Capacitor upgrade in crossover - You CAN'T handle the TRUTH! - Part 3

[ctrl]

So, after the topic capacitor upgrade has already been discussed by me in two parts, there is no reason to talk about it anymore - or is there?

The full mini series:
Capacitor upgrade in crossover - Is it audible?
Capacitor upgrade - part two
Capacitor upgrade - part three
Replace resistor by low-inductance resistor - Is it audible?
Different Binding Posts - is it audible?
Audible difference in high-end capacitors? - ABX samples


I would like to go into more detail on two points. But this is the last part promised!
1. Hides the "bad sound quality" as not measurable distortions?
2. You CAN'T handle the TRUTH! GR-Research proves crossover components upgrade works



1. Hides the "bad sound quality" as not measurable distortions?

Quite a few will think that the good man simply cannot accept that the quality improvements of a capacitor upgrade are simply not measurable.
Since there are no magic air vibrations that a microphone cannot record, but should be audible to the ear without any problems, the above "explanation" is really hard to accept.

In order to find these "mysterious distortions" we will take a closer look at the multitone measurements made in part 2. Where a 35 year old electrolytic capacitor from Visaton (nominal 100µF) was compared to a film capacitor - each case in series with a midrange driver. The multitone distortion of the driver was measured in the near field with each of the two capacitors.

The sonic improvements or deviations that result from the use of a film capacitor, which are supposed to be "well audible" (at least if you read capacitor reviews), should therefore be "hidden" below the measured multitone distortions.
In the diagram the limits for -50dB and -60dB distortion damping were entered.

The "hidden signal change" which is supposed to lead to the allegedly clearly perceptible sound change when an electrolytic capacitor is exchanged for a film capacitor, is in our example at least -50dB, usually even below -60dB.
To get an idea what this means for the overall signal, just play white noise at -60dB and listen to music at 0dB. Only with high sound pressure and corresponding "pauses" in the music signal is this perceptible.
It is impossible that the absence or presence of such a signal in the entire music signal can lead to such auditory impressions as often described:
http://www.humblehomemadehifi.com/Cap.html

...the Janzten Silver Gold Z-cap is very detailed and sometimes too detailed for my liking, ...They do well in upgrading existing crossovers of speakers that are a bit on the laid-back side of neutral, they will pump some life and energy into such a system...
Verdict: 11-

The maximum frequency response change when replacing an electrolytic capacitor with a film capacitor was 0.2dB - see part 2.
The maximum frequency response change when replacing a normal film capacitor with a high-end film capacitor was 0.03dB - see part 1.
Please keep these dimensions in mind for the next section.

All well and good, but on Youtube there is a video and audio samples that prove the opposite - take that MF!



2. You CAN'T handle the TRUTH! GR-Research proves crossover upgrade works

Already in 2010, Danny Richie from GR-Research equipped a loudspeaker with two switchable crossovers.
Here are a few quotes from him from the thread:

This is for all the objectivist that think parts are parts, wire is wire, all capacitors are the same, yadda, yadda, yadda....
...
Okay, I measured the values used in one of the stock crossovers and made an identical crossover to the stock one using Sonicaps, a Mills resistor and an Erse XQ air core inductor. ...
All switches up towards the stock binding post, and the stock binding posts and stock crossover is used. All switches down, and the lower tube connectors are used with the higher quality crossover parts.
...
The cables and crossover have been playing now for several weeks and have now settle in. That's right, they needed burn in time. ...
The difference between the two networks is very apparent (night and day), so if the nay sayers can't hear this level of differences then they probably really should consider hobbies outside of high end audio.*
...
Vertical and horizontal off axis measurments were posted in that first thread. The upgraded crossover did not change any of that.

*Highlighting by me.

In a nutshell, one could say: Not measurable but clearly audible!

