• 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!

Can someone give me a link of cable blind test research papers?

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
We have tests on this site which show that competently designed DACs sound the same but that there are many DACs, including some (previously) well thought of named DACs, on the market which apparently are not competently designed and they sound not only different but worse.

There are at least a couple of cable tests on this site:

https://www.audiosciencereview.com/...-high-end-usb-cables-make-a-difference.11272/

https://www.audiosciencereview.com/...sis-plus-oval-digital-xlr-cable-review.11271/

Like DACs it is surely possible that cables sound different because of the competency of their design so I would suggest it is still necessary to test cables to find the good ones.
If you look in the database we have all these categories:

1584704383028.png


I've attached the paper @solderdude referenced. It's a strange one that leaves me with more questions than answers. No reference to blind testing.

As long as we're going down this route I would recommend the equally strange chapter on speaker cables in (2019) Philip Newell & Keath Holland - Loudspeakers for Music Recording and Reproduction, 2nd ed., Chapter 6 "Amplifiers and Loudspeaker Cables - A General Review". I think there is some reference to blind testing, but only in an off-hand way.

Philip Giddings' Audio Systems Design and Installation is a much more serious book on electrical design and has information on cables throughout, including a dedicated chapter somewhere in the middle of the book (I don't have it on hand). No reference to blind testing.
 

Attachments

  • (1991) Fred Davis - Effects of Cable, Loudspeaker, and Amplifier Interactions.pdf
    6.9 MB · Views: 986
Last edited:

database

Member
Joined
Jan 2, 2020
Messages
53
Likes
54
Location
VA
Burden of proof is on claimants, always remember that. If someone tells me that they have a machine that runs at 110% efficiency, it's not up to me to disprove it, it's up to them to prove it.

I agree, but I conducted (single) blind tests on an HDMI cable and found 3 skeptical users passed with a 100% score. I don't have the means to measure the cable itself, but sine wave sweeps in REW didn't show any measured differences coming out of the speakers themselves.

Of course, no one (besides me and the test participants) accepts my results, and they don't really have to. A few unsubstantiated tests by one user doesn't prove anything, maybe the tests I performed were flawed somehow. But then how should I proceed from here? In the end, people believe what they want, so some people dismiss my result simply because the conclusion doesn't agree with their beliefs, but there's no science in that line of thinking.

I have read before that measurements of sine waves can't capture differences you would hear in music, because sine waves are much simpler than the wave forms of music. Sure, that claim might just be marketing from a company selling snake oil. But one instance where I was reminded of that claim was when I learned how humans perceive absolute polarity -- http://archimago.blogspot.com/2019/06/musings-and-listening-on-absolute.html . For some reason, for certain wave forms only, the wave form sounds louder to the ears with one polarity compared to the inverted polarity. This doesn't happen with sine waves, and no looking at any wave forms compared to their inverted counterparts would ever predict this. But it does happen with specific wave forms included in the link, as well as certain types of music. This is what made me start to feel that it could be hasty to make sweeping conclusions based on measurements of sine wave sweeps alone.
 

BDWoody

Chief Cat Herder
Moderator
Forum Donor
Joined
Jan 9, 2019
Messages
6,949
Likes
22,628
Location
Mid-Atlantic, USA. (Maryland)
maybe the tests I performed were flawed somehow. But then how should I proceed from here?

They were indeed flawed.

You should proceed by reading about clever Hans, then learning how to do the test properly, then try to repeat your results in a way that gives it any validity.

Then, once you've controlled for everything BUT secret woo, and can still pick out the differences, you will be taken more seriously as someone who is willing to go through that minimum standard, than someone who makes whatever version of "but even my wife in the kitchen" could hear the difference types.

I would welcome your results! But not some well meaning but completely flawed approach to the comparison.

You asked...
 
Last edited:

database

Member
Joined
Jan 2, 2020
Messages
53
Likes
54
Location
VA
They were indeed flawed.

You should proceed by reading about clever Hans, then learning how to do the test properly, then try to repeat your results in a way that gives it any validity.

Then, once you've controlled for everything BUT secret woo, and can still pick out the differences, you will be taken more seriously as someone who is willing to go through that minimum standard, than someone who makes whatever version of "but even my wife in the kitchen" could hear the difference types.

I would welcome your results! But not some well meaning but completely flawed approach to the comparison.

You asked...

Thanks. I see how the Clever Hans effect can invalidate the results of a single blind test.

In my test, I had a computer with two HDMI outs going to two HDMI inputs on a receiver, with one HDMI cable on each. I gave each test participant a chance to listen to the test track on each cable sighted, to learn how the test track sounds on each.

Then I moved to another room, blindfolded the participant, virtually flipped 10 coins and recorded the results as the correct answers. For each trial,
I manually selected the proper HDMI audio device according to the coin flip in foobar, selected the correct receiver input, played the first 15-30 seconds of a test track in foobar, and had the test participant state which cable they thought was being used, and recorded their answer. I didn't say anything in response to their answer. In between every trial, I used a clock to wait around 30 seconds before starting the next trial. After 10 trials, I revealed the results.

If you have any suggestions on how to improve the test method, feel free to let me know. The biggest weakness is obviously that I had to manually conduct the test while knowing the answers. I don't see how to avoid this without writing some specialized software, which doesn't seem to be worth the effort if people will dismiss the test results anyway.
 

Wes

Major Contributor
Forum Donor
Joined
Dec 5, 2019
Messages
3,843
Likes
3,788
double blind the test so the participant cannot see or hear the person who knows the cable ID

be sure to check that SPL is not changed by a different cable somehow
 

BDWoody

Chief Cat Herder
Moderator
Forum Donor
Joined
Jan 9, 2019
Messages
6,949
Likes
22,628
Location
Mid-Atlantic, USA. (Maryland)
, I used a clock to wait around 30 seconds before starting the next trial. After 10 trials, I revealed the results.

Switching also needs to be almost instantaneous, with possibly shorter samples. 30 seconds I believe is on the long side...

As @Wes said, check output levels to be sure they are matched...

I'm sure there will be more...
 

database

Member
Joined
Jan 2, 2020
Messages
53
Likes
54
Location
VA
double blind the test so the participant cannot see or hear the person who knows the cable ID

Yep, I (the tester) was in another room.

be sure to check that SPL is not changed by a different cable somehow

As @Wes said, check output levels to be sure they are matched...

As I mentioned, I took REW sweeps (nearfield) and they were identical. Is this sufficient for level matching, or do I need to measure voltage instead?

Switching also needs to be almost instantaneous, with possibly shorter samples. 30 seconds I believe is on the long side...

This is another challenge. Since I was switching manually, 30 seconds was the fastest I was able to go. Do you know why instant switching is required between trials? I know auditory memory is limited to a few seconds or so. But Foobar ABX tests let you take as much or little time as you want between trials, and they also let you select to listen to A and B at any time during any trial. My test was much harder, since users had no opportunity to listen to A and B ever again except before starting the test; I played them a sample once per trial and they were required to give me their answer immediately. To me, the fact that they were able to pass a harder version of the foobar ABX test with a 100% score suggests that the difference was large enough that they knew what to listen for (which is what they told me after completing the test). Perhaps that logic is flawed somehow.

I'm sure there will be more...

If someone is able to provide a source with a comprehensive list of requirements, I would consider retesting if we're able to come up with an acceptable testing method. I'd prefer not to go through the effort of retesting after addressing 2-3 issues only for that new test to get thrown out with several more issues identified. I appreciate the feedback you've provided so far.
 

Wes

Major Contributor
Forum Donor
Joined
Dec 5, 2019
Messages
3,843
Likes
3,788
the tester was in another room all the time??

who swapped the cables?
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
If the differences are that audible they should be very easy to record.

So you could redo the test (2 different HDMI cables) and record the analog output on 24/192 for the 2 cables. Preferably the exact same music excerpt where you heard the differences most obviously. Easy to do and could be analyzed as well as listened to.
 

database

Member
Joined
Jan 2, 2020
Messages
53
Likes
54
Location
VA
the tester was in another room all the time??

who swapped the cables?

