Voiceprint Technology (II): Fundamentals of audio signal processing

2.1　 To understand voiceprint , Learn audio first

In terms of discipline classification , Voiceprint technology is a branch of speech signal processing , and Speech signal processing belongs to the category of audio signal processing .

Voice signals and sound signal , The difference between the two is ：

• Voice signals It refers to the voice with social significance when human beings speak ,
• sound signal It generally refers to all the sounds that human beings can hear . For example, the sound of an instrument , Sounds made by animals , The sound of a car engine , And people snore 、 Sneeze 、 The sound of coughing , These are audio signals in a broad sense , But they are not voice signals , Therefore, it is usually not within the scope of voiceprint technology research .

Many basic concepts and knowledge in audio signal processing , It is very important for learning voiceprint technology .

Any voiceprint system , No matter how advanced the model is , How sophisticated the algorithm is , Can not do without dealing with sound . Only when the correct audio signal is connected , The meaningful feature representation is extracted from it , The latter model can play its role to the greatest extent .

So this chapter , We will specifically and systematically learn these concepts and knowledge related to sound . This chapter covers a wide range , Involving human auditory perception 、 Audio interface 、 Coding technology 、 Discrete signal processing and many other sub fields . At first glance, these sub areas , It doesn't seem to have much to do with each other . However , When we really embark on research or engineering projects in the field of voiceprint , You will find that all the knowledge in these sub domains will inevitably be used . In an enterprise or research institution , Yes