Only this year a video with audio samples was published by NewRecordDay in cooperation with GR-Research - link is below.
The viewer is asked to listen to the audio samples and to try to hear the differences between the crossovers. Judging by the comments, the audio samples have convinced many that replacing normal components with high-quality ones is clearly audible.
(Core inductor coils show high distortion at high voltage levels, but these are no influence at normal room volume - this shall not be discussed here either)

In the video Danny says "...these things measured exactly the same, there is no difference in the measurements but they sound very different..."
The measurements linked under the video are not really identical.


So, what do we expect from the audio samples? Right, these should of course be identical, since the measurement conditions have not changed after switching the crossover. There may be fluctuations due to air movement, but on average these should not have any effect.

So what do we do?
First, we extract the exact same 32 seconds of audio from the audio samples for each crossover. Then we look at the corresponding frequency analysis of both crossovers.
Here is the sample for the crossover A:

It is the sound clip at 01:07 for crossover A. For crossover B we proceed identically and extract the exactly same 32 seconds for each crossover.

As already mentioned, we now look at the frequency analysis of both crossovers. Since up to about 400Hz the room modes dominate, we concentrate on the area above.

That's strange, it was only switched between the two crossovers and nothing else was changed, and as we were assured, the differences are practically impossible to measure due to the exchange of the crossover components - you can only hear them.

Perhaps our eyes deceive us, so we export this data and look at it again.

Once again a difference can be seen. What follows now should already be known from part 1+2.
We normalize to the frequency response of a crossover to see the differences better.

The deviations are from 1 - 2.5kHz on average +0.7dB and from 2.5kHz to 9kHz about -0.2dB *** (see below, updated content)
If you have an equalizer at hand you can reproduce the deviations and switch back and forth. Then check your listening impressions with those in the video.

It seems that Danny didn't take the selection of the component values so carefully after all and there were deviations in e.g. the capacity values.

So what we are likely to hear are not improvements due to more expensive components, but differences due to small component value deviations.
In part 1 it was shown how much even the smallest deviations in the capacitance values affect the frequency response.

*** Update:
In addition, here is the evaluation of the frequency analysis (with higher resolution than in the diagram above) of all three music clips. It always shows the deviations from clip B to clip A when normalizing to clip A.
Therefore, Clip A is shown as a red straight line and the deviation from Clip B as a blue curve. The scaling is 0.1dB, so the y-axis shows +-1dB.
Sample 1, 2, 3


Guess that Ron from NewRecordDay mixed up the crossover versions A and B in sound clip 2.
If A and B are mixed up, the frequency analysis will match sound clip 1 and 3.

But according to the Youtube comments nobody seems to have noticed that

If the measurements were carried out carefully, it can be assumed that there are measurable differences between the crossover A and B.


Update 2022-03-23: Added full mini series links

 

[dualazmak]

Hello ctrl and friends,

Thank you for these interesting threads, part 1, 2, 3. Although I do not have skill and tools for full objective measurements like ctrl described, last summer I experienced great improvement of total sound quality by renovating LC-network+attenuators of my SP, YAMAHA NS-1000. I can say "it is audible for sure!". You would please refer to my post here.

At that time, all of the capacitors, coils and attenuators were replaced by new ones; with high-grade audio film capacitors, with air core thick coils and with fully overhauled/cleaned attenuators. The new alignment of the coils was carefully designed to minimize the possible electromagnetic interactions. Furthermore, I took all the LC-network+attenuators out of the SP cabinet and installed them in an outer network box. All the cabling/connections were done by careful soldering using AWG12 cables.

At that time, I used two of 47 uF (microF) electrolytic chemical capacitors (MUNDORF Ecap100v-47.0uF) for woofer, but in February this year, I changed them to film capacitors MUNDORF MCAP250-47t (white big ones), and again the improvement of woofer's sound was clearly audible.

Through my above experience of renovation of LC-network+attenuators, the total sound quality was much improved in;
- cleanliness and transparency
- speed, especially in low frequency
- sonority (so called..?)
- sound resolution (e.g. I can hear individual violinist in orchestra, individual soprano singer in large chorus)
- 3D sound perspectives (really amazingly)
and "it is audible".

