Cable Myths Continued

[DrWho]

Mike, I am sure we disagree on small bits here and there. But, it seems to me we agree that this whole process is riddled with subjective analysis and individual beliefs, and that objective data has only a small input to the process. For instance, if we have already decided that a triode is the right answer, we will strive with objective data on transfer functions to make the triode operate as linear as possible. That's a good and proper use of the objective data, and it has excellent correlation with subjective listening data. e.g. A properly operated triode will have less distortion and sound better than an ill-operated one. But even so, the designer is faced with a thousand more choices on associated bits, succeeding stages, topological considerations. Each one begins with a belief and subjective choice and is only buttressed by objective data after the selection. The design remains an art form.

I agree that we agree on most points? Haha....it's the differences that I find interesting to discuss....

I personally have found that I can place a much larger emphasis on the 'objective data' than what I think you are suggesting - and that's a big shift from where I started as an audiophile. It's probably why I'm so interested in defending/expressing my perspective here. What I mean by this is we can set objective targets before starting a design, and those targets will usually indicate when a design will be perceived as an improvement. This process is absolutely intertwined with all sorts of listening and pre-existing bias, etc... and usually starts with a few experiments identifying what measurable attributes of existing gear are likely the cause of the subjective opportunity for improvement. However, there are still many parts to the design process that just have to be cold-hearted solution driven engineering (like your computer engineering example). When implementing good engineering practice, I think the art lies in the realm of generating the objective targets. How else do you decide when the design is done? Relying on subjectivity alone is dangerous - at least for me. I am continuously more and more aware of the biases I hold and sometimes it takes a lot of conscious effort to be truly open-minded. I think the hardest part of this hobby/profession/industry is learning to be aware of our biases - as many of them are dominated by external influence.

To put it another way, I won't pursue changes or 'improvements' that don't also have an objective basis for improved accuracy....but you better believe I'm probably going to stick to topologies that have a more enjoyable sonic signature when it's not affecting the goals of the design.

It's also a pet-peeve of mine when people poke around in a design without understanding what they're doing, but then try to guise their ignorance with claims of: 'higher levels of understanding that aren't chained down by the cold-hearted objective world'. Maybe it's incorrect to try and fall back on science to reveal their ignorance? Is there even merit to calling these people out?

As far as projects for myself, I find that I never fully commit because there is always some compromise being made and I want to completely understand the sonic impact of every variable. It also takes time to finish a design and I usually learn something new in the meantime, which means I end up starting over from the beginning; reprocessing every decision in light of the new information because the only real motivation is personal exploration in the first place (long sentence, but it's really a single thought in my head). Probably a topic for another discussion, but at what point do you just draw the line and decide to profit from your ideas? It's a huge personal investment when you commit to a design and sometimes it feels like ripping out part of your soul because you're more aware of the compromises than anyone else in the world - or at least I'd argue you should understand a design to where you know where every sonic signature's lever is to be pulled.

[DrWho]

Wouldn't the Aural exciter start in the category of 'creation' and not 'reproduction'?

To Mark's earlier point, I absolutely agree that the playback system is incapable of the 'documentation' of a live event, but I would suggest that we want a playback system that is 'accurate' so we can fully experience the intent of the original artistic creation. It's the same reason for picking attributes of light that reflect the original intent of the artist for the painting hanging on the wall.

The reality is that no playback system is 'accurate' so it's important that the compromises minimize the impact of the original intent....that's absolutely where subjectivity gains its immense value.

[DrWho]

I only set subjective targets. I think objective targets are useful only in the rudimentary FUNCTIONS For instance, in designing a phono preamp, the gain and the RIAA equalization curve accuracy are totally objective targets. They will be fixed no matter what topology is selected. But in the end, the unit must meet a subjective sonic criteria, which for me is simply, "better sounding than the last one I designed."

