How to equalize your headphones: A Tutorial - Page 3

No, we should not be EQing reality, nor does the resonance that I speak of occur when listening to live sound. The resonance occurs with headphones, earbuds, and in-ear monitors specifically due to the acoustic impedance mismatch phenomenon.

I just performed a very interesting test. I put my headphones down on my desk, and turned SineGen up to its loudest output. I ran through the entire frequency range, and as far as I can tell, my headphones have a nearly perfectly linear frequency response. It is only when I actually put them on, so that the transducers are closely coupled with my ear canal and eardrum, that the frequency response becomes uneven.

I don't know what this is supposed to be about. The experiments that we are using to find the resonance peaks are repeatable and verifiable, and confirm the data contained in the graphs I posted.

If you're equalizing for both peaks and notches, that should take care of itself during the equalization process, as you attempt to achieve a uniform response to the sine wave and pink noise sample.

Those are equal loudness curves, which describe the sound pressures needed to perceive a uniform response to sine waves at a particular loudness, with the very high and very low frequencies needing more and more boosting as the overall loudness level drops... and also have absolutely nothing to do with what I am talking about. Perhaps you should read this quote, carefully:

This is the problem I am talking about, and the problem I am trying to fix. It is an inevitable part of the headphone experience, and does not occur with speakers or in live situations, to my knowledge. The point of equalizing your headphones to flat is so that they give the purest possible reproduction of music. If you are equalizing properly, you will not have to worry about equal loudness curves, because that will take care of itself during the process. At any rate, the deeper problems of psychoacoustics are beyond the scope of this tutorial.

I only just barely understand the concept of acoustic impedance as it is (too much math!), so I cannot provide a good technical reply to this. There are ways to test for this though. The best way I can think of would be to use a pair of studio monitors that you know to have a flat output, and a pair of headphones that you know to have a flat output (such as STAX, for example.)

Use the SineGen program with both the speakers and the headphones and note where the perceived amplitude of the sound increases. With the speakers, the sound will be more or less uniform, with an increase around 2.5-3kHz due to the pinna. If the speakers are truly flat, there will not be any sharp peaks in the upper mids or highs. With the headphones, however, the response will start to become increasingly jagged above 2kHz, with a sharp peak in the 5-8kHz range. You can verify the presence of the problem by taking the headphones off, placing them on the desk and turning the volume up to where you can hear a 1kHz tone clearly. If you run the frequency sweep again, it will suddenly appear to be much more uniform in amplitude. Put the headphones back on, and the problem comes back.

The human auditory system does not normally create such a large peak in that specific frequency range. It can only be due to the unnaturally close coupling of the headphones to the ear canal and eardrum. When I listen to music on my studio monitors, it does not have the exaggerated high frequency energy that it does when I listen with my headphones unequalized, even though my headphones have a technically flatter response.

Unfortunately, the analyzer does not know what you are hearing, so equalizing based on what it tells you does nothing. If you are interested, you should start with SineGen as soon as possible and see what you can find out.

Awesome! Now that's what I'm talking about. Looks awfully familiar too...



My response.



Linkwitz's response. Notice that the HD414's had some problems in the 2.5kHz range, too. But look at the response he discovered for the high frequencies. I'm beginning to see a pattern here.

The Burn in Wave generator does not allow for the sweeping of the sine wave by hand, though, which is important for determining where the peak is and how loud it is relative to other frequencies.

It will work with any pair of headphones. I've even equalized cruddy pairs of grocery store headphones, just for kicks. To tell what is a peak and what isn't, just put the headphones on, load up the SineGen program (which I provided a link to), and press the "Power" button. Use the frequency slider to move the tone slowly up and down the frequency range. If there is a peak, you will perceive it as a sharp and sudden increase in volume on a particular group of frequencies.

You know, I think I will post some samples of music that I have heard that easily reveal the presence of the anomaly. They will be helpful demonstrating the problem to others.