[konut]

Is it not possible to play a 10 second file of music, record the file,  then overlay each file for cancellation?

[JohnR]

Yes, that would show that a particular treatment made no difference - bearing in mind that acoustic measurements have a much lower signal-to-noise ratio than electronic ones.

So perhaps it would be useful for some things

[neekomax]

Here's a guess as to why not.

If you record to a DAW from the LP before and after a certain treatment(s) is/are added, reverse the phase of one of the waveforms, then sum them to find the differences, you would probably find that the resulting sound artifacts (the difference) are significant.

Problem is, was the mic in a slightly different position? Was the volume level of one of the tests slightly different for some reason? Were you standing in a slightly different place relative to the mic or the speakers, changing the acoustics slightly?

There are probably more variables that that, but my point is that for measuring the overall acoustics of an entire room, seems like a null test would be VERY sensitive, and it would be hard to be sure that all other variables besides acoustic treatment were unchanged.

Or maybe not. Maybe with care it could yield some interesting and telling data. Would be cool to hear exactly what a certain treatment eliminates  .

Just thinking out loud, so to speak  .

[JohnR]

Yes, they'd have to be done right after each other, useful for a single item but not a complete set of treatments. Opening doors and windows will affect the measurement too.

Good point about you being in the same position.

[konut]

I would agree that it would require stringent protocols to be maintained to obtain repeatable and accurate results. The people that make treatments charge a fair amount of cash. This should be doable.

[InfernoSTi]

Interesting conversation.  My question is this:

Assume that you could do two identical measurements and only change one variable such as changing a single bass trap from one to another.  And suppose you could measure a difference.  Would you be able derive from the data which sounds better? Or would you merely be in a position to say this is this and that is that?

I would argue you would have to then listen to the two, write down qualitative descriptions, and assign these subjective descriptions to the two measurements.  This is because the objective with music is beauty (in the ear of the listener), a subjective goal.  If it were purely objective, we would continuously be narrowing our audio equipment choices as every manufacturer and DIY'er zeroed closer and closer to the purely objective target.  Yet as we gain more and more measurement capabilities, our choices are increasing and more diverse than ever before.  This is truly a renaissance of audio with greater choices and knowledge than ever before.

I believe this is what others are trying to say with their comments regarding famous concert halls that each sound different but each sound "right" as well.  It is the subjective beauty that is relevant, not the outcome of the measurement.

Relevance is important. If you can't state a hypothesis that results in relevance, then there is no logical reason to pursue the test.  Every test should start with the idea that it will prove something meaningful, either in the positive or the negative (which is easier: here is a measurable difference and nobody could blind A/B tell the difference in sound quality or half preferred A and half preferred B).   What do we learn if we develop a test and we are successful in conducting the test: that is the question at hand. 

Imagine I came up with a measurement test that measured the color of my amplifier and it was one hundred percent accurate.  Would that be a relevant test to determine the sound quality of the amplifier?  No, even if the test is accurate.  So I would dismiss developing such a test without seeking a way to do it accurately because the logic doesn't support it being relevant.

Now if all you are saying is that you want to measure the difference so you can begin to assign subjective sound qualities to the various results, then is it any different that simply listening to each system and picking the one you like better?  Why test with equipment something that is only relevant if you can hear and prefer one over the other?  Keep it simple instead!

There, I'm done with my rant....   

John

[neekomax]

Inferno John, points well taken. A test like that would not be able to show qualitative differences before and after treaments were applied. Although it would be interesting to be able to actually HEAR the bass being trapped by a bass trap , which is what I assume would result, for example.

All that would be shown is the relative quantity of change in the sound of the room, I believe. It would be a subjective call to decide if the resulting sound were 'better'. The null test merely proves differences, not desirability of said differences.

BTW, I pinged Ethan Winer to see if he would like to comment here. I assume that's cool. 

[JohnR]

Assume that you could do two identical measurements and only change one variable such as changing a single bass trap from one to another.  And suppose you could measure a difference.  Would you be able derive from the data which sounds better?

I'd say so, in as far as the question makes sense. You wouldn't be attempting to measure which bass trap "sounds better," but to ascertain properties that will help to figure out how to get to your goal.

[konut]

Thank you neekomax and InfernoSTi.  I would imagine that, depending on the test, different conclusions could be drawn. For example, if a test tone(s) is used, relative effectiveness could be plotted to determine the appropriateness of a given treatment for the amount of reflection, absorption, diffusion, trapping, etc. desired or required. If music is used, it might be possible to correlate preferred outcomes in the same way that different codecs are used to compress audio files. IOW, data generated could be used to accurately predict the most pleasing result for most listeners.
         In order to get baseline and repeatable data, the protocols would have to be standardized so that treatments could be compared between manufacturers in an anechoic chamber. It would then be possible to determine the best solution for a particular situation. 
         Exactly right JohnR.

