Eilex Binaural

Eilex Binaural is a process which converts recorded materials intended for stereo speakers to Binaural Audio ideal for reproduction with headphones/earphones.  It maximizes the performance of any headphones/earphones while maintaining their tonal characters.

Eilex Binaural is suitable for 2-ch movie soundtracks and internet audio reproduction.

  • Binauralize Standard Stereo Source in Real-Time
  • Natural Headphone/Earphone Sound
  • Proprietary HRTF
  • Precise VIR Filtering of HRTF and Middle-Ear Frequency Response
  • Minimum Fatigue for Extended Listening
  • Applicable on All Headphones/Earphones
  • Preservation of Headphones/Earphones Tone Character
  • Proprietary APVD (Acoustic Power Volume Density) Measurements
  • Hi-Res Audio Compatibility

1. Problems of Headphone Sound Reproduction

Headphones/earphones always deliver an artificial sound image, regardless of their quality. Compared to speakers, the disadvantages of headphones are:

  • A vocal in the center moves to the middle of your head.
  • Panned sound shifts to the extreme left or right.
  • Details are over-emphasized, damaging the tonal balance.
  • The amount of reverb is exaggerated.
  • Different parts of the sound from the same instrument come from different directions.
  • Background instruments move from behind a singer to the sides.  
  • The placement of instruments in an orchestra get mixed up.

These problems do not happen with speakers.  The cause of the problem is reproducing materials meant for speakers with headphones.

When listening to a stereo system, both ears can hear both the L and R speakers.  When listening through headphones, each ear only hears one channel.  This is the reason why the headphone audio produces a strange sound image and an uncomfortable feeling.  (See Note 1.)

This problem has been studied by many researchers for long time.  It is known that feeding a part of the sound from the opposite channel reduces the problem.  This technique is called cross-feeding.  

But cross-feeding creates a new problem: it tends to muffle the tone, and we cannot simply boost the highs to get the right sound.  When using speakers, the L and R sound reach both ears, the sound reflects and diffracts around the head, outer ears, torso and shoulders.  This adds quite complex filtering to the sound.  We have to include this frequency response of the reflection/diffraction to mimic real world circumstances.  This is called HRTF (Head Related Transfer Function).

Another important factor is that there is a time delay between the sound arriving at the left and right ears from a speaker.  This time difference must be included in the cross-feed, because it directly relates to the perception of sound in terms of directionality and spatial image.  

2. Binaural Recording

There is a recording technique called Binaural recording, which is intended to be reproduced by headphones.  To mimic the physical auditory situation, the Binaural recording is performed by using a dummy head and torso with microphones buried in the ears.  Thus the Binaural recording automatically includes the cross-feed, HTFR and time delay.  (The Binaural recording is the most suitable for reproduction using headphones, but not for speakers.)

3. Solutions by Eilex

Eilex Binaural converts regular stereo materials into Binaural format in real-time.

Binauralization makes high quality headphone sound reproduction promising, but there are still some remaining problems, such as correct HRTF measurement; determination of HRTF frequency range; treatment of higher frequencies over the HRTF range; and management of personal frequency response of outer and middle ears.  All such filtering must be accomplished by using one or two filters.  The total frequency response of the combined filter is quite complex.  The conventional FIR or IIR filter cannot be a practical solution.  Eilex’ proprietary VIR (Variable-resolution Impulse Response) filter solves this problem, providing highest accuracy and resolution using minimal computational resources.

Eilex has studied headphone audio from both the physical and psychoacoustic aspects for many years.  Based on these studies, Eilex has developed a headphone signal processing algorithm called the Eilex Binaural process.

The Eilex Binaural process consists of a VIR filter which compensates for the high frequency response of the ears, and a pre-process which takes care of the psychoacoustic characteristics based on the HRTF.

Fig. 1 and 2 show the actual operation of the cross-feed.  The signal source is the first 40 seconds of The Beatles’ “Here Comes the Sun”.  The upper and lower traces are L and R channels, respectively.  As seen in Fig. 1, the song starts with a guitar on the L-ch, while the R-ch is totally silent for 15 seconds.  The guitar from the left speaker sounds just fine for both ears.  But on the headphones the guitar is heard only by the left ear, which creates an odd feeling because this never happens in the natural world.  

Fig. 2 shows the result of cross-feeding.  A signal of almost the same shape and strength appears on the R-ch, as it does when listening to a speaker.  An appropriate time difference is added to the cross-feed to preserve the directionality and spatial information.  

Here Comes the Sun (Original)
Fig. 1  “Here Comes the Sun”, Original Click on image to enlarge.
Here Comes the Sun (Cross-Fed)
Fig. 2  “Here Comes the Sun”, with Cross-feed Click on image to enlarge.