I believe not only the new capacitors but also the new coils, the coil alignment in larger space, careful soldering connections with thick wire, new attenuators, and moving all of them into outer box, are all together contributing the significant improvement of the total sound quality.

 

[A800]

Not everything can be measured.

 

[ctrl]

Hello ctrl and friends,

Thank you for this interesting thread. Although I do not have skill and tools for full objective measurements like ctrl described, last summer I experienced great improvement of total sound quality by renovating LC-network+attenuators of my SP, YAMAHA NS-1000. I can say "it is audible for sure!". You would please refer to my post here.

At that time, all of the capacitors, coils and attenuators were replaced by new ones; with high-grade audio film capacitors, with air core thick coils and with fully overhauled/cleaned attenuators. The new alignment of the coils was carefully designed to minimize the possible electromagnetic interactions. Furthermore, I took all the LC-network+attenuators out of the SP cabinet and installed them in a outer network box. All the cabling/connections were done by careful soldering using AWG12 cables.

At that time, I used two of 47 uF (microF) electrolytic chemical capacitors (MUNDORF Ecap100v-47.0uF) for woofer, but in February this year, I changed them to film capacitors MUNDORF MCAP250-47t (white big ones), and again the improvement of woofer's sound was clearly audible.

Through my above experience of renovation of LC-network+attenuators, the total sound quality was much improved in;
- cleanliness and transparency
- speed, especially in low frequency
- sonority (so called..?)
- 3D sound perspectives (really amazingly)
and "it is audible".

I believe not only the new capacitors but also the new coils, the coil alignment in larger space, careful soldering connections with thick wire, new attenuators, and moving all of them into outer box, are all together contributing the significant improvement of the total sound quality.

As I wrote about your post in part 1 and which is the whole quintessence of the three-part series, if you don't select the components exactly when you replace them the difference you hear is due to the change of the crossover.

I don't doubt that you will notice a difference in sound, only it is not caused by "expensive" high-end components, but by slight differences in the component values and the resulting changed crossover.

 

[RayDunzl]

Not everything can be measured.

Corollary: Not everything can be heard.

 

[A800]

Corollary: Not everything can be heard.

That too.
But if your system is capable enough a lot can be felt.

 

[ctrl]

But if your system is capable enough a lot can be felt.

...what are we actually talking about

 

[A800]

...what are we actually talking about

I have no idea.

 

[dualazmak]

I am very much impressed by ctrl's objective measurement approach, of course.

On the other hand, I like the Siegfried Linkwitz's words of "What is important to the eye is not necessarily important to the ear..." (I found it here), which is almost identical to my policy in my project; "The simpler, the better as far as my ears and brain would recognize so in my listening environment". I feel, at least in my project, "do not be objective measurement addict, but trust ears and brain; my final goal is to enjoy music, not simply sound, in improved quality".

In any way, ctrl's objective approach should be very nice and important references for my (our) rather subjective approach.

 

[krabapple]

Not everything can be measured.

Corollary: Opinions aren't necessarily changed by mere facts.

 

[March Audio]

Not everything can be measured.

If the difference is big enough to hear, the reason for it can be measured.

As always the big variable in these things is the person making the claim. Its the only mysterious "component" that defies understanding and quantification.

 

[A800]

If the difference is big enough to hear, the reason for it can be measured.

As always the big variable in these things is the person making the claim. Its the only mysterious "component" that defies understanding and quantification.

OK.
But what about the Behringer U-Phoria that measures relatively good but sounds like crap for example?
I'm sure there are more components that measure well but perform subpar.
Of course it goes the other way around also.

 

[krabapple]

OK.
But what about the Behringer U-Phoria that measures relatively good but sounds like crap for example?
I'm sure there are more components that measure well but perform subpar.

