T.H.E. Show Newport 2014 Highlight: New and Improved SPL Phonitor 2

Having been the first person to put an analog cross-feed circuit into a commercial headphone amplifier while at HeadRoom I find myself irresistibly attracted to the SPL Phonitor 2 headphone amplifier. It's cross-feed circuit is simply outstanding!

When you listen to a pair of speakers and turn the left speaker off both ears continue to hear the right speaker, but because the left ear is slightly farther away and in the shadow of the head it hears the sound slightly after the right ear (about 300-400uSec) and lower in level with some EQ changes. In headphones when you turn the left channel off the left ears hears nothing and gets no inter-aural time difference information. This is unnatural and leads to poor imaging (blob in the left, right, and center, with little in the betweens) and listening fatigue (because the brain keeps working to localize the sound without the proper cues).

The world of headphone virtual audio (making sound in headphones seem like it's coming from outside your head) is moving forward at a rapid pace driven by gaming, home theater, and smart headphone demands, but all those applications perform the virtualization process digitally, and there's all sorts of room for poor sound quality to creep in. Analog cross-feed circuits, on the other hand, allow the processing to be done with little damage to signals from upstream gear. In other words, if you've got a killer DAC or LP playback system, you really won't want to put it through an ADC, DSP, and DAC to get your headphone virtualization. Analog cross-feeds keep your signal as pure as possible.

The downside of analog cross-feed is that it can't reproduce the complex Head Related Transfer Function (HRTF) needed for a convincing, out-of-your-head listening experience. What it can do though, is fix the basic psychoacoustic problems discussed above when listening to headphones. The magic of the Phonitor is variable controls for the inter-aural level difference (labeled Crossfeed) and inter-aural time difference (labeled Angle) of the cross-feed circuit. Additionally, because the ratio of the sum and difference component of various recordings may vary, a Center knob is available on the Phonitor to adjust the intensity of the center image relative to the sides. I can't tell you how cool it is to have this adjustment—you'll just have to hear it for yourself.

SPL is a German pro audio company that makes a wide variety of recording studio electronics. The Phonitor is a result of their efforts to make a piece of studio gear that allowed sound engineers to mix and master accurately on headphones, which they sometimes just have to do. The Phonitor has been well received in the pro audio market. What SPL didn't expect was the enthusiastic response headphone hobbyists would heap on the Phonitor...heck, I doubt SPL knew we existed.

The Phonitor 2 is essentially SPL's response to the enthusiastic headphone hobbyist market as they clamored for just a few changes to the original Phonitor make it more friendly for high-end consumers. The Phonitor 2 adds unbalanced inputs, output switching, and remote control among a number of other subtle changes. I'll let Marty Druckman tell you all about it in the video.

For more info go to SPL's USA home page, and the Phonitor 2 product page.

Hjelmevold's picture

Did they have the SPL Phonitor Mini there, as well? I can't wait until mine arrives (ETA one month from now).

Here's an interview of SPL by Sound On Sound done at Musikmesse 2014 in Frankfurt, describing the Phonitor Mini

A quick clarification on a common misconception made in the video in this Innerfidelity article:
- The difference between +4dBu and -10dBV is not 14dB. As always when using decibels: It's all about the reference value, and here the reference value is different. 0dBu = 0.775V, while 0dBV = 1V. So after doing the math, the difference between +4dBu and -10dbV is really 11.8dB. There is not too much difference between 14dB and 11.8dB in practical terms, but now you know.

As for the new SPL Phonitor 2: It's still one of the best heaphone amps out there, even if they took away both the dim/pad feature and the PPM option from the VU meter. It's also nice to see a video that actually demonstrates the motorized volume pot in action, so users who can't try this in person can see how fast/slow it moves.

Hi-Reality's picture


First of all, thank you for the seminar "Future Directions for Active Headphones" at T.H.E. Show. I think it was great. I thought there I had coined "Augmented Audio Reality" until I did a search and found it is actually a popular research area.

Will this Phonitor enable me to narrow a binaural recordings sound-stage so that my headphones can better simulate a 2-channel speaker system?

When I listened to Chesky's new binaural recording I found an issue: the sound-stage's width is too large and not realistic (is is also like, as you wrote, gaps in the sound-stage). I would like to narrow it in order to create a more realistic sensation. Can Phonitor enable me to do that?

If I have understood you correctly; Phonitor can fix the sound stage issue with the headphones but not the inside-your-head issue? for that it takes HRTF-optimized headphones like the NAD VISO HP50? or are they trying to do the same thing?