Fig. 3 shows the original waveform of the whole tune of “Flying”.  Fig. 4 is after cross-feed is applied.  The L and R waveforms become very similar, yet a normal stereo image is maintained.

Flying (Original)
Fig. 3  “Flying”, Original Click on image to enlarge.
Flying (Cross-Fed)
Fig. 4  “Flying”, with Cross-feed Click on image to enlarge.

4. Types of Headphones and Earphones

Depending on how they are acoustically coupled with ears, headphones and earphones fall into the following four types:

Headphones

  • Around-ear (Circumaural)
  • On-ear (Supra-aural)

Earphones

  • Earbuds (Concha)
  • In-ear (Canalphones)

The frequency responses the headphones and earphones require are significantly different between these types.  Our ear is an acoustic transmission line consisting of the outer ear, ear canal and eardrum.  When a sound from the natural world (where the acoustic power frequency response is flat) arrives at the outer ear, the frequency characteristics start to change.  The sound continuously changes until it reaches the eardrum.  Therefore, depending on the location of the headphone or earphone on the ear, the appropriate frequency characteristics must be applied to create the perception of a flat response.  

Figs. 5 through 8 show examples of the frequency responses of different types of headphones and earphones.  These are actual measurement results of well-known brand products.  It should be noted that each type of device has a unique frequency response.

Headphone makers actually add some types of HRTF to their products to compensate for frequency response.  Cross-feeding is not considered in this HRTF, because they do not expect the typical headphone players to support cross-feeding.  HRTF characteristics differ significantly even among the same brand’s products, providing totally different sounds and characters.

Fig. 5 Around-Ear (Circumaural)
Fig. 5  Around-Ear (Circumaural)
Fig. 6  On-Ear (Supra-aural)
Fig. 6  On-Ear (Supra-aural)
Fig. 7  Earbuds (Concha)
Fig. 7  Earbuds (Concha)
Fig. 8  In-Ear (Canalphones)
Fig. 8  In-Ear (Canalphones)

5. Sound Improvement by Eilex Binaural

The Eilex Binaural process makes headphone/earphone sound identical to reference speakers. Reference monitor speakers are equalized to a flat APVD (Acoustic Power Volume Density) frequency response by Eilex PRISM. The flat APVD speakers reproduce a true-to-source sound in the room.

The Eilex Binaural process is calibrated by multiple experts to match the sound of headphones and earphones to the reference monitor speakers.

The sound difference between the headphone/earphone models mainly comes from the difference of high frequency sensitivity and characteristics. Eilex Binaural process has High Frequency Level control to handle the wide variety of sound character of unspecified headphones and earphones.

Listeners can adjust (boost or cut) the high frequency level to what sounds most natural. The sound should be closer to the APVD flat speakers. (Note: This process is unnecessary for speakers because listeners hear exactly the same sound.  Headphones and earphones provide a different sound to each individual’s ears, so different adjustment is necessary.)

When it is required to tune a particular pair of headphones/earphones to perform at an ideal performance level, the optional Custom EQ (Ref: 6. Custom EQ) is used.

As with speakers, the tonal character and taste of each brand’s headphones is often as important as the sound quality itself.  The compensation coefficients of Eilex Binaural Headphone are carefully composed to preserve the brand’s intended character for their products.

6. Custom EQ

Eilex Binaural’s Custom EQ is for the specific headphones and earphones which CE manufacturers often combine with their products.  Eilex makes measurement of the given headphones’ APVD frequency response and create a filter to match the ideal characteristics.  Then, by multiple experts the compensation curve is tuned to match the sound with the APVD-flat reference monitor speakers.  The end user selects the Custom Mode, then finds the best compensation level for the higher frequencies by High Frequency Level control.

7. Movie Reproduction

Headphones and earphones are typically used when movies are played on a PC, laptop, notebook, tablet, smartphone or mobile phone.  Such movie soundtracks are normally 2-ch stereo.  Eilex Binaural is ideal for reproducing 2-ch soundtracks; it improves the sound quality and restores the sound image.

Eilex Binaural improves the sound quality by matching the frequency response of headphones and earphones to the user’s ears.  It improves sound clarity, and details are more easily heard.  Speech intelligibility improves and dialogue becomes clearer.  Sound effects are more vivid.

2-ch soundtracks, down-mixed from multi-track materials, tend to reproduce uncomfortably for headphone users.  The components which were originally a part of the surround sound, such as explosions with continuous mid-low frequencies, ambient noise, rain, thunder, and applause sound unnaturally close to the ears.  Such close sound to the ears masks more important sounds, including dialogue.  Also, the sound image for sound effects expands too wide, creating location discrepancies between the sound and picture.  Eilex Binaural’s Cinema Mode corrects these types of errors and restores the sound image as if it originated from two front-firing speakers.

