>
</div>
<div className="absolute inset-0 bg-gradient-to-t from-slate-900 via-transparent to-slate-900 pointer-events-none"></div>
{/* Avatar Container */}
<div className="relative w-[600px] h-[600px] flex items-center justify-center z-10">
<div
className="relative w-full h-full flex items-center justify-center"
style={getAvatarStyle()}
>
{/* Main Character Body */}
<img
src={avatar.imageUrl}
alt="Avatar"
className="w-full h-full object-contain drop-shadow-[0_0_15px_rgba(168,85,247,0.5)]"
style={{
// Use clip-path to hide the right-side assets from the main view, keeping only the main character
clipPath: 'inset(0 25% 0 0)' // Hides the right 25% (where assets are)
}}
/>
{/* Dynamic Eyelids (High Fidelity Sprites) */}
{avatar.leftEye && avatar.textureClosedEye && (
<Sprite
imageSrc={avatar.imageUrl}
sourceRect={avatar.textureClosedEye}
className="absolute pointer-events-none z-20"
style={{
left: `${avatar.leftEye.x * 100}%`,
top: `${avatar.leftEye.y * 100}%`,
width: `${avatar.leftEye.w * 100}%`,
height: `${avatar.leftEye.h * 100}%`,
opacity: trackingData.isBlinkingLeft ? 1 : 0,
transition: 'opacity 0.05s linear',
}}
/>
)}
{avatar.rightEye && avatar.textureClosedEye && (
<Sprite
imageSrc={avatar.imageUrl}
sourceRect={avatar.textureClosedEye}
className="absolute pointer-events-none z-20"
style={{
left: `${avatar.rightEye.x * 100}%`,
top: `${avatar.rightEye.y * 100}%`,
width: `${avatar.rightEye.w * 100}%`,
height: `${avatar.rightEye.h * 100}%`,
opacity: trackingData.isBlinkingRight ? 1 : 0,
transition: 'opacity 0.05s linear',
}}
/>
)}
{/* Dynamic Mouth Animation */}
{avatar.mouth && avatar.textureOpenMouth && (
<div
className="absolute pointer-events-none flex items-center justify-center z-10"
style={{
left: `${avatar.mouth.x * 100}%`,
top: `${avatar.mouth.y * 100}%`,
width: `${avatar.mouth.w * 100}%`,
height: `${avatar.mouth.h * 100}%`,
}}
>
{/* Skin Patch - Hides the static closed mouth when speaking */}
<div
className="absolute w-[120%] h-[120%] transition-opacity duration-75"
style={{
backgroundColor: avatar.skinColor || '#fcd3bf',
opacity: trackingData.mouthOpen > 0.1 ? 1 : 0,
filter: 'blur(4px)', // Blends edges
borderRadius: '50%'
}}
/>
{/* Mouth Sprite - Scales based on mouth openness */}
<Sprite
imageSrc={avatar.imageUrl}
sourceRect={avatar.textureOpenMouth}
className="w-full h-full"
style={{
opacity: trackingData.mouthOpen > 0.05 ? 1 : 0,
// Scale open mouth based on volume
transform: `scaleY(${0.8 + trackingData.mouthOpen * 0.5})`,
}}
/>
</div>
)}
</div>
{/* Status Indicator overlay if tracking is lost */}
{(!cameraReady) && (
<div className="absolute inset-0 flex items-center justify-center bg-slate-900/80 z-20 rounded-xl backdrop-blur-sm">
<div className="text-cyan-400 animate-pulse font-mono">INITIALIZING CAMERA LINK...</div>
</div>
)}
</div>
</div>
{/* Control Deck */}
<div className="h-24 bg-slate-800 border-t border-slate-700 p-4 flex justify-center items-center gap-6 z-20">
<div className="flex flex-col items-center">
<span className="text-xs text-slate-400 mb-1 font-mono">MOUTH</span>
<div className="w-24 h-2 bg-slate-700 rounded-full overflow-hidden">
<div className="h-full bg-cyan-400 transition-all duration-75" style={{ width: `${Math.min(trackingData.mouthOpen * 100, 100)}%` }}></div>
</div>
</div>
<div className="flex flex-col items-center">
<span className="text-xs text-slate-400 mb-1 font-mono">HEAD ROLL</span>
<div className="w-24 h-2 bg-slate-700 rounded-full overflow-hidden flex justify-center relative">
{/* Center marker */}
<div className="absolute w-[1px] h-full bg-slate-500 left-1/2"></div>
<div
className="h-full bg-purple-500 transition-all duration-75 absolute"
style={{
width: `${Math.abs(trackingData.rotationZ * 50)}%`,
left: trackingData.rotationZ < 0 ? 'auto' : '50%',
right: trackingData.rotationZ < 0 ? '50%' : 'auto'
}}
></div>
</div>
</div>
<div className="flex flex-col items-center">
<span className="text-xs text-slate-400 mb-1 font-mono">BLINK</span>
<div className="flex gap-2">
<div className={`w-8 h-2 rounded-full ${trackingData.isBlinkingLeft ? 'bg-pink-500' : 'bg-slate-700'}`}></div>
<div className={`w-8 h-2 rounded-full ${trackingData.isBlinkingRight ? 'bg-pink-500' : 'bg-slate-700'}`}></div>
</div>
</div>
</div>
</div>
);
};
export default Studio;