服务端流式返回音频数据，每个chunk经过decodeAudioData后，将其中的buffer进行拼接，块与块之间会存在卡顿。请问是服务端返回的音频数据格式有问题吗？还是拼接方式有问题？

const audioCtx = Taro.createWebAudioContext();
let tempList = [];

const task = wx.request({})
task.onChunkReceived((res) => {
  audioCtx.decodeAudioData(
    res.data,
    (buffer) => {
      tempList = [...tempList, buffer];
    },
    (err) => {}
  );
});

setTimeout(() => {
  const buffer = audioCtx.createBuffer(1, tempList.reduce((pre, item) => item.length + pre, 0), 44100);
  let seek = 0;
  for (let i = 0; i < tempList.length; ++i) {
    buffer.copyToChannel(tempList[i].getChannelData(0), 0, seek);
    seek += tempList[i].length;
  }

  const source = audioCtx.createBufferSource();
  source.buffer = buffer;
  source.connect(audioCtx.destination);
  source.start();
}, 5000);