eli5 How does the brain separate out frequencies it’s hearing so that we can make out different instruments in a song? It seems like some insanely complex analysis behind the scenes??
In: 330
The cells in your inner ear are what convert sound from air pressure waves to electrical signals in the brain. They physically vibrate in response to the air pressure and induce a signal in the specific part of the auditory nerve to which they’re attached. This isn’t so different from how cells in your retina work, or how an actual microphone works (it just uses a little magnet and wires instead of cells and nerves).
The important part for your question is that each individual cell responds to a narrow range of frequencies. So your brain decodes the information essentially from which cells are stimulated.
The brain is an expert at pattern recognition. The brain is efficient et recognizing faces, shapes and other familiar things and sounds are no different. The brain categorizes sounds it hears individually with respective instruments; it learns sounds just like it would learn words. When listening to a song, it listens to parts of the song where only one or just a few instruments play and from that it’s easy to recognize the instruments. In music there are often moments where just a few instruments actually make a sound at a brief period of time. Even sounds that have a few instruments playing at a time can be identified by “simulating” or combining instruments in your head to match what you are hearing thus getting the instruments used. This is all done “automatically” or subconsciously. The latter method is called bruteforcing and is inefficient.
Our brain is doing something that we call Fourier Transform. It is designed to compute the frequencies that the sound consists of. FT is quite time consuming for the computers, but our brain evolve troughout millions of years to do it apparently quickly.
The cells in your inner ear are what convert sound from air pressure waves to electrical signals in the brain. They physically vibrate in response to the air pressure and induce a signal in the specific part of the auditory nerve to which they’re attached. This isn’t so different from how cells in your retina work, or how an actual microphone works (it just uses a little magnet and wires instead of cells and nerves).
The important part for your question is that each individual cell responds to a narrow range of frequencies. So your brain decodes the information essentially from which cells are stimulated.
Our brain is doing something that we call Fourier Transform. It is designed to compute the frequencies that the sound consists of. FT is quite time consuming for the computers, but our brain evolve troughout millions of years to do it apparently quickly.
The brain is an expert at pattern recognition. The brain is efficient et recognizing faces, shapes and other familiar things and sounds are no different. The brain categorizes sounds it hears individually with respective instruments; it learns sounds just like it would learn words. When listening to a song, it listens to parts of the song where only one or just a few instruments play and from that it’s easy to recognize the instruments. In music there are often moments where just a few instruments actually make a sound at a brief period of time. Even sounds that have a few instruments playing at a time can be identified by “simulating” or combining instruments in your head to match what you are hearing thus getting the instruments used. This is all done “automatically” or subconsciously. The latter method is called bruteforcing and is inefficient.
This is called selective auditory attention, or the cocktail-party effect (where you can pick out one person’s voice from many others in a loud environment), and it is my understanding that how it works is not particularly well known. That said, the task is considerably more difficult if you only have one ear, so that’s a factor.
This is called selective auditory attention, or the cocktail-party effect (where you can pick out one person’s voice from many others in a loud environment), and it is my understanding that how it works is not particularly well known. That said, the task is considerably more difficult if you only have one ear, so that’s a factor.
Your ears have structures that physically detect sound [frequencies at different locations](https://en.m.wikipedia.org/wiki/Place_theory_(hearing)), the information that you heard those frequencies is then sent to the brain separately and [processed in separate specialized](https://en.m.wikipedia.org/wiki/Topographic_map_(neuroanatomy)) portions of the brain.
For example, you have low frequency detectors in your ear that send info about low frequencies to a part of the brain that just processes low frequencies. From there the brain can combine the separately processed information and interpret it.
Edit: This process is complex and this is a simplified version of just one theory of how it works.
Our brain is doing something that we call Fourier Transform. It is designed to compute the frequencies that the sound consists of. FT is quite time consuming for the computers, but our brain evolve troughout millions of years to do it apparently quickly.
Latest Answers