You're sure of that? You're sure the right measurements were done, the right way? Conversely, you're sure that the components were found to 'sound like crap' in a non-biased way?

I wouldn't be.

 

[A800]

You're sure of that? You're sure the right measurements were done, the right way? Conversely, you're sure that the components were found to 'sound like crap' in a non-biased way?

I wouldn't be.

Measurements are always difficult but it got measured here so I think they got done quite reasonably.
I wasn't biased before comparing it to other cheaper DACs but afterwards... maybe a little.

 

[March Audio]

OK.
But what about the Behringer U-Phoria that measures relatively good but sounds like crap for example?
I'm sure there are more components that measure well but perform subpar.
Of course it goes the other way around also.

Does it? You say that like its a statement of fact. Either of the opinions therein expressed (objective or subjective) could be wrong. I havent seen any information to support either conclusions.

By making the statement you just have, you have pretty much demonstrated my point regarding where the big variable lies.

 

[dualazmak]

Hello ctrl,

Let me just add and ask one point.

Now, we, including you and I, know and confirmed that total sound quality can be improved by some renovation of SP's LC-network+attenuators.

We are also well aware, however, that LC-network circuit and attenuators consume (or waste) significant potions of amplifiers power and "efficiency" before actually driving the SP units, and this is my main reason and motivation to move forward into "multi-channel, multi-amplifier audio system using software crossover and multichannel-DAC" by completely eliminating/avoiding LC-network.

I assume that your ultimate goal would be same as mine, "to hear music in improved sound quality". Do you have any interest or plan to go into such direction of multi-amplifier audio system driving SP units directly without LC-network?

 

[A800]

I havent seen any information to support either conclusions.

So you didn't read my earlier post?
Also if you check reviews about the U-Phoria on thomann.de for example they basically say the same about it.
Too bad I read them afterwards I noticed the bad SQ.

 

[ctrl]

Now, we, including you and I, know and confirmed that total sound quality can be improved by some renovation of SP's LC-network+attenuators.

You put words in my mouth I've never said.
Once again, my measurements have shown that replacing intact components does not improve the sound quality.

This was shown in the first part by comparing a high-end film capacitor with old standard foil capacitors. There is not the slightest indication that the replacement would somehow improve the sound quality.

In the blue frame we see an extremely enlarged representation in the frequency range 8-9kHz with a scale of 0.01dB. No smoothing was performed, it shows directly the untreated measurement.
The red curve of the high-end capacitor shows no "magic oscillation" or especially "fast response".

The acoustic measurement method via the tweeter is even so sensitive that it shows the very low capacitance difference of the high-end capacitor compared to the standard foil capacitors.
High-end capacitor 4.68µF, the standard film capacitors 4.67µF. Therefore the red curve is in this frequency range always slightly above the other two.

Do you have any interest or plan to go into such direction of multi-amplifier audio system driving SP units directly without LC-network?

In fact, I'm currently working on an active DIY project that has to meet specific radiation requirements. The simulations are promising. If it doesn't sound like a crap, I'll (in a few weeks) introduce it here...

 

[krabapple]

Measurements are always difficult but it got measured here so I think they got done quite reasonably.
I wasn't biased before comparing it to other cheaper DACs but afterwards... maybe a little.

Ah, I didn't realize Amir had measured one. I'll stipulate that his is a competent measurement suite . He rated it Recommended. So, that leaves the evidence of 'crap' sound that you refer to.

You are always subject to bias unrelated to sound, whether you are conscious of it or not. That's why sighted comparisons of the 'sound' of digital components like the U-phoria aren't considered valid scientific evidence. Got anything better?

 

[March Audio]

So you didn't read my earlier post?
Also if you check reviews about the U-Phoria on thomann.de for example they basically say the same about it.
Too bad I read them afterwards I noticed the bad SQ.

Read what about what? Amir's review linked above is one of his early ones on the old lower performance AP with very little information.

All subjective uncontrolled reviews by any chance? You are missing the point.