Is there any way to minimize bias in a blind test when you know which speakers are playing by hearing them?

[goldark]

I own quite a few speakers that I've listened to for years and am familiar with how they sound on a list of reference test tracks. Lately, I've been pondering trying to set up a blind test for myself as a way to pass the time when we should be staying indoors and socially distancing ourselves (wash your hands, people!)

Given that I can mostly tell which speakers are playing based on sound characteristics, is there a way to truly be non-biased in my selection of which sounds "best" to me? I suppose I can try new reference tracks, but even then, I think I'd be able to tell at least most of the time which ones are playing. This isn't even taking into consideration equalizing for bass extension differences, which is another issue I have to think about how to address.

Most of my speakers lean towards being neutral so it would be an interesting experiment for me to do. The lineup I'm thinking of using right now include: Revel F206, Infinity R152. Ascend Sierra 2-EX, Ascend CMT-340, Infinity IL30, Energy Connoisseur C3, (soon to be receiving) Salk WOW1, and DCM TP160 as a budget neutral reference.

Because of my experience with these speakers, would a blind test be automatically compromised if I'm generally familiar with audible differences among them before even testing? Or is there a way to salvage it?

 

[solderdude]

What exactly is the goal of the blind test ?

 

[goldark]

What exactly is the goal of the blind test ?

To rank my speakers in order of preference based entirely on sound quality - nothing too serious, just a little bit of fun and hope to gain some insight about my biases. I'm just afraid my biases having heard all of them might make the exercise futile.

 

[Soniclife]

Can you manage the blind part of the test? Be interesting to see if you can indeed tell them apart with your existing test tracks before going further.

 

[solderdude]

You will need to obscure the speakers from sight without covering them so you could actually use a blindfold so you cannot see the speaker.
Someone else must bring in speakers and connect them.
You should not have multiple speakers in the room at the same time.

Also turn down the volume to 0 each time the speakers are switched.

There are pitfalls but as this is for fun and not hard science I would say ... interesting experiment that would only say something about your pereference in your room.

 

[ctrl]

Because of my experience with these speakers, would a blind test be automatically compromised if I'm generally familiar with audible differences among them before even testing? Or is there a way to salvage it?

So I assume that this is a "real" blind test.

You could rate the speakers on everyday sounds and environmental noises - tram, church bells, leaves rustling, ... but also individually played musical instruments.

You can download many single music and percussion instruments here:
https://tech.ebu.ch/publications/sqamcd

This way you should find it very difficult to identify your speakers.

 

[Soniclife]

You should not have multiple speakers in the room at the same time.

Why?

 

[solderdude]

Because other speakers may damp resonances (act as a bass trap) or cause resonances when not connected to anything. Certainly when close together.

 

[goldark]

Can you manage the blind part of the test? Be interesting to see if you can indeed tell them apart with your existing test tracks before going further.

I'll do my best. I'll have some help changing which speakers are playing of course. Will also have to level match them.

 

[Soniclife]

Because other speakers may damp resonances (act as a bass trap) or cause resonances when not connected to anything. Certainly when close together.

Is there any evidence of the impact of this? Seems like it's going to be similar to adding any other bit of furniture to a room.

 

[solderdude]

Just put 2 different speakers next to each other and sweep and measure one with and without the other speaker and you'll know.

All cabinets + woofer, that is electrically not connected, (not damped) form a resonator.

 

[Wes]

w/o having someone else connect them - you could gin up a relay to switch and have it randomize the selections, as well as record which it is

 

[Soniclife]

Just put 2 different speakers next to each other and sweep and measure one with and without the other speaker and you'll know.

All cabinets + woofer, that is electrically not connected, (not damped) form a resonator.

If very close I can see an effect, but if a few feet away?

 

[Ron Texas]

My whole life feels like a blind test lately.

 

[Mnyb]

This would actually test your assumption that you indeed know speakers or maybe not ?

 

[goldark]

This would actually test your assumption that you indeed know speakers or maybe not ?

I want to think that I'm pretty familiar with how they sound, but that can't be known for sure without a blind test.

 

[Mnyb]

God luck and have fun doing it

 

[Blumlein 88]

If there is space, maybe create a lazy Susan for speakers. One where you could at least have two on it. Hook up things to both speakers, adjust for level differences. And your assistant could simply rotate between each speaker (or pretend to do so). See if you can identify each. I'm pretty certain you'll hear a difference, but will you correctly identify each? Play one channel at a time obviously. But this would let you rapidly switch back and forth without connecting/disconnecting speakers. At least initially just put up with the interference between having both speakers in the room at once.

BTW, pink noise filtered to have 500-2000 hz is suggested for level matching.

Method #2, if your room is symmetrical, place the speakers in mirror image positions. Play a test file which plays in only one channel, and then the other. You can have a variety of snippets of music. When satisfied identify the speaker. Or alternative method, describe how the two speakers differ from each other. You'll need an assistant to set up the speaker for each channel ahead of time, and you'll have to be lead blind folded into the room to the listening position. Again you'll have to live with any interference between speakers.

Method #3. Have an assistant setup one speaker. Lead you into the room blind folded. Listen to snippets of music. Then provide a description of the sound. Leave room, let assistant change speakers, and repeat. At least once without knowing have the assistant repeat an earlier speaker and see if your descriptions are consistent.

 

[richard12511]

I had this exact problem with a blind test I did with myself and a couple friends last year. I was able to pretty easily recognize the two speakers in the 1v1 blind that we did. The blind still worked for my friends, though.

My suggestion - and what I would do if I were to do it again - is to find away to introduce 2 or more new speakers into the blind that you've never heard before. Maybe use crutchfield for their $10 return shipping, or use a company that offers free returns. If you can setup a blind with say 5 speakers, 3 of which you've never heard, I think it will be much more difficult to identify the ones you know, especially if you listen to music you've never heard before.

 

[Chrispy]

Seems like taking several of your favorite foods and for some reason ranking them.....now the process of making and eating them is highly enjoyable, the ranking not so much.