一尘不染

有没有办法从音频文件中获取类似分贝级别的东西,并将该信息转换为json数组?

json

这样我就可以使用这些信息来协调页面动画,例如随着分贝级别越高,使元素越亮


阅读 470

收藏
2020-07-27

共1个答案

一尘不染

这种方法将在Chrome / Safari中起作用:

+function(){



  var ctx = new AudioContext()

    , url = 'https://cf-media.sndcdn.com/OfjMZo27DlvH.128.mp3?Policy=eyJTdGF0ZW1lbnQiOlt7IlJlc291cmNlIjoiKjovL2NmLW1lZGlhLnNuZGNkbi5jb20vT2ZqTVpvMjdEbHZILjEyOC5tcDMiLCJDb25kaXRpb24iOnsiRGF0ZUxlc3NUaGFuIjp7IkFXUzpFcG9jaFRpbWUiOjE1MTUwNDM5Njd9fX1dfQ__&Signature=FfmL2qUssAKs3Z7EPoYo0Yq8-SAg8rKLPs65EasXwuVkfsOB4joFqeCvVR2elpaG-lJaV4hXpXFiRCDWXNOYyAtO4Oz~sexiPwIoSk8-jWiVbGQRS8TMmUmj7TJzxemMOIj7ugWJKk6PHsrUdgqs9woDpHzxmkGCzk6sfqJEIsdeZJ4rWUFAh4iGWn9M6b0xfzTgndAJmytkNj9raCpWCBVmdr5u-r9nt~q5uF1easNSW9oaFilM4s1Hq2ei~VJye8zW9bzvrGm8idVdy-tiPeMWAKcE8J2VuaS1Ret6jRTRaHTDuiNgA5sZvgTzNpEpKtWI7UmAWI5TrqNVSlxpgQ__&Key-Pair-Id=APKAJAGZ7VMH2PFPW6UQ'

    , audio = new Audio(url)

    // 2048 sample buffer, 1 channel in, 1 channel out

    , processor = ctx.createScriptProcessor(2048, 1, 1)

    , meter = document.getElementById('meter')

    , source



  audio.crossOrigin = 'anonymous'



  audio.addEventListener('canplaythrough', function(){

    source = ctx.createMediaElementSource(audio)

    source.connect(processor)

    source.connect(ctx.destination)

    processor.connect(ctx.destination)

    audio.play()

  }, false);



  // loop through PCM data and calculate average

  // volume for a given 2048 sample buffer

  processor.onaudioprocess = function(evt){

    var input = evt.inputBuffer.getChannelData(0)

      , len = input.length

      , total = i = 0

      , rms

    while ( i < len ) total += Math.abs( input[i++] )

    rms = Math.sqrt( total / len )

    meter.style.width = ( rms * 100 ) + '%'

  }



}()


#meter {

  width: 0%;

  height: 15px;

  margin: 2px 0;

  background: green;

  -webkit-transition: width .05s;

}


<div id="meter"></div>

重要的事情发生在这里:

processor.onaudioprocess = function(evt){
  var input = evt.inputBuffer.getChannelData(0)
    , len = input.length   
    , total = i = 0
    , rms
  while ( i < len ) total += Math.abs( input[i++] )
  rms = Math.sqrt( total / len )
  meter.style.width = ( rms * 100 ) + '%' 
}

基本上,您每隔2048个样本就获取一次原始PCM数据(值从-1到1),然后遍历它们,从而计算给定时间段内的平均信号电平。

然后,您可以使用该值制作动画。

编辑: 更新为使用RMS,正如Jason指出的那样,RMS是一种更有意义的度量。

2020-07-27