• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

speaker cable listening comparator

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
I've built a cable comparator. It's a bit basic but seems to be effective. I'd be interested in views of the concept.

I run the two speaker cables for comparison from each channel of a stereop amplifier. Both cables connect to a box at the speaker. In the box there is a two pole change over power relay. The output from that is fed by short cables to the speaker. One of these boxes exists at each speaker.

The relays are driven from a common control from a 12V battery and box at the listening seat. I placed an led in series with the 12V supply - serves handily as an indicator of when the relays are on, ..........and as a fuse ....

The relays are controlled by a 12 way rotary switch. Every second way turns the relays on. A round knob is used with no marker or indication.

Operation is simple - by the time I'd soldered the speaker boxes together I'd forgotten which plugs are nc and which no - but that doesn't matter - it's just another level of obfuscation. Connect the speaker leads being careful to make sure the same type go to the same connections on each side.

Don't look at the led, hide the box, close eyes, whatever.....- rotate the knob randomly a few times - by then you have no idea whehter the relay s are energised or not, let alone which cable is selected. Listen as long as you wish. Swap to the adjacent switch position (the relays change state - the other cable is selected). Listen again. Swap back and forth as much as wished. Try to identify which cable is which by whatever differences are detected. Make the decision and check the led. Record the state - on - or off. Rinse and repeat. Do this however many time is needed for a statistically valid test (which is how many? - gets a bit boring past about 12). Finally, if indeed anything better than random result is noted, with the switch in a known state determine which lead is which by unplugging one.

The amplifier is a high quality SS design with low output impedance. Speakers are B&W 800D3 so not an easy load. The user can choose any music via Roon.

There's nothing absolute about this except to demonstrate whether a user can tell the difference between the two compared speaker cables. in my set up

Does this make sense or have I missed some point that invalidates everything? Is there any value in doing so anyway?
 

tvrgeek

Major Contributor
Joined
Aug 8, 2020
Messages
1,017
Likes
566
Location
North Carolinia
Fun to see someone re-do what has been done thousands of times over at least 40 years. You can then prove to yourself some cables are audible, but eventually will figure out for yourself 16 gage zip cord is as close to a perfect speaker cable as there is. Different does not always mean better.

When people did this before, the usual argument is the relay will mask those great differences with the $500 /foot cables on organic harvested wood posts through cryogenically aligned micro-crystal 9-9s cables made from non-conflict copper mines and insulated with the silk of worms only feed by mulberry trees growing in one district... How could that impure relay not mask the .000001% distortion being played through speakers with 2%? Then there is the paranormal argument that regardless of the randomness of your selection, you will "know" the relay position through an alternative cosmic plane. Yes, I really did see this as a rebuttal to AB blind switching! People are so funny, I don't have to make it up.

An actual valid argument is to have three possible connections. Two the same. When it changes, you assign a number to which you believe is playing. Then do the math. This is actually a valid blind test method. You could easily program an Arduino to keep track of the connection and your response. Amazing what an $8 computer can do for fun.

Hey, it is a hobby, have fun!
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
My simple "solution" to test speaker or interconnect was to use one cable for left, and the other for right, and see if my ear was drawn to or wanted to avoid one side or the other.

I had lots of time and lots of different kinds of scrap cable from work to play with.

There were no amazing revelations.

Any differences were pretty subtle, if there at all.
 

somebodyelse

Major Contributor
Joined
Dec 5, 2018
Messages
3,682
Likes
2,962
I don't see a problem in redoing what's been done a thousand times - see the motto of the Royal Society.

There are plenty of articles on the statistics of ABX testing - including on this site. How many tests is enough depends on how confident you want to be, and how well you can tell the difference. Being right every time you can be quite confident quite quickly, but if you're right 70% of the time you need more tests to be confident it's not just chance. If you're testing a lot of cables then the chance of getting a false positive mounts (chance of throwing a 6 in one throw vs. throwing a 6 in 10 throws), so you might want to redo any positive results to check that it's repeatable.

You may need to soundproof the relay box - in at least one paper they noted some subjects saying that they couldn't hear the difference between the things being tested, but could tell which was which by the sound of the relays switching. The blindness of the test could be questioned too - how random is the initial setting really, and do you start to take shortcuts as the number of repetitions goes up? After all it does get boring and repetitive, and we're only human. The Arduino suggestion is a good option, or you could use something like the switch box from ESP if computers aren't your thing.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Re. silencing the relay -- in the primordial past when I set up a similar test, using a big power relay, the problem was not just the switching sound, but also that there was an instant of no sound as the relay switched. I got around both problems by always switching the relay; sometimes it went back to where it started, other times not. I used a very simple random number generator to create an ABX system and track the results. Eventually (after a few unstable amps were discovered; techies will know what that implies) I switched to a make-before-break relay for no gap (and to keep the amps happy) but retained the circuitry that always switched the relay.

One other thing I did, but people did not like, was added a one-shot circuit that would randomly vary the time between switching. Listeners would know that it switched, but never exactly when it would happen. I had the idea that it would further randomize the testing but the subjects hated it because they wanted to control how long they could listen. Ultimately the circuit was pretty much always left in manual mode (listener hit a button to switch the relay whenever they were ready).
 
