Quote:
Originally Posted by audioengr /img/forum/go_quote.gif
No Gauntlet here.
|
Good, thanks for the clarification Steve. An inherent problem with forums and/or email is that the non-verbal content of a message is missing, so things are often misinterpreted. Your post seemed a little aggressive to me, and I have no desire to get into an exchange of barbs, especially with other manufacturers. I think it is unfortunate that this has occurred as frequently as it has in this and other forums. I was unsure of the gist of your post; thus my query into your intent.
Quote:
Originally Posted by audioengr /img/forum/go_quote.gif
And congratulations on what is evidently a great product.
|
Thank you! And my congrats to you also; I have a lot of respect for your product. You've done a lot to help create this product niche and deserve credit for that.
Quote:
Originally Posted by audioengr /img/forum/go_quote.gif
The difficulty is that I have seen similar posts elsewhere and they can lead many to believe that analog and digital are really the same, which we both know are not.
|
No, but far more worrisome to me is that many people are convinced that they're unrelated, almost as if they behave according to two separate sets of physical laws. Typically, this devolves into the tired "bits is bits" argument that has been trotted out so many times when talking about digital audio. Of course if that argument were true, there wouldn't be much justification for the premium products either of us is offering.
In my view, the similarity or difference really depends on the level of abstraction. The point of my original text you quoted was that at a sufficiently low level (say, the electron level), all of the physical effects exerted by something like a capacitance will be the same regardless of whether that electron is part of what we happen to call a "digital" or an "analog" signal. Certainly from a signal integrity point of view, it's all analog, all the time. This is really one of the central messages from key people in the SI field like Johnson, Archambault, Bogatin, and others, though they may not articulate it quite that way. I agree that digital and analog circuits are generally treated, designed, and optimized differently, but in order to do high-speed digital design well, one has to have some command of the decidedly analog effects on a signal which represents digital content. The transmission interface doesn't know whether a signal is analog or digital; in the case of our products the term "digital" only has real meaning on the receive side, where the signal is processed according to an arbitrary set of rules which, by convention, endow it with a particular meaning. In the telecom world they're primarily concerned with SI problems that would give rise to bit-level errors, while those would be vanishingly rare in our digital audio world of comparatively short signal paths. We're concerned with signal integrity - primarily jitter and noise - inasmuch as they contribute to an inaccurate digital to analog conversion. Some of the origins of those problems might be digital in nature (such as a PLL divider problem), but most are analog in nature.
Quote:
Originally Posted by audioengr /img/forum/go_quote.gif
These are certainly foreign terms to me. Educate me. What is "whitening" and a data slicer? Is this from telecommunications circuits? I've only been in computers.
|
Yes, the terms are most frequently seen in the telecom world. Very briefly, a data slicer is basically a one-bit A/D converter - essentially a comparator. It accepts a "digital" RF signal and decides what is a 1 and what is a 0. To whiten the data (which is very much a digital process) means to process it by XORing it with a pseudorandom sequence, in order to reduce the maximum symbol length, or the maximum possible length of contiguous ones or zeros. Much like Manchester encoding (or the bi-phase encoding we see in a SPDIF signal, which is basically the same), this serves to effect more frequent bit transitions and reduces or eliminates the DC content of the signal. That helps linearize the data slicer's operation. It also has the effect of producing a more Gaussian spectral shape of the RF signal.
Relating this back to my original post, my point with this example (and perhaps it was a poorly-conceived one) was simply that calling a signal digital doesn't make it immune to analog problems that afflict various parts of the overall signal transmission chain.