Help with my University Research Project

Thanks for the input so far. The way I plan to do the experiment is as follows:

First I will get some musicians from the conservatory to perform a 3 or 4 minute work. I will record this at 24/192 and use it as the ref. I will then play it back through various configurations of cables/components and monitor it with what is basically a modified Spectral analyzer. I will then try to see if a pattern becomes apparent. I will not be considering speakers as that would make everything too complex (I would have to worry about placement and room acoustics etc)

I will be playing the file from the computer - out to the components/cables - and back into the same computer. After I have become aware of each componets/cables sonic characteristics (assuming there will be some...) I will then conduct a double blind study with some of the local "golden ears" as well as various students. I will use the same speakers but change the components or cables hopefully reproducing the effect I recorded in the lab.

Then we can be on the way to creating a rating system so that people can match their components properly... then the industry will quickly adopt our system and the world will be perfect...

I hope everyone picks up on the sarcasm in that last paragraph "because I was laying it on pretty thick" - David Spade

I will be talking to local audio shops soon. What I was thinking was starting with is the middle of the MIT line and comparing it to the middle of the Kimber line. Also, I plan to start with ConradJohnson and Rotel. All of these companies are well represented at my local shops.

I do not plan to nail down the wave characteristics associated with "proper soundstage" or any of the more
popular terms. What I hope to do is to determine that there are changes each component/cable etc makes on a signal. I just wanted some suggestions so that I could possibly have easy measurements. Those measurements will be stored in the computer. For instance, there would be a file containing all the cable info. I would then run a test on a certain amp/pre-amp combo. Based on the deviations from that test, the computer would suggest a cable that might bring the signal closer to the reference.

PLease read all this post carefully. It is a little long but I hope it clarifies some things and brings people up to speed with our currrent state.

I had thought this thread was dead... until I can complete my initial research and can post the results. But recently I have recieved a few emails (one about a month ago and another yesterday) and have noticed a few new posts as well. I think that I should restate a few things. Not because I wish to argue any points, the recent posts and emails are completely correct, but because I would like to clarify some things that we hope to accomplish.

I will not be using any sort of "human compass" test measurements initially. First I would like to establish a reference recording. Then I will play this recording from source (computer) through the cables in question and then into the measurement computer. If there is a difference then it will show up. I might not be able to say what that will correspond to if it were played through a "perfect amp" and "perfect speaker" and in a "perfect room", but I will be able to say that there is a difference.

Of course all of that is an oversimplification. However we believe that any time problems, or any frequency or harmonic anomolies will be detected. Again, what this translates into or how to really quantify the anomolies is not our initial goal. If it turns out that we can assign a set of characteristic anomolies to a certain cable (and our preliminary stuff says YES) then we will begin to set up a very controlled environment in which to go deeper. Of course there is always the possibility that a cable only acts the way it does when connected to Levinson or Rotel stuff and not with VTL or Spectral. If that is the reality of the situation... and we suspect it is, then this research will not validate it. What it will do is prove that cables can impart changes.

Our project is shaping into this:

We will use all Mark Levinson gear (No.39 CD, No 380 Pre, and a 336 Amp) going to Thiel speakers in a rectangular room. We will then use some friends as subjects. Again it is not our goal identify what cable produces the widest or most accurate soundstage but only that it does indeed change it. I think that anyone shoud be able to tell, on well recorded music, if there is big enough change.

Of course none of this will help anyone on this board when deciding on cable purchases. What we hope it will do is prove that it can be done. What we think needs to happen (and what we know never will) is that there needs to be some sort of standard test for all cables. Maybe this won't be "hey this cable gives a 2.3 whosawhatchit soundstage". But it might be something like this. "hey this gives a 2.3 whosawatchit which means it might cause timing or harmonic anomolies with equipment in this blah blah blah arbitrary catagory"

We also should mention that we in now way believe that we are doing research that has not been done before. What we do believe is that it has not been published. It is probably done once a day for the last 10 years in research labs at Madrigal or Spectral or TagMcLaren etc. etc.

So in summation, we will first look at a few cables to determine their reaction with a reference system. We will then try and document any measurable changes the cables impart on the system. The last stage will be to try and describe (in very, very basic audiophile terms) what those changes might be. If this can be done then we will have answered a question that many industry engineers already know. Will this mean that Joe Bloe with his $10,000 cables really does hear a better soundstage than Joe Schmoe does with his RatShack stuff... NO. But it will give some validity to these types of claims "I noticed a difference between cable A and Cable B"

Rayhall -

When I made the statement regarding lost waves I did not mean that literally. That was extremely simplified and ultimately a bad phrase. I wish I had a picture of what our software looks like graphically...

Try to imagine the (x) axis being the frequency band from 0 to 20,000 and the (y) axis being the amplitude. If I were to record myself saying the phrase "social science is an oxymoron" it would register my voice throughout the spectrum. The average of say the first syllable "so" would be the tone of my voice which a guitar tuner might register as "C#". However the reality of that syllable is the vibration of the air at many different frequencies. Different cables, because they have different physical properties can alter some of those frequencies. The audible result might not be a change of my voice from (C#) to (D) but it will change something.

The physics behind what makes a person perceive a "soundstage" as being wide or deep as opposed to shallow are actually known. It is a kind of mathematical physics that is best done on computers. As you probably know reverberation and or perceived reverberation is a major factor. When a signal gets altered one of things that can be lost is that sense of space or perhaps the exact placement of instruments. It is this kind of stuff that all those cheap DSP programs on A/V Receivers try to accomplish (the addition of space). The only good implementation of this technology that I have EVER heard is the LOGIC 7 stuff in Lexicon gear. In a way this is what a lot of people seek from upsampling technology. Some components do this very poorly (they add a sense of space but they alter the signal too much, obscuring other things). The really good components manage to add a sense of space without making a detrimental impact on the original signal. Of course this is ultimately a matter of taste but I think there is a general consensus that some components do accomplish it better than others.

All those references to "lifting a veil" are actually great analogies IMHO. Our goal is to find out what that veil looks like, does it have characteristic look (mathematically)? Is it merely a decrease in certain frequencies whose effects could be countered by a cable that amplifies those frequencies?

Again, all of this is really speculation since we do not have nearly enough data to say statistically whether or not our model is valid. Not to mention the fact that we have not done any listening evaluations.

I hope this was helpful and/or that I explained my project in more detail