TOPIC:

I'm trying to perform some headphone and amplifier comparisons by recording songs on my E.A.R.S. and then comparing them using Audio DiffMaker .

As a baseline, I recorded the same song twice in a row with no change on the source. My expectation is that the diff would essentially be null, but instead I actually hear some audio that sounds like a difference in recorded level between take 1 and take 2. Here's the audio diff itself.

Before I go too far down the road of figuring out if I'm doing something wrong, I'd like to know--am I expecting too much of the E.A.R.S. jig in terms of being able to repeatably record the same source material in the same way?

Interesting application. Even taking the headphones off the EARS and putting them on again will change the measurement. So generally people (e.g. Tyll of innerfidelity) recommend taking an average of measurements.

One thing you might want to try is loading a portion of the original track into Audacity and taking the spectrum, then doing the same with the diff file.

Just wondering why not use measurement sweeps?

For comparing amps, in general I would suggest doing that without the headphones in the loop - same thing with normal amps, i.e. I wouldn't be trying to measure difference in amps while measuring speaker output because the non-linearities in the transducer swamps those in the amp. With that said, for something like response changes with high output impedance amps, it might work. (But I think I'd measure it with an REW sweep).

Thanks John!

Even taking the headphones off the EARS and putting them on again will change the measurement.

For this particular experiment I'm keeping the headphones on the EARS without touching them.

taking an average of measurements

Good idea, I'll try that.

For comparing amps, in general I would suggest doing that without the headphones in the loop

Hehe. What I'm trying to do is somewhat different from the usual. I'm not actually trying to measure the amp, I'm trying to measure the effect of different sources (amp, DAC, etc) on the transducer. In particular, I've found that audio reviewers frequently discuss synergy between amps/DACs/DAPs and headphones, or this idea that a headphone "scales well" with improved amplification. Other than testing for impedance mismatches, I've never seen any attempt to back up such claims with measurements.

So, the hypotheses I wish to test are of the form "transducer T sounds noticeably different when driven by source S1 vs source S2".

why not use measurement sweeps?

One might make the argument that a sine wave doesn't sufficiently stress the system enough to discover meaningful differences when listening to actual music. Hence, my desire to test with actual music

I was actually recently able to produce a couple of trials on the same chain which were only barely different. I'm going to see if I can come up with a methodology for doing this repeatably.

pwjazz wrote:

Even taking the headphones off the EARS and putting them on again will change the measurement.

For this particular experiment I'm keeping the headphones on the EARS without touching them.

It's interesting that you would get that difference then... I don't have an explanation for it. The diff test is quite sensitive (i.e. just 0.1 dB difference will result in diff signal only -40 dB down from the original) but still... I'd have expected difference due to ambient noise etc... but not signal. Hm...