Sound source localization
With stereo speakers you have a different staging than with headphones. The localization of phantom sound sources works with headphones due to the lack of room acoustics with an impressive sharpness - but the stage is different.
In head-related localization is based on several localization stimuli. With headphones, the result is almost always an in-head location of the instruments - but this stage can vary greatly from model to model. The human ear has several complementary ways of locating a signal.
- The best known effect is based on intensity stereophony. If the sound level of the right and left ear is different, the brain uses the level differences to locate the sound source.
- Another possibility is running-time stereophony. If the sound arrives at the listener's side, the sound components arrive at the ear phase-shifted. The human ear is very sensitive to these so-called intraaural differences in transit time. Sound time differences of approx. 0.04 milliseconds can be resolved.
- Another effect is based on the external ear transfer function, also known as the Head Related Transfer Function (HRTF). The sound components to which the ear is most sensitive have a wavelength of 1 - 20 cm and thus move in similar orders of magnitude to the head and ear geometry. If an object (head, ear) is in the order of the wavelength, this object influences the wave field. The sound arriving at the ear is given a directional "signature".
Due to filter effects on the ear-geometry, diffraction and shading effects by the head, etc., the signature of sounds gets strongly modified - different for the left and right ear. The corresponding direction-dependent frequency bands are known as blue-touch bands.
With loudspeaker stereophony, the effect is not a big problem. The ears will add the "distortions" (=HRTF) to the incomming sound like it should, the brain will "decode" the sound like always, so the sounds can be located correct: between the speakers. Instinctive turning of the head will improve the localisation.
Figure 1: Stax electrostatic headphones with feeding unit
Tuning of neutral colored headphones
If headphones are used for mixing, from the developer's point of view the impact of the HRTF raises the question of how the headphones should be tuned. If the headphones are tuned to a linear frequency response, it will sound completely discolored with a standard stereo speaker recording. After all, the human brain has been accustomed since birth to the sound being pre-filtered by the HRTF.
- The most common approach - diffuse field equalization. In most listening rooms the diffusefield-sound is louder then the direct sound from the speakers. That means: Construct a plastic human head with microphones in the ears - then place it in an average room with omnidirectional speakers in opposite positions and play testmusic. Afterwards you put headphones on the plastic microphone-head and use an equalizer to change the frequency response of the phones until the measured sound matches the excitation from the external speaker. Done - you have an diffusefield tuned headphone setup.
- If you put that plastic head in front of two neutral loudspeakers in a anechoic chamber - you will get the headphones free-field equalized analogous to above.
Another approach is based on listening to a good stereo system: Here the plastic head is placed in front of good and neutral stereo speakers in a room with great acoustics. After the routine from above two further steps are needes for tuning: First: The bass of headphones is generally perceived as too quiet, as the bass is only heard with the ears and cannot be "felt" with the whole body. Therefore, the bass is adjusted a little bit more until the subjective listening experience fits. Second: the ear deals differently with incoming direct and diffuse sound when hearing the tonal colour, this is also subjectively corrected. With a similar approach the so-called "Harman listener target curve" for headphone tuning was defined - the naming coined by the Harman company.
Those three approches for tuning headphones sound logic. But there is a "but"...
Are you thinking, headphone measurements are more art then science? Yes? If there's still any doubt, we'll take the next step right away...
Figure 2: Measurents with two headphones (left) and one speaker (right)
Above we can see measurents from two pretty good and natural sounding headphones on the left - and a natural sounding speaker on the right.
The first thought might be here: The headphone measurements are easier to interpret. The radiation of the speakers is not measured, so its not possible to understand the interaction with the room acoustis - but the headphones dont radiate in the room.
But far from it: the measurement of the headphones is even more intransparent. The range of several kilohertz is primarily influenced by resonance effects of the HRTF. The headphones go very low in the low bass and will therefore be closed (=are in fact closed), which means that the measurements here depend heavily on the correct fit on the head.
If you move the headphones slightly on the artificial head the curves will probably start to wander.
Therefore: The interpretation of headphone measurements is often pure esotericism. But if you want to have that kind of fun, here are some links to help you:
To make the interpretation of the measurements even crazier: the frequency response of the headphones not only affects the timbre but also the stage.
