You said:
"Dan,
You used some of the following words:
truncation
conversion errors
not a full 24 bits
errors piling up
Or expressed those ideas for problems, right?
So what do you need at the chip level
for it to be done the right way?
In other words, couldn't the chips be made better so these errors don't exist and would not that be better? I dunno?"
Of course the chips can be done better, and the idea is to know where the next thing is to improve. It is a very big question!
Some of it has to do with the standard itself. The limitations of 16 bits (red book) call for processes such as dither to avoid truncations and noise modulation, but when you go for say 24 bits (DVD Audio), you can forget about such dither. With 441.KHz (CD) you are better off to noise shape the dither, but with say 88.2KH or 96KHz, you have a huge range where to “park” the truncation problems. So your noise shaping problem become an easy one to accomplish.
Some of it has to do with processes and resources. Say you need a hack of a lot of digital computations. It is more likely to be able to do a good job when you take an IC that is “specialized” at doing huge amount of processing and nothing but that, and put it together with another IC that does say just the DA function…
You see, the IC makers often try to cram a whole bunch of stuff into a single IC solution. The whole thing must be limited in space to something the size of a pin head, and it is running “boiling hot” at say 2 watts or so. Than someone “glues it” into a board that fits on a 19 inch rack… Well, yes, there are some support IC’s and often a display and some large connectors… But fundamentally, there is the question at hand: why cram a whole function into an single IC?
Much of the answer is about mass production and selling a complete cost effective solution. I am often amazed at who well the IC makers do things, given their constraints. But is it the best way to go? For mass production and cost effectiveness, the answer seems to be yes. For highest quality? Probably not so.
You see, The IC maker must produce solutions WHILE BEING CONSTRAINED to their own technology. True, their technology offers a lot of advantageous, but also there are a lot of things they simply can not do inside the IC. Anyone familiar with the game of Chess knows that there are only a few basic pieces (King, Queen, Bishop, Knight… Pawn). In electronics we have resistors, capacitors, indictors, semiconductors… I (an equipment designer) can use a whole range of resistor, capacitor and inductor values, and I can also choose them according to properties (the material they are made of and so on). They (the IC designer) are playing the game with a lot less flexibility. They do not have ANY inductors inside the IC. They have almost no space for capacitors inside, only for very tiny value caps, of certain limitd material. For the most part, they often can not handle high voltage or current or much power and so on. Some of their world does offer advantageous the equipment designer is longing for (such as – all parts are so nearby, thus at almost the same temperature, thus they track well, laser trimming for great matching of parts, and so on).
We all end up knowing that the best design is a hybrid of IC and non IC technology. Going for quality, say a design calls for some poly capacitor. The smart IC designer will provide a pin or two to connect an external poly cap (The Cap may be much bigger than the whole IC) .The best IC technology for say a sigma delta DA may be best done with some semiconductor process, 5V supply and so on, that is NOT best for some other function. So you have a different IC process for an OPamp, say for a +/-15V supply, and so on…
To complicate things, all of it, inside and outside the IC, is based on theory, understanding of networks, circuits, signals, math and more. These days, I believe you will get the best quality when the IC guys do what they do best, and stop short of trying to go beyond their limitations. At that point, the equipment designer takes over, and with freedom from the limitations imposed on the IC designers and chip makers, goes for a complete quality product, whatever it takes.
Obviously, this is not what takes place when the main goal is mass production and low cost solutions. So I do not expect the IC maker to lead us in the direction of highest quality. Their bottom line is driven by large quantity production! It costs a huge amount to make a new IC, and they need the volume. I am not sure we should wait for the IC makers to provide all the leadership. I was sad to see them sell 192KHz audio sampling AD and DA. They know it is BS from science and engineering standpoint. Often it is the little guy (like I) that has to fight those king kong size companies to send them in the right direction. I made a lot of progress, man does it take a lot of non paid for work
BR
Dan Lavry