[Chris J]

Quote:

Placebo effect. You can't really hook these up to a machine and analyze what is truly better, so people's biases get in the way. 
 
I'd say that many, if not most, audio equipment is highly highly overpriced. 

 
Sure you can, you put two pieces of equipment thru a diverse and comprehensive battery of tests and determine which one should perform better.
Just don't expect one set of tests and one machine to do all the testing you want.
I suspect a lot of these truly great audio companies do testing we are not even aware of.
 
For example, how does a power amp react to highly reactive loads? Very low impedance loads?
How does a power amp recover from clipping?

 

[Chris J]

Quote:

Placebo effect. You can't really hook these up to a machine and analyze what is truly better, so people's biases get in the way. 
 
I'd say that many, if not most, audio equipment is highly highly overpriced. 

 
I've give you this much, it is quite possible that we have not figured out how to measure everything.

 

[Steve Eddy]

Quote:

 
I've give you this much, it is quite possible that we have not figured out how to measure everything.

 
You don't necessarily have to measure anything. The null test for example. The beauty of it is you can implement it with any set-up and can even use music signals for your source.
 
se

 

[kiteki]

 
You don't necessarily have to measure anything. The null test for example. The beauty of it is you can implement it with any set-up and can even use music signals for your source.

 
Do you mean recording the same music twice, switching something in the chain the second time, compile, subtract, and listen to the difference?
 
I suppose that will work in a lot of cases.

 

[Steve Eddy]

Quote:

 
Do you mean recording the same music twice, switching something in the chain the second time, compile, subtract, and listen to the difference?
 

 
Well, you don't have to go through all that. You can null the outputs of two different devices running at the same time.
 
se

 

[kiteki]

It wouldn't work for finding the sound quality of an A/DC converter though.

 

[billybob_jcv]

Quote:

 
Well, you don't have to go through all that. You can null the outputs of two different devices running at the same time.
 
se

 
 
So - this provides the relative measurement of two devices that are both physically in your possession?  Would you need a universal calibrated reference standard that would be used by all reviewers?

 

[Steve Eddy]

Quote:

It wouldn't work for finding the sound quality of an A/DC converter though.

 
The null test has nothing to do with assessing "sound quality." The null test is all about differences between two devices. If the residual output of the null is not audible, then there is ipso facto no audible difference between the two devices.
 
se

 

[Steve Eddy]

Quote:

So - this provides the relative measurement of two devices that are both physically in your possession? 

 
Yes. The null's output is whatever difference there is between the outputs of the two devices.
 
Quote:

Would you need a universal calibrated reference standard that would be used by all reviewers?

 
Effectively, yes.
 
se

 

[kiteki]

It wouldn't work for finding the sound quality of an A/DC converter though.

 
The null test has nothing to do with assessing "sound quality." The null test is all about differences between two devices. If the residual output of the null is not audible, then there is ipso facto no audible difference between the two devices.

 
It won't work for finding the difference between two A/DC's either, or transducers for that matter.  If a transducer is a TV screen, let's say OLED versus LCD, you record both, and play them on an LCD, that doesn't work, same principle applies to the A/DC section.

 

[Steve Eddy]

Quote:

Wouldn't you would be back at the same issue:  is the difference audible and is the difference good, or bad?

 
Well, if you do the null on two components and you can't hear the residual output, where else is there to go?
 
se

 

[Steve Eddy]

Quote:

 
It won't work for finding the difference between two A/DC's either, or transducers for that matter.  If a transducer is a TV screen, let's say OLED versus LCD, you record both, and play them on an LCD, that doesn't work, same principle applies to the A/DC section.

 I don't see how your OLED and LCD example applies to ADC's.
 
se

 

[billybob_jcv]

Subtracting two signals does not tell you which is "better", only that there is a difference from your standard.  You can't then "listen" to the delta signal - to do that you would have to feed the delta signal through additional components, thus changing the result, and you would not be able to tell if the delta signal is discernable in the presence of the original signal.  I must be missing something, I'll go read-up on "null testing" of audio gear.

 

[kiteki]

He's saying null testing is only to prove a difference, i.e. if the result is pure silence, there was no difference to start with, so it's a scam. =]

 

[billybob_jcv]

OK, after reading this thread:
http://www.diyaudio.com/forums/everything-else/13415-null-difference-testing.html
 
I think it's exactly what I assumed it was - an attempt to subtract two signals and then examine the result.  The thread does what all these threads do - it degenerates into the objective vs subjective argument.  Note the discussion of the Bob Carver vs Stereophile challenge.  This makes perfect sense to me.  If you can modify your signal to provide a -70dB null against the reference, then you have effectively modeled your reference and you won't hear a difference - assuming that the rest of your metrics are also close enough to not matter.  I think what that shows is what we all know - that the frequency response is the major contributor to the sound we hear.  If you match that to a good reference, and don't completely f-up anything else, you will have yourself a nice system.