[Ethan Winer]

how does one go about conducting, what seems to be the gold standard, a null test?

You're already getting good advice, but I can add a bit. Especially since I see my name. 

A null test is not appropriate for measuring the improvement after adding acoustic treatment. A waterfall plot is the best way to assess the improvement after adding bass traps. For mid and high frequencies I'd use an RT60 measurement as well as frequency response. Both of these are features in popular software such as Room EQ Wizard, FuzzMeasure, and ETF.

I'll also mention that it's critical for the measuring microphone (or SPL meter) to be in the exact same location for Before / After tests. Not just in the ballpark, but ideally within 1/4 inch in all directions. I mention this because such tests are usually done hours or days apart.

--Ethan

[konut]

You're already getting good advice, but I can add a bit. Especially since I see my name. 

A null test is not appropriate for measuring the improvement after adding acoustic treatment. 

I had a feeling you were going to say that.

[Rclark]

Not to hijack, but to hijack: Ethan, if you have a quick moment, would you please weigh in here??

http://www.audiocircle.com/index.php?topic=98771.0

 I plan on doing treatments as the next stage, or at least get things going. Neekomax I am also looking at your diffuser thread.


 ... back to testing thread 

[bpape]

Is it not possible to play a 10 second file of music, record the file,  then overlay each file for cancellation?

From a frequency response perspective only, yes.

From a time perspective, how would you measure decay over time when you have a 10 second clip where the stimulus is constantly changing and being added...

Also, how would one see changes in imaging from overlaying clips?  One of those things that cannot be measured.  Channel separation is not sufficient nor would it be measured in this type of test anyway.

Bryan

[neekomax]

Not to hijack, but to hijack: Ethan, if you have a quick moment, would you please weigh in here??

http://www.audiocircle.com/index.php?topic=98771.0

 I plan on doing treatments as the next stage, or at least get things going. Neekomax I am also looking at your diffuser thread.


 ... back to testing thread 

Build a skyline diffuser! 

[konut]

From a frequency response perspective only, yes.

From a time perspective, how would you measure decay over time when you have a 10 second clip where the stimulus is constantly changing and being added...

Also, how would one see changes in imaging from overlaying clips?  One of those things that cannot be measured.  Channel separation is not sufficient nor would it be measured in this type of test anyway.

Bryan

If there were differences in decay, they would be contained in the null result as artifacts. The trick would be to be able to separate the frequency differences from the decay differences.
        Lets keep it simple and leave imaging and separation for a different acoustic discussion.

[bpape]

At the longer end of the decay, it would be masked by the stronger current stimuli.

My point on the imaging was more of a general one that one cannot FULLY glean the benefits of a particular acoustic treatment in a specific place in the room (a whole 'nother area we've not even touched on yet....) purely by a null test.  Imaging and separation are part of what they can bring to the table.

What if one purely added diffusion into the room where there would be minimal change in frequency response in many cases?  How would one deal with that in a null test using music over time?

Not trying to be difficult - just bringing the whole real world problems into the discussion.

Even test tones will differ.  If you run a sweep from say 20-300Hz and then run another from 20-say 5kHz and window it to 20-300, your results will be different.

Bryan

[konut]

Never said it was going to be easy. Only, referring to my initial post, asking if it had been done, and if not, why not. To assert that it is "not appropriate" lacks scientific specificity and rigor. How do we know it is "not appropriate"? The only way to find out is to do it.

[Ethan Winer]

Ethan, if you have a quick moment, would you please weigh in here??

It's easier if I reply here, so I don't forget to look in a section I don't usually visit for follow-ups. In fact, PM me if a further reply is needed. I'm hardly visiting any forums these days as I work on my book to meet the publisher's deadline.

Anyway, you definitely want to face the top wall with the window as you sit, with the speakers firing toward the bottom angled wall. That keeps the front of the room symmetrical, which is what matters most. If you can leave the door open, that adds bass trapping in that corner. Though you'll of course need "real" treatment too.

--Ethan

[Rob Babcock]

Nice to see you still pop in occasionally, Ethan!  Good luck on the book.

[Rclark]

It's easier if I reply here, so I don't forget to look in a section I don't usually visit for follow-ups. In fact, PM me if a further reply is needed. I'm hardly visiting any forums these days as I work on my book to meet the publisher's deadline.

Anyway, you definitely want to face the top wall with the window as you sit, with the speakers firing toward the bottom angled wall. That keeps the front of the room symmetrical, which is what matters most. If you can leave the door open, that adds bass trapping in that corner. Though you'll of course need "real" treatment too.

--Ethan

 Thank you Ethan, that was very useful. And I appreciate the invitation to ask you further questions later. And please do come through the forums more often.

 If you need an editor, hit me up. Haha.