GRollins- You're hitting right on the problem, but I can't accept the conclusion that measurements in general are of no help. What's no help is measuring the same old c*r*a*p that we've been measuring for decades. I've little doubt that the differential techniques I like would show differences between your amps of differing sound stage. The problem is, all that could be said is they were different. There's no way to analyze the differences and make any kind of prediction ahead of the listening session as to what we might hear. My frustration is that lots of much smarter people have looked at these sorts of issues for a long time, and reached no consensus at all. In spite of all the inexpensive computer based tools we now have, allowing tests that were impossible a while back, everyone seems to have given up. IMO, maintaining the status quo is probably better for those with a financial stake in selling equipment, than if we really could test and evaluate with certainty.
I have two boys approaching their first birthday, a daughter who's four, a wife, a house to maintain, a writing career (soi disant) that I need to get back to, and a thousand other things to do. What little time I can spare for electronics, I will spend at the bench building circuits, not trying to come up with a thousand ways to test things. If you come up with something that works, please tell us all how it's done, but it's a really, really low priority to me at this time. My ears still work, and that's all I need for now.
Grey
Grey
Grey, I think that you have the right idea. We have been working on this problem for many decades. Nobody has a real solution. Charles and my biggest gripe is the quest for one more decimal point in distortion by the use of more negative feedback. We just have not found that it is that important, yet we have been successful.
There are even more exotic measuring techniques, but beyond this discussion, I am sorry to say. Some day are really accurate differential subtraction test that will automatically separate linear from non-linear distortion will evolve, but it doesn't seem to be here yet.
There are even more exotic measuring techniques, but beyond this discussion, I am sorry to say. Some day are really accurate differential subtraction test that will automatically separate linear from non-linear distortion will evolve, but it doesn't seem to be here yet.
SpittinLLama said:
SL, you have hit an important nail on the head. We need BOTH listening tests and creative lab tests. Together, in an iterative back-and-forth approach, we can possibly move the ball forward and create better designs, while at the same time learning HOW to create better designs the next time.
Listening tests and lab tests both have their limitations, and there will always be things we don't quite understand and where the correlation we seek is missing. There will be times when we make a change, for whatever reason, and it makes the sound better, and then we go back and measure and find no difference. That's just life. It doesn't mean we should stop trying.
It is unfortunate that so many in audio fall into one polarized camp or the other.
Many of those who depend mainly on listening think that those who are good at measurements and appreciate the importance of measurements do not listen.
Many of those who are heavily into measurement think that those who depend mainly on listening are hearing things and don't understand measurements.
Both of these types of characterizations are unfair and ignorant. Each side seems to think that the other side cannot walk and chew gum at the same time.
Cheers,
Bob
Charles Hansen said:
At the molecular interaction level, you get a hysteresis like curve of 'do nothing' at the ac crossover point, under DC bias or no bias, just pure AC. The molecular complexity of the resistive elements, being of a dissimilar nature (oxides, etc). It ends up sounding alot like transistor shot noise. As you have well discovered the physical origin points of the complex HF LCR of a SMT resistor do definitely play into the sonics presented. Clean, pure elements tend to sound the best, due to these aspects. However, when doing it right.. it can be a case of people not getting what they expect in sonics and dissing the lack of highs... not realizing that they've spent the last 40 years listening to transverse originated phase noise not actual transients. One can go broke trying to edjamacate them, via lack of sales...
Some folks are deaf and like high feedback designs, what's a guy gonna do? Next thing you know he'll be buying cables with boxes on them and the whole world goes to hell.
the biggest point that can be stressed here, when attempting to correlate measurements to hearing, is to measure digitally,as i am sure you guys 'up there' do, onthe Ap sets..and have someone write you a software program that calculates and shows graphically the distortion characteristics..but ONLY taking the leading edges of all transients into account in this weighted system. The human ear is non linear, misses most of the back wave and creates harmonics and fundamentals, level, etc, via listening to the leading edges of structure,and the timing between them,and the level of those transients. (The length of time of the leading peak area of the transient plays in as well, very important!) The proof is in that many or most feel that horns sound great..when they have 25 to 50% distortion of the waveform. this is due to the human mechanism,as stated. You'll have to measure at 500khz sampling, minimum and 20 bit depth or so. Then re-weight the measurements according to the stated criteria. Then the numbers will exactly correlate to what the ear hears, for the first time.
On second thought, about a 1mhz sampling rate will be nesessary.
Under such a measurement critera as usually enountered, which is whole signal, the distortion between two seemingly identical signals could be 0.0003%. And a difference is heard between the two. Then re weight, and you might end up with 0.25 or even 3% distortion. Remember, that tiny area is 100%!!! of what the ear works with and hears. It largely ignores the rest. It is also the very nature of the human voice, single ended, all positive and transients. Our ear- brain was designed to decode this.
AJinFLA said:
If you'd like to know, be prepared to drop preconceived notions based on emotions. A long story to explain that remark, longer than you imagine. Basically, truth rarely comes in a package you want, recognize, or are even remotely ready to handle.
Read the recent release 'PSI Spies', By Jim Marr, the story of The US government and Military experiments in psychic research. They did it for over 25 years, that we know of -- and came up with positive results, every single time. They are still doing it (the programs have expanded considerably and gone completely black), but have disinformation junkies out there to villify even their prior heads of research, witness James Randi's absolute ******** attack on Hal Puthoff. He's a paid man, that James Randi. Totally uneducated and hack at best. Why is this done? They do this to get you off the trail of understanding these subjects. You are to remain as a mushroom....
I'm no resistor snob- those Digikey 1/4W MFs are fine for most of what I build, but I tested several resistors for use in scientific instruments and found two I'm wary of. Thick films can have very high excess noise. Miniature metal oxide resistors can have a large voltage coefficient. I'm sure they can be used successfully, but I avoid both of 'em for audio. For low values I wind my own from Manganin wire using one of the non-inductive techniques. IMO, the bulk metal foil resistors are about as good as one can get.
I like the transient explanation, as the differences certainly aren't in continuous tones. Do I see a differentiator as part of the new test setup?
I like the transient explanation, as the differences certainly aren't in continuous tones. Do I see a differentiator as part of the new test setup?
- Status
- Not open for further replies.