10 Jun 2017

Headphone correction

News Comments Off on Headphone correction

You might be perfectly happy with your current cans. Prior to purchase you studied reviews, evaluated the headphones by careful listening, got suggestions from friends, and you’ve paid a decent amount of money for high-quality, studio-grade headphones.

Does that mean you can trust them? Do they provide a reliable, honest representation of the audio content you are listening to? How consistent is the sound with other headphones? Does your mix translate correctly between headphones and loudspeakers? Read on to gain a better insight on headphone frequency responses and what we think constitutes good headphones.

The questions above led us at ToneBoosters to study this topic more rigorously. How much variation exists in headphone frequency responses? Are expensive cans more reliable than cheap ones? What is a good frequency response for headphones? Is it flat, or perhaps something else?

To better understand how we can quantify the quality of headphones, we started out with a thorough literature survey. In perfect agreement with our measurements of tons of headphones (see below), we found that the frequency response of acclaimed headphones when measured on a professional dummy head is typically not flat. Think about it. Why not flat? Aren’t we taught that flat is the ideal response? Yes… For loudspeakers! The golden reference for audio production and playback has always been a loudspeaker with a flat on-axis response. If such spectrally flat loudspeaker has the ideal response, it is easy to understand that for reproducibility and consistency of levels, EQ, and timbre, a headphone should provide a response that is equivalent to such an ideal, spectrally flat loudspeaker. Now, it turns out that if we measure the response of a perfect loudspeaker on a dummy head, with an actual head, ears, shoulders, etcetera, the resulting response at the ear drums is in fact far from flat. The change from a spectrally flat loudspeaker in an anechoic environment to a response measured at a dummy head is often referred to as a ‘head related transfer function’, or HRTF in short. These HRTFs have very specific spectral features; one of the most prominent ones is a strong peak at around 3.5 kHz that is in part the result of resonances in our ear canal.

Next we measured tons of headphones with scientific accuracy and in accordance with international ITU standards that describe how these measurements should be done. We quantified the variation across the measured responses, and correlated this with headphone type (circum-aural, supra-aural and in-ear type), and retail price. We basically wanted to understand if more expensive models would provide a more consistent, accurate and truthful reproduction than cheaper models.

Variance in frequency response across headphones

Mean and variance of headphone response curves as a function of frequency. Top panels: lower, middle, and upper price quantiles. Bottom panels: CA, IE, and SA headphones. RMSD denotes the root-mean-square deviation across frequency (full frequency axis and below 100 Hz only).

What do you think the outcome was? Read it in our latest publication in the acclaimed and peer-reviewed Journal of the Acoustical Society of America (link below). We’ll give you a hint on the conclusions here…. Our study indicates that basically all headphone categories (from cheap to very expensive) will benefit from headphone correction tools such as Toneboosters Morphit!

Link to our scientific paper: http://asa.scitation.org/doi/10.1121/1.4984044

Comments are closed.