I beg to differ; although not exactly the same but our first controlled listening experiment (singleblind) back in 1984/1985 was about the difference between two foil capacitors with different dielectrica (Polyester vs. Polypropylene) used as coupling capacitors in a phonopreamplifier. As said before, having read Dan Shanefield´s article about blind testing, i found his reasoning convincing and we tried it. Of course quite amateurish at that time, not knowing much about sensory tests and their difficulties, but 5 trials with correct identification needed in each trial to reject the null hypothesis.
Obviously a popular belief, but apparently not true in general. That shouldn´t be _that_ surprising as our listening sense is a nonlinear device......
Did the coupling capacitors form a significant part of the LF rolloff? If so, then it might just be possible to hear the difference in dielectric nonlinearity. Did extra stray capacitance from the (presumed physically larger?) polypropylene affect HF response? Were the caps exactly the same value i.e. did you measure them? I note that the test was single blind.
People like bad cables. There must be a reason for this. It could be cost alone, but the effect of extra interference is my preferred explanation. Anyway, my point is that small degradations may be perceived as a change but this change may sometimes be misperceived as an improvement - especially in situations (e.g. almost all the time) when no comparision with the original sound is possible.
People like bad cables. There must be a reason for this. It could be cost alone, but the effect of extra interference is my preferred explanation. Anyway, my point is that small degradations may be perceived as a change but this change may sometimes be misperceived as an improvement - especially in situations (e.g. almost all the time) when no comparision with the original sound is possible.
I think there is another factor to cables. We all want something that is unique and our own. Look at cars - there are so many combinations of options that it is unlikely that you meet someone with exactly the same car as yours, so you 'build' your car to be unique to yourself.
Same with audio. What are the options to 'build' your audio as unique by yourself? Exchanging a DAC filter chip or changing the amplifier display logic is not really for Joe Consumer.
What remains? Tube rolling, opamp rolling, capacitor rolling, and yes, cable swapping. No or very little actual knowledge required, and endless possibilities to make it uniquely yours.
Your perception apparatus, in combine with your ego circuits, will make sure it sounds better and better 🙂
Jan
Same with audio. What are the options to 'build' your audio as unique by yourself? Exchanging a DAC filter chip or changing the amplifier display logic is not really for Joe Consumer.
What remains? Tube rolling, opamp rolling, capacitor rolling, and yes, cable swapping. No or very little actual knowledge required, and endless possibilities to make it uniquely yours.
Your perception apparatus, in combine with your ego circuits, will make sure it sounds better and better 🙂
Jan
As far as I know my car is standard, so there could be hundreds like it. I regard a car as a vehicle, not a personal statement. Maybe the fact that my car is standard is my personal statement?
No, the capacitors were just for coupling the MC Headamp to the MM amp, input impedance was 47 kOhms//80pF, nominal cutoff frequency was ~1 Hz.
No, the stray capacitance of the (not so much larger MKP) didn´t affect HF response as far as i could see, but measurement was restricted as only a 20 Mhz scope was at hand.
Yes, we measured and both were within 2% of their rated capacitance (bridge basic accuracy +- 0.1%)
Valid hypothesises but just that..... 🙂
Last edited:
Sure. Lots of people also have unmolested, bog standard audio stuff just as it came out of the box. But IF you want to make a statement, rolling some easy replaced parts - and cables! - is about all that you can do. And at that point the price also becomes part of the statement.
Jan
Ok, you have covered the obvious things. Do you have an explanation for a coupling cap with little signal voltage across it somehow affecting the sound? Which sound was more accurate i.e. like the original?Jakob2 said:
A google search reveals results that suggest the increased uF is not dangerous 😀
...it's 0.223% greater (approx.), i.e. outside the 20% range.
Last edited:
Buy a Mega328 component tester off ebay for $10 so you can measure stuff roughly.
Nichicon is notorious for measuring 10-15% below-spec from new anyway.
😀
Nichicon is notorious for measuring 10-15% below-spec from new anyway.
😀
Google again. A 2200 +/- 20% can be max 2640. That is almost 50% more than 1800!
Jan
Suprsising, isn´t it? 😉
Everything else equal it seems that a better dielectricum delivers better sound quality (correct identification was needed for the rejection of the null hypothesis, but identification was based on differences accompanied by strong preference).
But or course the "everything else equal" part is debateble as we did no further research with a capacitor manufacturer to ensure that construction was identical and therefore the dielectricum the only independent variable.
But, also different construction as reason looked back then quite strange, as according to the manufacturers they weren´t that different.
Later, we noticed the socalled audiophile capacitors (wondercaps, sidereals and so on) doing only one more controlled test for confirmation purposes and sighted listening as standard procedure, i got the impression that it would be better to have no capacitor at all.
What was more accurate is hard to say (mainly because no original was available 🙂 ), but the preferred capacitor delivered a better approach to realism.
Unfortunately i couldn´t work out an idea to do a true test for transparency.
By 'better dielectric' do you mean more linear or some other parameter?
The biggest difference between capacitor and no capacitor is a little LF phase shift. This could be audible.
Transparency means signal out = signal in, apart from the unavoidable high pass filter formed by the capacitor and whatever follows it.
It is a pity your tests were only single blind, as that limits their significance.
The biggest difference between capacitor and no capacitor is a little LF phase shift. This could be audible.
Transparency means signal out = signal in, apart from the unavoidable high pass filter formed by the capacitor and whatever follows it.
It is a pity your tests were only single blind, as that limits their significance.
Sorry! ...still at least several contributors have now confirmed this, makes me feel more secure when you all back each other up in this way!
Cheers!
Don't assume that because people agree on an audio website this means they are right. People all telling the truth will all agree, but so will people who have all swallowed the same myth. Fortunately, the truth about circuit theory is widely available for anyone to learn.
Given the limited voltage swing across and the quite similar temperature niveau, the dissipation factor seemed to be the relevant variable, "lower" = "better" .
I knew, as we have auditioned the differences between highpass cutoff frequencies at and well below 1 Hz, for me/us everything below 0,3 Hz was safe (and of course it depended on the bass extension of the speakers).
But that we could trace back to the cutoff frequency by comparing servos to RC highpass. The remaining difference between cap and no-cap still was perceptable.
It is interesting that you asserted even at cutoff frequencies that low a difference could be audible, as the literature afaik gave group delay numbers of several milliseconds in the range from 55 - 200 Hz.
I know, why do you mention it?
While that apparently is true for any independent observer not present, we were very strict, although as already admitted not knowing much about proper sensory tests.
Starting point was sort of qualitative description of the sound differences we perceived during sighted listening, given by a colleague and myself we both agreed on quite specific reproduction details. In fact we both thought the difference was that large that a controlled listening experiment must show it too. We did A/B comparison, including "same trials" (as said little did i know at that time; correct identification of the capacitor was needed), but did solder in and out every time to get the time span in the same region and to not allow any hint due to noise. Source was analog; my colleague started first and failed.
Then i did the test, trying to remember the difference and comparing it to the delivered sound. The first 4 trials i was sure he exchanged the capacitor every time but in the final trial the difference suddenly hits me and so i knew, that he did use the same capacitor for the first 4 trials and only exchanged it in the last. I changed my scores accordingly and after finishing the last trial he "deblinded" the trials and my results were correct.
That experience raised my interest in sensory testing. This episode illustrates the difficulties one can have under specific test conditions. And of course sometime later it became obvious that using such a small number of trials isn´t generally recommended as the risk of type 2 errors gets way to large.
- Status
- Not open for further replies.