Localisation with headphones
Like mentioned: You can record music with a plastic head microphone. If you hear the resulting recordings with a neutral headphone, you will (in theory) have neutral coloring and even a normal soundstage without inside-of-your-head-localisation. There are a lot of samples on youtube - grab a headphone and listen if you want. Search for "Binaural Recording". An example: https://www.youtube.com/watch?v=uwDNhBvotks
If you want to hear chartmusic with your headphones: You will have to accept a inside-of-the-head soundstage. Some ways to escape this...
- Computers with Creative Soundcards: They have a digital headphone-postprocessing for games that will emulate a speaker-soundstage. Works great with a normal headphone if you are into gaming, the results with music and movies are not that impressive.
- Beyerdynamic Headzone: Also based on digital intraaural sound postprocessing, but also compensates the position of the head by a movement sensor (=Headtracking). I tried it once, with my ears the effect less impressive then Creatives X-Fi Cards during gaming sessions.
- Smyth Realiser: Sensor based Headtracking combined with a personal measurement of the ears: You can put tiny microphones in you ear and calibrate the Smyth-Realiser to the sounding of a high-end recording studio or homecinema. You can save several setups on SD-card. The system has the price of a used car but is described as pretty amazing. I did not have a change to test it myself.
If you are ok with inside-of-head soundstage: Normal music recorded for speaker-stereophony can be heard quite ok on good free/diffusefield/Harmancurve equalized headphones. The equalization/tuning of the phones will replace parts of the HRTF during headphone-listening, so tuning of the headphone will not not only change the soundcolor but also the soundstage.
Diffuse field equalization depends heavily on the type and correct fit of the headphones.
- So-called in-ear headphones are pushed into the human ear canal - accordingly the natural HRTF of the ear is bypassed to a large extent, the InEar has to balance the full HRTF with its tuning. Cleanly tuned InEars are therefore difficult to find - and the evaluation of the timbre depends strongly on the listener. The use of equalizers can be useful.
- A counter-example is the AKG K1000, where the sides of the headphones could be folded away from the head. This means that the sound reaches the head in a way comparable to hearing via a stereo system - and the HRTF of the ear geometry allows you to run without bias. The tuning of such headphones is accordingly more simple.
- High-quality headphones often rely on large diaphragms (larger than 40mm), which allows the complete ear to be used as complete as possible. Often the drivers are slightly angled and a few centimeters away from the ear to allow the sound to arrive as naturally as possible.
- Electrostats and magnetostats are often very expensive, but usually have much larger diaphragms than conventional dynamic headphones. Those headphones can sound amazing.
If you want to tune headphones to a very sharp and present soundstage that is as present as possible, the tuning of the headphone frequency response based on the HRFT based "Blauert ribbons" shown below is recommended.
Figure 3: Tuning of headphones that are as present as possible
IfA correspondingly tuned headphone will have a much more stable and presentable stage image than conventional models. However, the trick for tuning is well known - many sound engineers use such filters as standard when mixing music, after all, this effect also works with normal loudspeaker reproduction. The combination of tuning the music music during recording and via speaker/headphone can be a bit too much.
Headphones will have a "wrong" stage while playing music recorded for normal speakers. Tuning the frequency response for better staging is ok for more accurate reproducing phones to get "closer" to the real sound.
In addition, the effect only works with music that covers the entire spectrum in broadband. With narrowband signals, the effect only has a limited effect. Developers of headphones therefore always need experience.
How do I find a neutral headphone?
If you are using the phones for mixing or just want natural soundcolor: Please hear several phones against neutral monitores and check if the soundcolor is natural with your ears. If possible: look at measurements while or after hearing the phones. Using an equalizer is ok to get the sound closer to your normal speakers.
I bought my personal loudspeakers mainly based on measurents - and I dont regret it. I have never done so with headphones, and I will never do. Just hear some phones, look if they are confortable, enjoy them. You will get amazing sounding heaphones for below 200€ - A comparable stereo system costs many times more. And requires considerable intervention in the room acoustics to achieve a similarly sharp stage image.
*click* Next section *click*