Originally Posted by mobycat
I REALLY hate to bring in wikipedia, but regardless...
Mindbender, please explain this:
Those vertical lines are where digital samples. Between those lines, digital does NOT record. It interpolates what is there.
Analog does not interpolate that. It records the entire wave.
That's partially correct. Digital (CD digital anyway), records at 44,100 samples per second. People like you claim that's not enough and that's why analog is better. The Nyquist theorem states that the sampling rate has to be (a minimum of) twice the highest frequency you wish to record. Since human hearing (of an infant) extends to 22KHz, 44.1 KHz was chosen (for commercial CDs) as the sampling rate. In part, this was a compromise so that at least 70 minutes of stereo music could fit on a CD.
Analog does not "record" the entire wave any more than a Xerox machine makes an exact copy of a document. Analog reinterprets and reformulates the entire wave through electro-mechanical means. That's why each time you make a copy, you have severe generational loss. You also have errors that aren't generational. For example, how well does the stylus track an LP groove? If the stylus can't accurately track it, you get distortion of the original wave form. If the stamper used to press that waveform into the vinyl is worn, you get distortion. If the vinyl itself has flaws, you get distortion. (And by distortion, I don't just mean harmonic distortion, like when you turn the volume up too loud --- I mean any variant of the original wave form.) In analog recording, if the level is set too high, you get tape saturation - a distortion which some engineers will use for creative purposes, but distortion nonetheless.
There are digital systems with higher sampling rates. Some systems sample at 96,000 times per second instead of 44,100. For an original live recording, especially an acoustic recording, it can sometimes make a difference that can be perceived. I have a CD-recorder that is capable of recording 96/24, but frankly, when making copies from LPs, even LPs that were recorded in analog, I can't hear any difference whatsoever.
However, digital does interpolate in the Y axis of your diagram. CD digital is a 16 bit system. In a 16 bit system, there can be 65,535 different values of "level" (actually voltage). When an analog signal is converted to digital, one of those 65,535 levels must be chosen even if the actual level falls in-between. This process is called quantization. When the exact level isn't "chosen", there is what is called quantization error. If the process is upgraded from 16 bit to 24 bits, there is much higher resolution in the voltage domain - 1.67 million different values or 256x the voltage resolution as compared to a 16 bit system.
(There are other factors, such as what happens when not enough bits of the system is used and noise is generated. "Dither noise" is introduced to compensate, but I won't get into those issues for purposes of this discussion. Analog has its counterparts, such as Dolby noise reduction.)
There are plenty of people who swear that analog is a better recording process, but at least insofar as recorded content is concerned, if I playback a vinyl LP recorded by analog means and a CD-R of that same album in sync and switch between them, you will not be able to tell which is which, especially in a double-blind test. Now there are those "subjectivists" who don't believe in double-blind testing, but I don't want to get into that argument here.
I am a vinyl fan and still have several hundred vinyl LPs in my living room, but those who claim that analog is always superior to digital are fooling themselves, IMO. All I remember back in the 70s, before the advent of the CD, is how much we used to bitch about poor pressing quality and the lousy sound of LPs, especially in the U.S. Every time I hear a CD that to my mind, sounds inferior to the way I remember the vinyl, I go back to the vinyl and listen and inevitably, the vinyl sounds far worse. There are those who swear that the original British pressings of Beatles LPs, for example, sound far superior to the recent remasters on CD. I don't happen to think the remasters are as much as an improvement as was claimed, but I also don't think they sound worse than the original LPs.
Yes, under absolutely perfect conditions ($20,000 turntables, $50,000 speaker systems, perfect pressings), vinyl can sound terrific, but we have to deal with the real world.
I think what we're remembering when we think analog sounds better are those terrific tube-based amplifiers and original model loudspeakers from the peak of the hi-fi era, from such manufacturers as AR, Advent, McIntosh, Dynaco, etc. as well as our emotional response to hearing that music for the first time. There's nothing wrong with digital audio recording, except for the way it is used. Because every group wants to be the "loudest", the dynamic range of most CDs is far less than it was in the vinyl days, even though CD is capable of 96db dynamic range. We're usually using only about 25 to 35db of it.