Music Coding

 LPC-based codecs model the sound source to achieve good

compression.

 Works well for voice.

 Terrible for music.

 What if you can’t model the source?

 Model the limitations of the human ear.

 Not all sounds in the sampled audio can actually be heard.

 Analyze the audio and send only the sounds that can be heard.

 Quantize more coarsely where noise will be less audible.-Music Coding

　LPC-based codecs model the sound source to achieve good

compression.

　Works well for voice.

　Terrible for music.

　What if you can’t model the source?

　Model the limitations of the human ear.

　Not all sounds in the sampled audio can actually be heard.

　Analyze the audio and send only the sounds that can be heard.

　Quantize more coarsely where noise will be less audible.