If this is indeed the case, then all the reviews, opinions, listening tests, and other types of sampling means absolutely nothing and there is no way to tell whether any design is good or bad or indifferent. In this scenario, you will always listen to a specific units and all opinions will be based on a specific unit. Furthermore, if there is audible variation among units, that would tell me that a manufacturer's quality control and manufacturing documentation are beyond horrible.

I have not personally experienced this, but it certainly is possible. The use of specially picked, sorted, laser-trimmed, etc. components by high end manufacturers implies there is a certain degree of variability.

Buddha,
In electronics, I don't have two of any identical units to compare. However, I do have two amplifiers of the same make and model. Unfortunately, one has been tweaked with upgrades and the other has not been. I acquired the second one which previously had been upgraded by the former owner. If the upgraded one had not been upgraded, it would be possible to compare in long-term listening tests, and in a double-blind test whether the two are the same or not. If it were found to be the same, then I would have made the tweaks to one of them and compare again. That would settle the question whether those particular tweaks make a difference or not.

As for speaker units, I have played pink noise in one channel or the other channel alone. Each channel sounds different. Here, room effects probably make the most influence. So I cannot make conclusions from this.

In another set of speakers, I swapped the left and right speakers. They sound different, as to imaging and balance of left/right cohesiveness (not knowing how to accurately describe what I heard) of the pair. In this case, I suspect the two units are not the same, perhaps in directivity, or some other attribute. Phase-wise maybe they were good, because out-of phase signals appear to come from either behind me, or off to the directly left or directly right.

I have compared two different CD players from different manufacturers. One is a budget unit from nearly 20 years ago with capacitor and wiring upgrades; the other is a more expensive, stock (as far as I know) unit from about 6 years ago. The test was done non-blind, with short-term listening to the same musical piece on two units back-and-forth, then another piece on the two units, and so on. After matching levels (at 1 KHz or pink noise, I can't remember which), I could not hear any obvious difference between the two. Maybe there is a subtle difference, but I could not be sure. The owner of the more expensive unit, said he could hear a slight difference in the favor of his unit.

Quote:

Second, when looking at parts, like capacitors or resistors, we often see, even for the "good stuff," as much as 10-20% tolerance or variation from part to part. Even cherry picking parts to match yields variability to the degree that finished units could vary considerably from each other.

Speaker drivers can differ from each other, as well. Then add the fact that crossovers contain electronic parts, and, whoa, Nellie, things can get complicated.

You have something here

If I cast my mind back to the mid 1970's when I began designing. The tollerances that were acceptible to a company with market economics in mind( such as the ones I worked for) were indeed around 10%-20%.

Easily 4 band 10% resistors where available and if you wanted to near triple the cost you could elect to try for 5%.

Wound components for crossover purposes and older polyester capacitors were always causing difficulties in maintaining concistency. For laboratory purposes we would sit some spotty 17 year old student on a Wheatstone bridge matching up pairs.

In the world of big power public address (JBL - Citronic - CAudio - Amcron/Crown) this work was pretty unnecessary -- no one cared less.

I dont work in this field anymore at component level - even us old farts now use software. But the ease with which sub 1% tollerance being readily available with little economic downside, seems to suggest that these days there isnt going to be much variance.

I have on my bench right now something the old guard would probably smile at with warm nostalgia, first ....and then rememberance of what a piece of shit it really was in its day.

A Cambridge Audio P110 (circa 1976) - everything in it was 20% tollerance. It was so bad even from new that each channel had a different character both in specifications under test and sound.

It has since seen my modifying hand changing everything to 1% and changing the, mexican made, bias taps to ceramic precsion Cermets - The power amp is pretty much the same with one small modification ... I gave the power transitors a heatsink to sit on --- I thought it important - Cambridge didnt.

Nice thread Buddah ....thx

Speaking of tight tolerance parts, I had done some capacitor mods in my preamp (and amp, tuner, CD player). Specifically in the preamp, I measured that the DC voltages were virtually identical between the two channels at corresponding points in the circuit. I don't remember the tolerances used in the stock parts, but my impression was that it was tighter than typical. So in replacing caps with polystyrene ones, I tried to get a pair nearly of the same value by measuring them and as close to nominal as I could find. Perhaps (speculation) that may be responsible for the particular character of the sonic improvement - instruments do not interfere with each other compared to before, among other improvements heard.

I think symmetry in design has helped as well as component tolerance. This way we dont have a Left/Right channel variances as was in my old Cambridge.

But this leads me on to another point I wish to test here...

If so much is discussed about the audio pathway when it comes to the inter-connects external to the active hardware (Signal cables and Speaker Cables) why is the majority of the pathway ignored?

The cables had their own pylons, classic. That is some funny stuff.

You bring up some interesting points, that are rarely discussed. I think what you said is a big part of it - that the external cables are more amenable to tweaking and interchanging than internal conductors.

Circuit traces on double sided boards do exhibit a capacitance when the traces cross. I've heard that this is more of an issue at higher frequencies, such as MHz to GHz range, and not so in audio range (I didn't do calculations to check).

Some internal components, such as capacitors and resistors, have leads that are magnetic - that is a magnet sticks to it. This is supposedly undesirable for audio. Other manufacturers make these same components with leads that are not magnetic. So in this sense, the internal parts do matter, if the premise is correct.

The same goes for leads from the RCA connectors to the board, or from the buttons and pots to the board, etc. I've used teflon insulated leads for the internal connections on locations where insulation is needed. From point-to-point on the board, un-uinsulated wire works, if risk of shorting is not present.