<Precursor to the sampling theory thread.>
We all know the real world is analog, right? All nice and smooth, no nasty discontinuities, no discrete steps in anything we see, touch, smell, taste, and most certainly hear! As an analog design engineer, even one designing data converters for a living, I know that better than most.
Or do I? No, I’m not going into quantum physics (I am a hairy-knuckled engineer, not a real scientist, after all). I went through pre-med, once upon a time, before deciding I preferred engineering. (Truth be told, I had simply had enough of schooling for a while and couldn’t face more years of medical school. Nor, after working my way through as an undergrad, did I have the funding it would take). Shocking stuff, that physiology stuff… I learned that our senses are communicated to the brain through nerve impulses, impulses that are generated by (gasp!) discrete little individual sensors in our ears, tongues, eyes, nose, fingers… Impulses that are transmitted to the brain via small electrical charges across a chemical barrier, charges that occur in discrete, individual little packets. Omigosh, we are digital!
Well, not quite: our brain takes all those signals and makes them continuous so we never really perceive individual neurons firing. A bunch go off, the brain filters everything, and we see (hear, touch – you get the idea) the world as nice and smooth. Once things get too small, there are not enough sensor signals generated to pass our brain’s “I care” threshold, and we do not or cannot process them. With larger signals, our brain takes many individual impulses from many sensors, and creates a continuous world-view for us. Many small discrete samples nicely integrated into a convenient whole.
And, that is why digital audio works.
Yes, I am going to go further, but first I wanted to get past the idea that digital is not, or cannot be, as good as “analog”, because there ain’t no analog. (Cue tears now, I still don’t really believe it!) If the digital signal has enough resolution (bits) and is sampled fast enough, we won’t be able to tell if it is digital or analog. How many bits and how fast is, and probably always will be, debated, but digital as a concept is not inferior to analog. Implementation, well…
In fact, digital can be superior. Digital numbers don’t change with time and temperature like analog components, nor vary with power supply voltage. I can design (well, with a lot of help) a digital filter that does not corrupt the phase in the signal band, has virtually zero noise and incredible dynamic range, and rolls off so steep it would scare my Jeep. I can perform billions of calculations a second on a little chip that outperforms (in spades!) an analog signal processor 100 times or more its size, and work wonders unheard/unthought of ten, twenty, thirty or more years ago.
And yet, I still like analog… – Don (as my turntable gently weeps)
Next up: sampling and the digital realm.
We all know the real world is analog, right? All nice and smooth, no nasty discontinuities, no discrete steps in anything we see, touch, smell, taste, and most certainly hear! As an analog design engineer, even one designing data converters for a living, I know that better than most.
Or do I? No, I’m not going into quantum physics (I am a hairy-knuckled engineer, not a real scientist, after all). I went through pre-med, once upon a time, before deciding I preferred engineering. (Truth be told, I had simply had enough of schooling for a while and couldn’t face more years of medical school. Nor, after working my way through as an undergrad, did I have the funding it would take). Shocking stuff, that physiology stuff… I learned that our senses are communicated to the brain through nerve impulses, impulses that are generated by (gasp!) discrete little individual sensors in our ears, tongues, eyes, nose, fingers… Impulses that are transmitted to the brain via small electrical charges across a chemical barrier, charges that occur in discrete, individual little packets. Omigosh, we are digital!
Well, not quite: our brain takes all those signals and makes them continuous so we never really perceive individual neurons firing. A bunch go off, the brain filters everything, and we see (hear, touch – you get the idea) the world as nice and smooth. Once things get too small, there are not enough sensor signals generated to pass our brain’s “I care” threshold, and we do not or cannot process them. With larger signals, our brain takes many individual impulses from many sensors, and creates a continuous world-view for us. Many small discrete samples nicely integrated into a convenient whole.
And, that is why digital audio works.
Yes, I am going to go further, but first I wanted to get past the idea that digital is not, or cannot be, as good as “analog”, because there ain’t no analog. (Cue tears now, I still don’t really believe it!) If the digital signal has enough resolution (bits) and is sampled fast enough, we won’t be able to tell if it is digital or analog. How many bits and how fast is, and probably always will be, debated, but digital as a concept is not inferior to analog. Implementation, well…
In fact, digital can be superior. Digital numbers don’t change with time and temperature like analog components, nor vary with power supply voltage. I can design (well, with a lot of help) a digital filter that does not corrupt the phase in the signal band, has virtually zero noise and incredible dynamic range, and rolls off so steep it would scare my Jeep. I can perform billions of calculations a second on a little chip that outperforms (in spades!) an analog signal processor 100 times or more its size, and work wonders unheard/unthought of ten, twenty, thirty or more years ago.
And yet, I still like analog… – Don (as my turntable gently weeps)
Next up: sampling and the digital realm.