Thursday, August 20, 2026

School series 73

 Voice Recognition — How Does a Machine Know Your Voice?

We can recognise a familiar voice even without seeing the person.

But can a machine do the same?

Yes—and it does so by studying patterns in sound.

When we speak, our vocal cords create vibrations. These vibrations travel through the air as sound waves.

A microphone captures those waves and converts them into electrical signals. A computer then analyses characteristics such as pitch, rhythm, pronunciation and the way particular sounds are produced.

When a voice-recognition system has been trained with examples of a person's speech, it can compare a new voice with the patterns it has learned.

This is why your phone or voice assistant can sometimes respond differently when it recognises a familiar speaker.

But there is an important distinction.

Speech recognition asks:

“What did you say?”

Voice recognition asks:

“Who is speaking?”

Modern artificial intelligence can perform both tasks remarkably well.

Our voices are therefore like a kind of sound signature—not identical to a fingerprint, but containing patterns that can help a computer distinguish one speaker from another.

So when a machine appears to “recognise” your voice, it isn't hearing the way we do.

It is measuring sound, finding patterns and making a very fast identification.

No comments: