psycho acoustics: the nightmare of listening tests

**JohnSwenson** · 2012-02-07, 17:28

I have no problem at all with the concept of doing a DBT as long as the protocol allows for some time for relaxed listening.

Here is a proposal for a DBT of the infamous TT3.0. This one is ideal because it's software only, no switching between hardware needed which should simplify the testing procedure.

A procedure gets written which can switch between TT3.0 and "standard" by loging into the Touch from a remote computer. (this should not be too hard to do) Three scripts are written which can be run from a remote computer which switches to "A", another which switches to "B" and another that does an "X" which may be either A or B. The sequence of X configurations is determined by an excrypted file. The person under test does not know the sequence. Another program generates the encrypted files which are emailed to the participants. At this point no human knows the sequences in any of these files. After the test is complete the same program will generate the human readable sequence in each file.

During the test itself the participants will have the Touch hidden from view (it's possible to tell which is which if you can see the screen durring the reboot). All interactions with the Touch will be done through one of the remote options (iPeng, Android app, server web page etc), none of these have anyway to tell if TT3.0 is active or not. The person under test can use the three scripts to select A or B or X as many times as they want, listen to as many songs as they want, take as much time as they want, do it in one day, do it in a week, whatever.

Some things for discussion, how many X changes should be done? Should the identity of A and B be known by the testers? The encrypted file could also be used to keep the identity of A and B secret.

Possible ways to cheat: look at the Touch when rebooting. For this to work we have to trust people not to do this. Logging on using SSH to see if TT3.0 is urned on. Again we have to trust people on that one. The sounds made durring boot could also be used to tell which is which, but these can be globally turned off. We can just have the scripts do that.

Any takers? Any discussion? Does this satisfy the DBT "rules"?

John S.

**pski** · 2012-02-07, 17:36

Originally Posted by JohnSwenson

I have no problem at all with the concept of doing a DBT as long as the protocol allows for some time for relaxed listening.

Here is a proposal for a DBT of the infamous TT3.0. This one is ideal because it's software only, no switching between hardware needed which should simplify the testing procedure.

A procedure gets written which can switch between TT3.0 and "standard" by loging into the Touch from a remote computer. (this should not be too hard to do) Three scripts are written which can be run from a remote computer which switches to "A", another which switches to "B" and another that does an "X" which may be either A or B. The sequence of X configurations is determined by an excrypted file. The person under test does not know the sequence. Another program generates the encrypted files which are emailed to the participants. At this point no human knows the sequences in any of these files. After the test is complete the same program will generate the human readable sequence in each file.

During the test itself the participants will have the Touch hidden from view (it's possible to tell which is which if you can see the screen durring the reboot). All interactions with the Touch will be done through one of the remote options (iPeng, Android app, server web page etc), none of these have anyway to tell if TT3.0 is active or not. The person under test can use the three scripts to select A or B or X as many times as they want, listen to as many songs as they want, take as much time as they want, do it in one day, do it in a week, whatever.

Some things for discussion, how many X changes should be done? Should the identity of A and B be known by the testers? The encrypted file could also be used to keep the identity of A and B secret.

Possible ways to cheat: look at the Touch when rebooting. For this to work we have to trust people not to do this. Logging on using SSH to see if TT3.0 is urned on. Again we have to trust people on that one. The sounds made durring boot could also be used to tell which is which, but these can be globally turned off. We can just have the scripts do that.

Any takers? Any discussion? Does this satisfy the DBT "rules"?

John S.

Jeez man. Have you considered real audio measurement ?

**magiccarpetride** · 2012-02-07, 17:38

Originally Posted by pski

I'm so glad you are here to help. This is why most of us pay attention and reply here.

Have you considered getting a dog ? They are slavishly attentive and have been shown to help with certain psychological issues.

Already have a dog. Next...

**magiccarpetride** · 2012-02-07, 17:40

Originally Posted by pski

Jeez man. Have you considered real audio measurement ?

The issue with me is that I can't even hear any difference between analog outs on my Duet and a TT3.0 SBT into an outboard DAC.

**pski** · 2012-02-07, 17:43

Originally Posted by magiccarpetride

The issue with me is that I can't even hear any difference between analog outs on my Duet and a TT3.0 SBT into an outboard DAC.

What model?

**magiccarpetride** · 2012-02-07, 17:48

Originally Posted by pski

What model?

Logitech.

**NoRoDa** · 2012-02-24, 10:55

Originally Posted by TheOctavist

see the tt 3.0 thread.

did loads of dbt today. no difference with or without, stock psu or linear.

i jumped on tt bandwagon hastily. it provided no benefit..so back to stock for me.

How did you manage a double blind test with stock PSU and linear?
Had to be difficult to arrange?

REGArds

**TheOctavist** · 2012-02-25, 02:16

Originally Posted by NoRoDa

How did you manage a double blind test with stock PSU and linear?
Had to be difficult to arrange?

REGArds

ill post the details in a bit!

**Mnyb** · 2012-02-25, 02:54

I remember a user named iPhone doing ADM measurements on the analog out of the old SB3 using stock or linear psu and the output signal was identical in either case.

No one done that for Touch but I suspect similar results .

Some thinker suspect that audibility of PSU is down to it's interaction with the rest of your hifi if your switcher is disturbing amps or DAC for example ?

So it's hard to get universal results here, add many cult designers affection for design principles that makes equipment more susceptible to these kinds of problems that would not bother normal equipment ?

OT:
I think modding here and there does not accomplish as much, the device must be correctly designed from the ground up improve one detail and you face the next bottleneck,

Hence the external DAC route

I have an 3rd party PSU I bought to my SB3 it fits the Touch too , SB3 in Eu came with a really shoddy PSU , I suspected it did not even conformed to standards ? The late logitech SB3 and Touch comes with an much better switcher ( if you can believe that ).

**jvanhambelgium** · 2013-01-25, 12:58

Hi there,

I do believe there are some very minor improvements to be made by changing for example the "stock" PSU adapter of the Touch with a better one ;-)
We tested last week actually.

A Touch with "stock" power and a ultra-lineair,battery driven solution.

Together (3 of us) could hear a very subtile improvement, and not in terms of "more detail" or "brighter" or so.
No, when playing complex songs, when all instruments start banging together, there was a pretty clear improvement on instrument separation and less "muddiness" at these passages.

Songs that have for example female vocals (eg. Eva Cassidy) did not sound different or more detailed as-such. Only when a lot of stuff is coming together we noticed a more clear "control" , "separation" when using the linear PSU.

The Touch is a pretty fine thing as a streamer/transport! And for a very affordable price!

Thread: psycho acoustics: the nightmare of listening tests

Thread Tools

Search Thread

Display

You claim Magnepans

Posting Permissions