Babak, Hi-Reality.org

Hjelmevold's picture

Although headphone crossfeed designs do use HRTF, and the NAD VISO HP50 has taken the frequency response aspect of HRTF into their design; neither of this does magically mean that their designs are all compatible with the HRTF of the dummy head and that they somehow can modify each other's attributes. HRTF is a collection of data (typically impulse responses) that can help you deduce how sound is modified when it reaches the human ear, but different designs use completely different approaches to how they modify the sound in order to simulate the HTRF.

I think the best answer to what will happen if you use the Phonitor crossfeed on a binaural recording, is that the signal coming out of your headphones will then sound as if the binaural signal was played back on speakers, not on headphones. It will probably have an effect of sounding narrower, but adding crossfeed will also add attributes to the sound that, simply speaking, will sound as if they've been added to the audio in a separate step.

I know that there are a few examples in the pro audio and research world of software that try to do what you're asking, but it's usually very experimental and might not work to an acceptable degree in your situation. One such product is the TB OmniSone VST plugin by ToneBoosters, although from what I understand, that one is meant for re-panning transaural (i.e. speaker-compatible) and not binaurally encoded audio. It will probably not sound entirely convincing on binaural recordings, but my guess is that it will sound decent enough for private listening purposes.

Hi-Reality's picture

Hjelmevold, thanks for your feedback. I try to understand the background to a number of issues I've observed (I've posed a related question; see the link below).

It looks like binaural recording is the right approach to get headphones produce more realistic audio than they do today (although I haven't yet heard DTS HeadphoneX or the gears you mentioned).

if you use the Phonitor.. if the binaural signal was played back on speakers, not on headphones. It will probably have an effect of sounding narrower,..

This was my question. So perhaps you meant "Short answer: Yes"; correct?

Here are a couple of problems I encounter when I listen to Chesky's new binaural recording album (through a Grado SRi60).

1- Imaging/soundstage is not realistic (seems not to represent that of the live performance); the width is too wide with gaps AND the height is too much up.

2- Out-of-Head effect is not realistic enough. Especially, the distance-to-binaural-head test clips can't trick my brain convincing enough. (the most realistic binaural recording I've heard to date is a less-known one in which a person is shaking a matchbox around and up&down the dummy head).

Hjelmevold, how would you go about to deliver the most realistic audio performance possible today (i.e. from recording to playback via headphones)?

Kind regards, Babak

Link 1: http://www.innerfidelity.com/content/ear-opening-experience-chesky-ultim...

Link 2: http://www.soundandvision.com/content/nad-viso-hp50-headphone#comment-50...

Hjelmevold's picture

Yes, crossfeed will make the audio sound as if it's played through speakers. No, crossfeed will not make the binaural audio sound more realistic (in objective terms), because the audio is then clouded with additional processing, and only traces remain of the binaural psycho-acoustic effect. And like I said in my post above: Altering the binaural encoding in a successful manner, is still unknown territory with today's knowledge and technology.

Have you ever listened to binaural recordings on speakers? If that end result is really what you're after, then crossfeed will help. But from my own experience, I don't think you'll find the result sounding any more realistic than on headphones. While many dummy head microphones are diffuse-field equalized in order to sound acceptable on speakers, on speakers you lose the special immersion effect that comes with binaural playback on headphones, due to crossfeed, acoustics, and other interference. You'll end up with a stereo effect that usually sounds nice, but it also sounds over-processed. In those cases, you'll get a more clean-sounding result if you listen to recordings done with other stereo miking techniques such as ORTF, NOS, XY, AB, M/S, Blümlein, and similar. Note that I say "clean-sounding" and not "realistic", as realism is an illusion that only the individual is capable of judging.

So what could be the reason why you find that the Chesky album sounds ultra-wide and too high up when played back on your Grado SR60i headphones?

One factor could be the headphones themselves. I don't know anything about Grado headphones, but looking at the measurement graphs on Innerfidelity, the SR60i has a sharp peak around 2kHz followed by a dip in frequency response from 2.5-5kHz before rising again towards 10kHz. This contrast could be the reason why you find that binaural audio played through your Grados sounds too far off to the side and too high up. These psychoacoustic effects are very complex and may even vary from one individual to the other, so it's hard to be sure until you've tried other headphones with a different frequency response. You mention the NAD VISO HP50 as an option, and I would say that they'd probably sound a little too dark to be immediately suitable for binaural, but you'll get used to them after a while.

Another factor could be in the manner that the album has been recorded. I haven't heard the Chesky album you're referring to, but I've heard several dozen albums that have been recorded binaurally. I would put these into two categories:

  1. Recordings that are modest, with little movement and usually recorded at a distance. The sound stage is for the most part narrow and static positioning, with only occasional elements that can be heard either outside of the established sound stage, or close to the listener. Examples include binaural recordings of classical concerts, but also field recordings done in nature.
  2. Recordings that are more extreme, and that try to exploit the capabilities of binaural audio as an effect. Usually full of moving sound sources, far off to the side and behind, but also close to the ears. Examples include the infamous virtual barber shop demonstration. I suspect that the Chesky album could fall into this category, but I haven't heard it yet.

