Johnny B wrote on Sat, 08 October 2005 05:17 |
Well you talk about "errors," but some errors can sound good, despite what the theory or generally accepted specs may say. Some errors we like, some errors we hate.
As far as you bringing up the cost aspects, the trade-offs and whatnot, you talk just like a million double e's I've dealt with, but the ones I liked the best were always the ones who did not let cost considerations get in the way...Granted, that was a luxury that not everyone enjoys.
As for the "road analogy," I like your concern for environmental values and preserving people's homes (in real life I share those values and concerns with you) but the data path as roadway seems apt. You seem to speak of a wide-lane, wide open highway. Like the Autobahn, where there is no posted speed limit...the actual speed depends on a variety of environmental conditions and other factors...but doing all those intermediate steps look like road blocks and stop lights...Removal of road blocks, yellow caution lights, and red stop lights would seem to have the potential to increase overall efficiency and system performance.
BTW, it may have been Bob K. who brought up an EQ which uses an 80-bit float. Might have been a little more than 80-bit...I already forgot...sorry...
|
Johnny,
Firstly I'm guessing that you must be the only person reading these threads who has missed the number of times I've talked about pleasant sounding errors and unpleasant ones. Unlike you I also happen to know what errors can occur where... IN BOTH DOMAINS.
Secondly I am a real engineer working in the real world, but one who takes a great deal of pride in the quality of his work from all aspects, including the sound quality of what comes out, some systems I do to cost and performance constraints.
Some I do to no constraints other than quality
I happen to know how to do both, I also know how to get the best performance from a system even if quality is the prime consideration.
This is what I do, it is my job, I am very good at it, I am what some people call an expert at it.
From the things you have said in this forum I can only conclude that it is not your job, you know nothing about it. You don't know what the audio algorithms are, you don't know the different ways they could be done, and you don't know about processor architectures, pipelines, caches, multiple execution units, SIMD processing or any of the other things involved.
You are guessing, I am not.
I am not perfect, I do make mistakes, I do have new things to learn, but the level at which you are implying my knowledge is, is just plain insulting.
Now the thing is, you don't have to take my word on this, you could go off and talk with other people who know and confirm the things I've said, you could even learn more about the subject yourself and confirm the things I've said.
Instead you simply repeat the same viewpoint again and again without reconsidering it once, and in doing so imply that I don't know what I'm talking about.
Your stance is that if we go to higher sample rates and wider data words, then things will magically improve. Despite the fact that I and others who know infinitely more about the subject than you do have pointed out the falacies in your thinking a thousand times, you continue to repeat the same view point, because you "have a feeling".
This is not a discussion, it is not a debate, we're not even beginning to drift into the areas of audio processing and conversion where there is a debate to be had.
If it wasn't for the fact that I care enough about this subject, and the industry, to not want people reading who are still learning (and unlike some are still open to learning) to actually believe and propogate the various flawed hypothesis you put forward, I wouldn't even bother.