"Here is a clue. One can obtain the same/similar specs by using a variety of parts types, and parts value. However, only one part type and value of each will give the most accurate music."
My head just exploded.
Don
Didn't mean to explode your head Don. I think I can expand and make my comment easier to understand.
1. By varying the bias on a tube or ss device, the distortion pattern will vary. The distortoin spectrum might not
be that different, but enough if it involved 2nd vs 3rd, or better yet 2nd vs 9th order. The higher the harmonic, the lower the % to be perceived.
2. One sees so many components whose specs state 0,1db from 20 to 20khz or +/- 0,1db 20 to 20khz (0,2db total variation). Very rarely is the actual spec going to be exactly 0,1db at 20hz or 20khz. It could be a variation, say -0,035db, -0,078db or -0,092db etc at one extreme and a different -0,xdb at the other extreme.
Remember that - 0,1db means the frequency response lowered only a small amount, about ~60db down from the midrange reference. However, the ear can perceive tonal changes (frequecy response changes nearly -100db down or more. Besides tonal changes, thin or fullness, we can notice changes in dynamics, sound stage width/depth, "cleanness" etc. Small changes is why so many notice sonic differences between different models, brands of components.
As such a -60db change is huge. All it takes is a small variation of resistance value/quality, and/or capacitor type and value/quality as well. (+/- 0,1db is only a -54db change in FR. that is quite pathetic.)
Never, never accept an article comparing capacitor brands as anywhere even close to being accurate.
3. As mentioned above, capacitors are notorious for being inaccurate. A capacitor is two foils separated by an insulator. Usually the foils/insulator are wrapped around and around, as an example let's say 100 turns.
If one measures with a scope, meter, you won't observe much of a difference between capacitors? Simply put, a scope or meter is not nearly sensitive enough, thus not sensitive enough to measure all parameters. (
The equation 1/2piFC equation (1/6.28 times frequency times capacitance in farads) doesn't really tell us a whole lot either. It is simply an equation that tells us the approximate capacitive reactance (AC resistance) at a particular frequency. It does not give us any information concerning the dielectric absorption (DA), nor the equivalent series resistance, the internal resistance and inductance of the foil turns.
a. In simple terms, the dielectric absorption problem is electrons sticking to the insulation when they should let go and "travel" in the circuit. Energy is stored when it should not be. Below is a graph test demonstrating this effect in voltage (a form of signal). The old fashion picture tube short test is another way. Short the picture tube, wait 30 seconds and watch for another spark when shorted. High DA causes a lack of dynamics, blurring etc, etc.
b. The termination of a capacitor's foils to lead wires affects the charging and discharging of a capacitor. A capacitor actually "fully" charges and discharges every cycle (perfect capacitor). 20khz per second is mighty fast charge and discharge. Any resistance in the foil and inductance affects this charging and discharging.
I hope this helps everyone (even though simplistic). Below is a general type test that shows a capacitor being discharged to zero, and over time the voltage gradually rises. As mentioned above, this is due to electrons slowly releasing from the insulator instead of instantaneously, and being released when it should not be. This slow releasing occurs for several minutes. In music, the milliseconds, two seconds, 30 seconds, 2 minutes later changes bias,
smears the sound etc.
Its a real mess that should be addressed if possible.
I hope this is easier to understand.
cheers
steve