
When all you have is a hammer, everything starts looking like a nail.
I'm about to start some cable listening tests. There's a bunch of cable makers out there, but I went with two fairly well known headphone cable makers because I mainly want to make the point that cables do make a difference, but not a big difference. (I think people are well served by communicating with other enthusiasts and the cable makers themselves to get a feel for which ones will best suit your particular application.) I find the changes made by cables subtle, yet often worth while. Once you have a source, amp, and headphones you like, the last thing you can do is get a gentle tweak in a preferred direction by changing the cables on the headphones. Audiophile icing on the headphone system cake, if you like. But there's a problem...well, many problems really.
I measure headphones all the time, and believe there's a fairly strong relationship between measurements and what I hear. So, the objectivist in me wants to see wants to see identifiable features in the data that correlate with my listening experience. Cables are a bit of a buggaboo though, in that I don't have the proper equipment to measure them well. The right tools for that job would be capacitance meters, time domain reflectometers, and stuff like that. I've got enough experience measuring things that I know measuring the differences cables make on my headphone measurement system is just the wrong way to go about it. However, I get the request to measure different headphone cables all the time, so I figure at some point I just needed to do it.
Another problem with measuring cables on a headphone measurement system is that the differences between cables using that measurement method are bound to be very small. When I swap cables on the headphones I have to remove the cans from the measurement head to replace the cable, and then have to put the headphones back on the head. This will result in a slightly different position of the headphones on the head from the prior measurement. These positional changes will cause changes in the sound measured, and those changes are likely to be significantly larger than the effects of the various cables.
Lastly, knowing with relative certainty that I was going to get a result that would tend to indicate that swapping cables made no measurable differences, I know I'm likely to just give fuel for objectivists to throw on the fire claiming that cables don't make a difference. I kinda don't like to do that because I do believe cables make a difference. Oh well...I guess I just have to let the chips fall where they may.
How I Conducted the Test
I used a pair of Sennheiser HD 650 headphones of recent manufacture. I have six cables to use in the test: stock, Cardas Headphone, Cardas Clear, Moon Audio Black Dragon, Moon Audio Blue Dragon, and Moon Audio Silver Dragon. I took the head out of the chamber and put it next to the tester and headphone amp as I didn't want to have the headphone extension cable I normally use in the experiment. (The silence the chamber provides is really only critical for THD+noise tests. These tests were performed with 90dBspl in the headphones, well above the roughly 50dBspl room noise level.)
Once I put the headphones on the head and adjusted the headband for proper fit, I taped the headband to ensure that setting remained the same. When I placed the headphones on the head, I had a mark on the top of the head to position the headband, and then I would slide the earpieces forward gently until I felt the ear just touch the rear of the cushion. This seemed to work quit well and repeatably.
For each cable I would take five frequency response and impulse response measurements. Frequency response measurements should give a good idea of any tonal changes, and the impulse response should give some idea of changes in time alignment. I removed and replaced the headphones on the head for each measurement. I also measured the stock cables twice to see if those measurements agreed more closely than when swapping cables.
Once the data was in a spreadsheet, the five measurements were averaged together for each of the four data sets (left and right frequency response, and left and right impulse response). The reason for doing this was to try to average out any small changes due to the positional changes of the headphones on the head. After each cable had been measure five times, I would swap to a new cable and repeat the process.
Once all the data had been gathered and numbers crunched, I copied the final results for each cable onto a master spreadsheet. There, any slight vertical offset difference in frequency response was compensated for by off-setting each curve such that all curves went through the same value at 800Hz. The impulse response curves don't need this as they naturally have 0 volts as a reference.
Okay, turn the page to see the results.