The only way to have a binaural recording sound less wide without destroying the illusion, is to modify the physical sound stage at the time of recording. Either by grouping the musicians and sound sources together, or by moving the binaural microphone farther away. In my first category above, this has been taken into account, and the point is to not make the listener be aware that it's a binaural recording. But in the second category above, people go crazy and don't mind if the listener gets spooked, and it's a point that the listener is made aware of (and impressed by) the binaural recording. Sadly, it's easier to sell "3D immersion" than "modest/natural sound stage".


Hjelmevold, how would you go about to deliver the most realistic audio performance possible today (i.e. from recording to playback via headphones)?

I don't want to use the term "realistic", as recorded audio is always an illusion compared to the original event. But if I were to create an audio experience that resembles an original perfomer-listener setup as closely as possible, I would do one of the following:

- Record using a binaural dummy head, such as the Neumann KU100 (expensive and best quality) or SoundMan OKM / 3Dio FreeSpace (cheap and decent).
- Mix and master using as little processing as possible, such as avoiding EQ, dynamic compression and similar effects
- Play back on flat/calibrated-to-sound-flat, comfortable-to-wear headphones at relatively low volume (<85dBA nominal), with an additional well-integrated subwoofer to provide rumble effect
- Ask the listener to close his eyes and not rotate his head during listening, or provide synchronised video showing the audio performance (the 3Dio FreeSpace is a very practical tool to make these kind of audiovisual recordings).

or B)
- Record using a surround miking technique, such as decca tree, Trinnov SRP, or similar.
- Carefully mix all tracks to synthesize a realistic sound stage, in surround format
- Play back on either a surround speaker system, or a surround headphone simulation with head tracking, such as the Smyth Realiser A8

As for headphones that I can recommend for binaural listening, expensive examples include Mr Speakers Alpha Dog and Shure SRH1840, while less expensive examples include Shure SRH940, the Superlux HD 681, and even the original Apple iPod Ear Buds. You'll just have to try a lot of different options.

Hjelmevold's picture

I forgot to add (seems like I can't edit my posts here on Innerfidelity):

A simple method that you can use to reduce the perceived width of any stereo signal, is to use Mid-Side processing, and reduce the level of the Side signal. I can name a dozen of pro audio plugins/gear that does this (and I can explain how to do the same thing manually inside a DAW), but if you're a regular enthusiast who doesn't use pro audio software/gear, I don't know of any specific "box" that does this automatically for you. Basically what you need is a stereo widener (like the BSG Reveel reviewed here on Innerfidelity) that reduces the level of the Side signal instead of the Mid signal. That should do what you've asked for, at least with some level of success.

Hjelmevold's picture

...I meant to say you need an M/S processing device similar to the BSG Reveel (or any other consumer "stereo widener"), but one that does the opposite ;-)

xnor's picture

I'm sorry for nitpicking again, but not correcting such things usually causes confusion.

Balanced and unbalanced per se have nothing to say about levels, in fact you can transmit exactly the same unbalanced signal over a balanced line. Balanced is used to reduce interference.

Decibel is just a ratio. +4 dB and -10 dB without reference don't mean much.
Two commonly used references are: 1V = 0 dBV and 0.7746V = 0 dBu

The nominal international studio level is +4 dBu (1.2V) and is usually balanced. Unbalanced consumer level is usually -10 dBV (0.3V).
The difference is not 14 dB but 11.78 dB, because they use different references.

But this doesn't really matter because any input usually has appropriate headroom to not clip with hotter signals. Some DACs for example output unbalanced 2V.

Btw, headroom in studios is more or less standardized so that you don't need 120V rails.

Hjelmevold's picture

Balanced and unbalanced per se have nothing to say about levels, in fact you can transmit exactly the same unbalanced signal over a balanced line. Balanced is used to reduce interference.

I think the reason why the salesperson felt like he had to "defend" SPL's use of balanced outputs, is simply because consumer equipment usually doesn't have balanced inputs, so buyers would then ask why they'd need to invest in a cable adapter. In my opinion, there's more to gain from running a balanced signal along a long length of cable terminating in a DI box, than you could hope to gain from investing in an expensive unbalanced cable. But I realize that it may be a hot topic, so I won't talk too much about it here.


Btw, headroom in studios is more or less standardized so that you don't need 120V rails.

Yes, I agree. In practice, this is correct. But in the HiFi world, anything that is measurably different, or event marketably different, will be exploited by salespeople. Nevertheless, even if some of the claims made by SPL are pretty far-fetched, they still make high-quality products that sound excellent.