So when picking a resistor value, how do you pick the "better sounding" one without any objective measure? Do you have a bin of random resistors and just randomly swap them out, labeling them as you go? "Here's my purple bubbly resistor. Here's my tree bark resistor, and this one is like tree bark with a heart etched in it with Dean's and Anarchist's initials. This one has Hilary Clinton and Mark Deneen's initials, but is an evergreen tree"

Don't you do things like make sure changing a resistor value doesn't affect the loop gain bandwidth and maintains a flat frequency response? Noise floor considerations, bias points and distortion, etc....there's all sorts of weird ways these active circuits get affected by any single one resistor value. I can't imagine random guessing - and I know that's not what you do, but I want you to admit that you do objective decision making in the process

Edited March 21, 2014 by DrWho

[djk]

"So when picking a resistor value, how do you pick the "better sounding" one without any objective measure?"

I suspect he means resistor brand or type, not the value.

I had to improve a Mark Levinson product that had different capacitor and resistor types in both channels. After making the capacitors match in both channels the amplifier still sounded wrong, because of the resistors. All the resistors were the correct value, just a different brand in the two channels. No difference could be measured from 10hz~100Khz. The audible differences went away after a 45 minutes warm-up period. The amplifier had no real stereo image during the warm-up period. The owner decided that although he didn't like it, he could live with it. The amplifier was totally unlistenable before the capacitor changes.

[DrWho]

"So when picking a resistor value, how do you pick the "better sounding" one without any objective measure?"

I suspect he means resistor brand or type, not the value.

I had to improve a Mark Levinson product that had different capacitor and resistor types in both channels. After making the capacitors match in both channels the amplifier still sounded wrong, because of the resistors. All the resistors were the correct value, just a different brand in the two channels. No difference could be measured from 10hz~100Khz. The audible differences went away after a 45 minutes warm-up period. The amplifier had no real stereo image during the warm-up period. The owner decided that although he didn't like it, he could live with it. The amplifier was totally unlistenable before the capacitor changes.

Care to share details about the types of resistors and capacitors and the particular circuit topology? Although even that will be anecdotal without the full layout of the circuit and corresponding power supplies.....

Most of the second order effects I've dealt with don't show up in the frequency response.

[DrWho]

I think this is tricky ground. If you take a particular phenomenon and it is examined by the Artist and the Scientist, you have two explanations. The Artist is often claiming that unknown processes are at work. The Scientist is often trying to force-fit an objective explanation that may only have partial correlation, but in the desire to explain all things, insists on the known science of the time. This is exactly what we see with wire and cable. The Scientists is insisting that only the known phenomena must be considered. I think that view is too tight. OTOH, we can't make use of fairy dust either. I am happy to sit with "no answer" on some of these things, and just accept the subjective analysis and wait and see what happens.

Wouldn't the scientist be insisting that "at least" the known phenomena be considered first? (not only)? I don't think any scientist is closed to the idea of new variables. Is your point that a scientific view has no category for the unexplained until a new proof is presented?

I would also humbly suggest that the artist not try to explain the "why" - the result of artistic thinking is more the "what" of the experience, or the ascribing of meaning....arguably more important than the why.

[djk]

The two channels were completely different in their passive parts.

The resistors were Roederstein in one channel, and Holco in the the other.

One channel had an electrolytic input coupling cap, the other has a Wima film type.

I thought I would be OK mixing resistor brands, but I was wrong. It did sound fine after an extended warmup.

It had a triple emitter follower (complemetary), with a cascoded J-FET diff pair, current mirror, Wilson CCS for both the input and Vas stages, The Vas was a Darlington connected pair with a cascode.

It had blown up twice driving a 4Ω load, and had problems overheating.

When I got done with it, it could drive a 1Ω load without blowing up or overheating.

The mechanical design of the heatsink assembly was terrible, and the execution was worse. Alan Hulsebus (Orca Design, Raven, Focal, etc), organized the mechanical re-design, I organized the electrical re-design.

Edited March 25, 2014 by djk