Note: When more dramatic sound effects are desired, Surround Mode can be used. Surround Mode no longer involves Binauralizer; a more aggressive matrix, delays and reverberation are used.

8. Internet

Internet audio is normally 2-ch stereo.  PCs, laptops, notebooks, tablets, smartphones, and mobile phones are used for reproduction.  The content consists of news reports, TV channels, radio, videos, movies, music, commercials, games, etc.  Headphones/earphones are more commonly used for audio than speakers, and exactly the same audio problems arise as with movies.

Headphones do a much worse job of reproducing internet sound than speakers; the inferior sound quality of internet audio is exaggerated by headphones.  This is because the noises and distortion caused by compression errors of low bitrate signals are heard very close to the ears.  Eilex Binaural pushes such noise and distortion away from the ears, leaving only the clearer, smoother sound.  Eilex Binaural very effectively improves all internet audio.  

9. HA Mode (Optional)

Hearing aids restore elderly people’s hearing enough for daily activities, but their quality falls short of high-end audio.  HA Mode provides natural sound for the hearing-impaired without the need for hearing aids.  No hearing aids correctly compensate for hearing loss from occupational noise-related hearing damage such as aircraft, machinery and loud music.

Eilex Binaural offers HA (Hearing Assist) Mode as an option, which enables the hearing-impaired to enjoy good quality sound with headphones or earphones, without hearing aids.

There are two types of hearing loss, age-related (Type A) and noise-related (Type N).  These require different types of compensation.  HA Mode has coefficients, consisting of HRTF and a special curve, to compensate for both types of hearing loss.  It also has three different levels of compensation to match the extent of the user’s hearing loss.

(Eilex is continuously developing more effective HA Mode coefficients.  The coefficients will be updated from time to time.)      

10. Compatibility

Eilex Binaural process has compatibility with the following:

  • Audio CD and all lossless and lossy digital compression formats, including MP3 and AAC  [Note: Refer to Eilex Harmony and Eilex Focus for sound recovery]
  • Down-mixed source materials from multi-channel recordings
  • Monaural Materials
  • Noise-cancelling headphones
  • Tone control
  • Equalizers (parametric and graphic)
  • Preset EQ (classic, jazz, pop, etc.) [Not recommended to use other than for marketing purposes.]
  • Hi-Res Audio [Note: Refer to Eilex HD Remaster for Hi-Res up-conversion]

Compatibility is not guaranteed in the following cases:

  • On-board 3D or surround post-processing within the MP3 player
  • Headphones with internal (electronic or mechanical) 3D or surround processes
  • Headphones for the hearing impaired
  • Use with hearing-aids

No compatibility (Eilex Binaural process must be turned off):

  • Binaural recordings
  • Headphones with electronic or mechanical (acoustic) cross-feed

11. Realization of Eilex Binaural

Eilex delivers to licensees or device makers a set of VIR filter coefficients and software to realize the Eilex Binaural process.

A licensee or device maker would install the Eilex Binaural firmware on the device, while the licensee develops the user menu (GUI).  The GUI lets a user toggle the Eilex Binaural process, choose a desirable high frequency enhancement level, and adjust the width of the sound stage and the location (distance) of center vocal.

The VIR filter coefficients should be stored somewhere in the DSP, CPU, flash memory, HDD or main memory.


Note 1:

Speaker and Recording Reproduction
Problems in Headphone Reproduction
Eilex Binaural Headphone

Note 2:

Even if perfect speakers existed, in theory there can still be no perfect headphones.  The reason is that there are absolute references in the natural world, but not for headphones.  The “absolute reference” is every sound we hear in the natural world, including speaker sound.  Multiple people can listen to these sounds together.  (There is no way to know how each person hears the sound, but that doesn’t matter.)

If a speaker’s sound is extremely close to the original sound, that sounds great to everybody.  The speaker can be perfect.  It is possible to make a perfect speaker for everybody.

Headphones are totally different.  When we wear headphones or earphones, some unique individual frequency characteristics appear which don’t exist in the natural world.  The problem is that the characteristics are significantly different from person to person, and even between the left and right ears of the same person.  In the natural world, there is essentially no auditory difference between the left and right ears of a person, except for the case of physical change by age, damage, etc.  Our ears are all asymmetrical by birth, thus so are the frequency characteristics.  Fortunately, while we are growing up, our brains learn to equalize the sound to hear it correctly.  Therefore the physical difference between the left and right ears does not alter our actual perception.  (There are similar stories in the visual perception.)

Even if there is a pair of perfect headphones for a person, they will never be perfect to others because everybody has different ears.  This is the most important difference between listening to speakers and headphones.

To make better headphones for everybody, makers can find the average frequency response of the human ear to use as a reference.  The products are fine-tuned by experts with accurate ears.  We all know there are obvious differences in sound between makers or even between products of the same brand.  This is indicative of the difficulties of making good headphones.