Help setting up blind test

[Dimitrov]

I need some advice. I want to test the effectiveness of an external power amp connected to an AVR.

AVR is a Denon AVR-3500H. Using the preouts on the Denon, Emotiva XPA-2 (300W into 8 ohms).

How would I perform a blind test between the two? Is it even possible?

 

[solderdude]

First you would need a switch-box that can switch speakers quickly and is break before make.
Then you would need to level match within 1% of the output voltage (under load) using a 250 Hz tone or something.
Then you would need someone else to help you operate a switch and keep logs.
Then you would need to make statistically enough switchings.
You would also have to figure out a way to test which you are comfortable with (you deciding the switch moment or the tester) and how to solve the problem of not switching but giving a small mute so you can't tell if switched or not.

Anything but easy to do ... properly.

 

[Dimitrov]

Great, was hoping to do a test and you've put me off.

I guess it would be easier if I had two identical AVR's or close enough models, one with a power amp attached and one without, then I could get someone to switch for me. Matching levels shouldn't be too difficult with a multimeter.

 

[solderdude]

My intend was not to deter but go about it in the right way which involves making or buying a switch box that is suited for switching speakers.
I would even recommend to use changeover relays and give the used amplifier a load when it is not connected (its what I built about 30 years ago)
My reply was just to show what is involved in something as simple as comparing 2 power amps.

When one tests blind one should do it properly with the right controls.

Not getting one of the controls right (within tolerances) means you may get 'skewed results' that make it appear you had a valid test and could.
discern differences reliably but was actually coming from one of the controls being incorrect.
Might as well test sighted in that case which will give very similar results.

From experience I can tell you that a properly conducted blind (DBT not really needed unless lives depend on it) is pretty hard to do properly and very educational when done properly, which is not really easy or quick to do.
Also one should train themselves with the many audibility tests around to find out what is audible or not.

The only things that can relatively easy be tested blind (properly) are interlinks and speaker cables but is very time consuming.

 

[Dimitrov]

My intend was not to deter but go about it in the right way which involves making or buying a switch box that is suited for switching speakers.
I would even recommend to use changeover relays and give the used amplifier a load when it is not connected (its what I built about 30 years ago)
My reply was just to show what is involved in something as simple as comparing 2 power amps.

When one tests blind one should do it properly with the right controls.

Not getting one of the controls right (within tolerances) means you may get 'skewed results' that make it appear you had a valid test and could.
discern differences reliably but was actually coming from one of the controls being incorrect.
Might as well test sighted in that case which will give very similar results.

From experience I can tell you that a properly conducted blind (DBT not really needed unless lives depend on it) is pretty hard to do properly and very educational when done properly, which is not really easy or quick to do.
Also one should train themselves with the many audibility tests around to find out what is audible or not.

The only things that can relatively easy be tested blind (properly) are interlinks and speaker cables but is very time consuming.

Thanks solderdude! Appreciate your response. I keep hearing people saying adding power amps make big differences in terms of sound quality and "control" - more definition etc.

So I wanted to test the difference it makes .... have both AVR and power amp play at the exact same volume - at a comfortable listening level but must be identical for both and hear (if any) difference it makes.

Perhaps a blind test isn't totally necessary here.

 

[solderdude]

I think the reported differences could be caused for several reasons.
There could be technical reasons or not so technical ones.
In all cases people can and will report improvements.

In your case I would say just connect a good power amp and if it works for you just use it.
When you can adjust the gain of power amp and have a decent AC multimeter (with 2VAC setting at least) and a 100Hz test signal one could level match. It should be within 1% if you want to find 'real evidence'.
This is not easy to do.

Only for absolute proof finding (or to further ones own knowledge as it were) I would recommend well conducted blind tests.

Making the test blind involves another person that swaps cables and you not knowing what is connected and at least 10 (preferred 20) tries in which the tester either swaps cables or not.

 

[Dimitrov]

When you can adjust the gain of power amp and have a decent AC multimeter (with 2VAC setting at least) and a 100Hz test signal one could level match. It should be within 1% if you want to find 'real evidence'.
This is not easy to do.

Are you saying level matching is not easy to do? Are there any multimeters you would recommend that would be good enough to measure within 1% without breaking the bank balance?

 

[solderdude]

Are you saying level matching is not easy to do? Are there any multimeters you would recommend that would be good enough to measure within 1% without breaking the bank balance?

Level matching within 0.1dB is not easy to do. 1% is very little.
The difference between 2.300V and 2.323V

Any multimeter with a 2VAC setting (or auto ranging with that range) will do. It does not even have to be calibrated.
a meter with a 2VDC setting may not have a 2VAC setting.

All of these meters will measure AC up to 400Hz accurately and even when 1kHz is used really doesn't matter as long as it displays something (but 1kHz playing loud through speakers is no fun).
In this case it is not about an absolute value but a relative difference between 2 values.
Check both L and R channels when using amps with volpots. L-R differences may be well over 0.1dB and may differ from volpot settings as well.

Use volume control before the signal is split to 2 amps to avoid false positives due to potmeter differences.