Some of you may immediately recognize this number. It’s 2 to the 16th power, a very important number for digital audio. It represents the resolution of a 16-bit audio source along the voltage axis. Resolution is the number of squares. Or, as is often seen in diagrams of digital audio sources, it represents the accuracy of the waveform’s step height.
Sound is recorded on both the voltage axis and the frequency axis, so let’s take a closer look at the CD format, 44.1 kHz/16-bit.
At 1 kHz 0 dB, the standard used when designing audio equipment, the voltage axis is represented by 65536 steps. But what about the time axis resolution? It’s just 44 steps. 65536 vs. 44. Doesn’t this strike you as overwhelmingly unfair? Although the height of the steps is more than 1000 times higher, allowing for extremely fine resolution (precision), only 44 steps are ultimately produced.
So what about 1 kHz -60 dB, which is said to be the (measured) playback limit of LP records? The voltage axis is 1/1000, so it’s 65 vs 44. It’s finally balanced. But at 10 kHz -60 dB, it becomes 65 vs 4.
As someone who absolutely loves the sound of LP records and designs non-feedback amplifiers for work, I’d like to say: isn’t 16-bit sufficient on the vertical axis for digital formats? I think that increasing the sampling frequency would achieve a better balance between the voltage axis and the time axis of the music signal. To begin with, even 16-bit precision is meaningless with a non-feedback amplifier. (lol)
“Since we can’t hear sounds above 20 kHz, there’s no point in increasing the sampling frequency. Increasing the bit depth is more important.”
I often hear this opinion, but when you look at it from the perspective of “resolution,” it looks completely different, doesn’t it? That’s what I was trying to say.