OP
M

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
Hmmm.... I'm not really trying to prove anything to anyone except myself, or take this to the nth degree. I posted my listening views about a certain $$ cable that has been subject to a lot of discussion here (now closed) recently. I thought I owed it to myself at least to confirm these views with a blind test. So I compared the cable to a twisted pair of 2.5mmsq cable from a local electronics shop.

Once set up tests are easy and fairly fast. I tested with a friend first - no one was cheating or interested in the click of the relay. Just whether he could tell one cable from another on various music. Tvrgeek is on the money actually. To my surprise in truth - my friend achieved 100% consistency (12 tests), though in about 50% of the tests he said he couldn't pick a difference so didn't want to choose. I then tried for myself and found that I really could tell a difference too - I tried 6 tests., and got 5 of 6 correct.. The differences were very small.

The interesting thing was that we both preferred the cheap twisted pair.

Another step in enlightenment for a reformed audiophile - if not those who have "already done this thousands of times over the last 40 years...." Some of us have to prove or disprove our own beliefs.

I plan to make some measurements next to see if there is any obvious difference in frequency transfer characteristic between the cables - hopefully there is something that correlates with the perceived difference. I'll post once done.
 

Zoomer

Senior Member
Joined
Oct 5, 2019
Messages
323
Likes
468
OP
M

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
So how does this fit in methodologically? o_O
We listened to try to tell one cable from another. As it happened we preferred the sound of one of them - we felt there was a little more "going on / slightly wider soundstage". This wasn't the object, just an outcome. We had no idea which was which at the time. Differences were only just noticeable - certainly not night and day.
 

Zoomer

Senior Member
Joined
Oct 5, 2019
Messages
323
Likes
468
Yeah, I get that.
But isn't this an impossible/freak result, given there is no scientific base for distinguishing between these 2 cables? Isn't that the crux of the argument that cables don't matter as long as they're implemented correctly technically?
 
OP
M

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
Well that's what I was trying to check. In the past I've believed I could hear differences between cables but many here tell me this must be proven to make any such claim. Hence the test. To my surprise actually we could reliably note a very small difference.

I made some measurements - and found fact the less preferred cable delivered about 0 - 0.15dB (frequency dependent) less than the preferred one. I measured the cable dc resistances and found the preferred cable had 1/4 the resistance of the other one. On calculation, the cable resistances closely account for the dB loss.

The speakers drop to about 3 ohms through the midrange but impedance is much higher at other frequencies - so the voltage drop is most evident in the midrange, but less elsewhere. I doubt that the cable impedance varies over the audio range, but the speaker impedance certainly does, so the effect of any cable with significant resistance will provide a non-linear frequency response - the degree depending on the ratio of the cable impedance to the speaker impedance. All this assuming the amplifier output impedance is essentially zero.

So - in short - we could reliably tell a very small difference between two cables, and the difference correlates with the cable resistance. More thorough testing and measurement might show more effects, but i expect lesser, and what we are dealing with already is on the limit of our assessment. The outcome is not what I expected from my earlier "listening experiences" - validating the claims of those around here who doubted/dismissed them. Oh well, case proven.

By the way we compared another pair of cables with less different resistance and couldn't pick one from another. For reference in all these tests one cable is the planar variety, the others are conventional flex.
 

Zoomer

Senior Member
Joined
Oct 5, 2019
Messages
323
Likes
468
Interesting.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
What would be interesting to test, since you still have the device, is to add a resistor of the difference in total cable resistance to the better conducting one. Then try again.

In this case you would have proven to yourself it is not the geometry, special 'properties' of the cable, etc but plain old resistance (gauge).

It is well known that cable resistance, certainly when combined with speakers that dip very low in impedance have measurable influence that when > 0.1dB could be detectable in AB tests.

This is not debated. What was debated is that 2 cables that are equal in resistance, yet differ in other properties have magical properties.
 
OP
M

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
Ok - fair enough - it is still set up - all I need to find is a couple of 60mohm resistors..... might have to think a bit to find or make those.,
M
 
OP
M

Murrayp

Member
Forum Donor
Joined
Oct 30, 2016
Messages
43
Likes
87
Location
Napier, New Zealand
Well that wasn't too hard - a length of lighting wire gave me 63mohm - near enough. With this duly placed in series with the low resistance cable to match the planar cable resistance - I couldn't tell any difference between the cables at all. I ran a quick sample measure and the voltage at the speakers is now the same with either cable, up to a few kHz at least - beyond that I'm not so sure as my meter (Fluke87) responds oddly to the Stereophile warble tones I'm using.

So in my simple experiment - cables of substantially different resistance, did make a detectable difference in listening and measurement with my low impedance speakers. When the resistance is adjusted to match - no difference was heard or measured..

So no mystery - now I'm happy! Boring though - no more lusting after flash speaker cables....
 

Speedskater

Major Contributor
Joined
Mar 5, 2016
Messages
1,610
Likes
1,333
Location
Cleveland, Ohio USA
beyond that I'm not so sure as my meter (Fluke87) responds oddly to the Stereophile warble tones I'm using.
Digital Multi-Meters (DMM) like your Fluke won't do well with warble tones. The meter samples several times a second. An old analog voltmeter will do better with warble tones, so will a SPL meter. Use sine waves with a DMM
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Kudos @Murrayp This is the sort of thing ASR should be all about.
 
Top Bottom