Both cables were plugged in at the same time into two different HDMI outs on the PC and HDMI ins on the receiver. I just remoted into the PC to switch output devices and receiver inputs.

If the differences are that audible they should be very easy to record.

So you could redo the test (2 different HDMI cables) and record the analog output on 24/192 for the 2 cables. Preferably the exact same music excerpt where you heard the differences most obviously. Easy to do and could be analyzed as well as listened to.

Thanks. That's actually the first thing I tried, and the recordings sounded the same to me in this test (but switching cables while playing the same recording sounded audible... :facepalm:). I don't think I was careful enough to record at 24/192 though, I will check that and see if redoing in 24/192 makes any difference.

I proceeded to test with my flawed single blind manual test to see if the recording method was missing something. Maybe my test was too flawed to make any conclusions, but if the recording method still demonstrates no differences at 24/192, I would like to perfect my manual test to compare results against the recording method.
 

Chrispy

Master Contributor
Forum Donor
Joined
Feb 7, 2020
Messages
7,757
Likes
5,918
Location
PNW
Both cables were plugged in at the same time into two different HDMI outs on the PC and HDMI ins on the receiver. I just remoted into the PC to switch output devices and receiver inputs.



Thanks. That's actually the first thing I tried, and the recordings sounded the same to me in this test (but switching cables while playing the same recording sounded audible... :facepalm:). I don't think I was careful enough to record at 24/192 though, I will check that and see if redoing in 24/192 makes any difference.

I proceeded to test with my flawed single blind manual test to see if the recording method was missing something. Maybe my test was too flawed to make any conclusions, but if the recording method still demonstrates no differences at 24/192, I would like to perfect my manual test to compare results against the recording method.

I'd start with a test of the levels involved. Most likely source of any audible difference IMO.
 

SimpleTheater

Addicted to Fun and Learning
Forum Donor
Joined
Jun 6, 2019
Messages
927
Likes
1,789
Location
Woodstock, NY
Page 10 of Stephen Lampen’s “Audio/Video Cable Installers Pocket Guide” has a story about a well-known speaker manufacturer assembling a group of hi-end aficionados to listen to high end speaker cables. The group chose the source material, the cables were laid out in plain sight, the amplification and the speakers. They started the testing with 12 gauge “zip” cord for base reference. The techs would then get each cable and install them. This was obviously a sighted test and while some members disagreed about which high-end cable was the best, they all agreed all the high end cables sounded significantly better than the base zip cord. The speaker manufacturer let them leave without ever telling them his techs only faked changing the cables and they were listening to zip cord the entire time.
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Page 10 of Stephen Lampen’s “Audio/Video Cable Installers Pocket Guide” has a story about a well-known speaker manufacturer assembling a group of hi-end aficionados to listen to high end speaker cables. The group chose the source material, the cables were laid out in plain sight, the amplification and the speakers. They started the testing with 12 gauge “zip” cord for base reference. The techs would then get each cable and install them. This was obviously a sighted test and while some members disagreed about which high-end cable was the best, they all agreed all the high end cables sounded significantly better than the base zip cord. The speaker manufacturer let them leave without ever telling them his techs only faked changing the cables and they were listening to zip cord the entire time.

I did a very similar test when installing a new (expensive) speaker cable on a high-end system (worked in a Hifi shop back then).
The guy had a great sounding system and knew it inside out.

Made it an experiment for him. Let him listen to his 'old' cable. Then pretended to swap the cable for the new one. He heard obvious differences (there were none, I changed nothing). Then swapped the cable for real and told him that as an experiment I put the old one back.
He found it sounded bad (new cable in reality) and asked me to quickly put in the new cable. I pretended to do that (again did nothing) and asked him how it sounded now. He found it sounded great (still new cable, no change). Still have his 'old' and inferior cables somewhere.

No more research papers were needed for me ...
 
Top Bottom