Each microphone is sampling a single point in space over time. It's an incredibly simple signal.
The problem in reproducing it exactly resolves to the problems of materials engineering and efficient coupling of your transducer to the air.
Meanwhile, your brain is doing an amazing job of extracting information from two point signals separated by the width of your skull. There's information that comes from the way that the shape of your ears distorts waves coming from different directions; timing information between the two ears; tilt information from kinesthetics and eyes. All that is handled by systems with a hundred million years or more of evolution in the Earth's biosphere, plus learned updates from your measly few years of life.
And 16 bits per sample, 44.1KHz sampling rate, will suffice to record it all as well as your ears can hear in any sort of normal environment. (A few more bits will suffice if you intend to capture both jet engines and soft breezes in the same recording.)
Microphones have a host of issues as well. But the reason people notice the difference between cheep speakers and good ones is speakers have to deal with an instantaneous signal, if you reincode based on the characteristics of the speaker you can get a lot from even fairly cheap speakers.
I'm not a 100% sure, but it probably involves convolution processing, as is used in some professional speakers - particularly for room optimization[0][1].
Comments
Each microphone is sampling a single point in space over time. It's an incredibly simple signal.
The problem in reproducing it exactly resolves to the problems of materials engineering and efficient coupling of your transducer to the air.
Meanwhile, your brain is doing an amazing job of extracting information from two point signals separated by the width of your skull. There's information that comes from the way that the shape of your ears distorts waves coming from different directions; timing information between the two ears; tilt information from kinesthetics and eyes. All that is handled by systems with a hundred million years or more of evolution in the Earth's biosphere, plus learned updates from your measly few years of life.
And 16 bits per sample, 44.1KHz sampling rate, will suffice to record it all as well as your ears can hear in any sort of normal environment. (A few more bits will suffice if you intend to capture both jet engines and soft breezes in the same recording.)
Microphones have a host of issues as well. But the reason people notice the difference between cheep speakers and good ones is speakers have to deal with an instantaneous signal, if you reincode based on the characteristics of the speaker you can get a lot from even fairly cheap speakers.
How would one go about doing that?
I'm not a 100% sure, but it probably involves convolution processing, as is used in some professional speakers - particularly for room optimization[0][1].
[0] http://www.genelec.com/products/dsp-products/glm-software/ [1] http://www.youtube.com/watch?feature=player_embedded&v=k...