Consider re-implementing THREE.VideoTexture to use `tick` since each one creates its own RAF

Description: We could reduce the number of requestAnimationFrame if we implement THREE.VideoTexture ourselves using tick. Here is the source of three.js VideoTexture

THREE.VideoTexture = function ( video, mapping, wrapS, wrapT, magFilter, minFilter, format, type, anisotropy ) {

  THREE.Texture.call( this, video, mapping, wrapS, wrapT, magFilter, minFilter, format, type, anisotropy );

  this.generateMipmaps = false;

  var scope = this;

  function update() {

    requestAnimationFrame( update );

    if ( video.readyState === video.HAVE_ENOUGH_DATA ) {

      scope.needsUpdate = true;

    }

  }

  update();

};

THREE.VideoTexture.prototype = Object.create( THREE.Texture.prototype );
THREE.VideoTexture.prototype.constructor = THREE.VideoTexture;

Start by locating the THREE.VideoTexture implementation and the existing tick mechanism, then compare them with the three.js source shown in the issue. Done means VideoTexture updates through the shared tick path instead of creating an individual requestAnimationFrame loop, while retaining its video-ready update behavior; no test file is named in the issue.