January 25, 2022
 by Paul McGowan

One of the touted advantages of headphones is their lack of crossfeed. Unlike loudspeakers, the left and right channels are always separate.

And so it surprised me to learn of headphone folks interested in adding crossfeed to their rigs—kind of like adding unwanted distortion.

First a bit of explanation. When listening to loudspeakers we are immersed in what is called crossfeed: the left channel is first heard in the left ear, followed a millisecond later by that same signal getting into the right ear (and vice versa). Thanks to that small delay between ears our brains can sort out the difference so that we hear two separate channels. Headphones eliminate this crossfeed and thus have an entirely different sonic signature: the left hears only the left channel while the right only what is fed to it.

In real life audio (as opposed to our reproduced version of it), crossfeed is natural. Extend your arm as far to the right as it will go then snap your finger. The right ear first receives the snap and a millisecond or so later your left ear gets a slightly modified version. It sounds natural.

One argument in favor of zero crossfeed is that we’re doubling down. First, the microphones pick up all this spatial information, and then through loudspeakers, we’re doing it all over again. Essentially double crossfeed. True enough but it’s worth considering that most recordings are made with a single microphone and then artificially panned without that delay.

The advocates of adding crossfeed to headphones suggest it sounds more natural and I suppose there’s a reasonable argument in favor of that. But, I wonder. Does it sound more natural or does it sound more like the loudspeaker version of natural we’ve all gotten accustomed to?

What’s your opinion?

Subscribe to Paul's Posts

34 comments on “Crossfeed”

  1. Thanks, Paul, for explaining a term I’ve not heard before. I’m not a headphone person and I assumed the extreme left-right mix in early stereo recordings was just how they liked to do things in the 1960s.

    In my experience of live music, the amount of directivity I enjoy depends on the nature of the show, so I suppose it is a sound engineering decision as much as anything.

    1. You are welcome. And yes, my guess about those early hard panned stereo songs was the old “if you have it use it” syndrome new technology is often subjected to.

      Blumlein, the inventor of stereo, would not have approved, but as you know he didn’t survive long enough to see what became of his invention.

      1. Indeed, Blumlein’s concept for stereo was motivated from going to early talking movies and realising that it enabled being able to have two people talking at the same time, which you can’t do on a mono soundtrack.

        Alan Shaw does not like pinpoint imaging, he thinks of sound as waves rather than laser beams of sound. I go with that. We were at the opera on Sunday, a bit too close (10th row) and so you get the instruments coming from different directions. That’s fine, if not desirable, for a string quartet.

        I would hope the crossfeed or blending is done in the mixing studio with sensitivity to the music rather than by toggling a switch on your hifi unit.

        1. It’s not about the recording, it’s how you listen to it, speakers, or headphones.

          The exception would be something like the few albums that were mixed in Q Sound that give you a surround sound experience with speakers.
          Binaural recordings made with a dummy head that give you an “out of your head” sound.

          Switch the playback system, Q Sound over headphones, or Binaural over speakers and you don’t get the effect.

          1. In my music room I have two systems, a 2-channel floor standing system and a 6-speaker ceiling system that sends sound out sideways rather than straight down, so has been described as headphones without wearing headphones. As I mentioned last week, it won CEDIA’s new hardware award for 2021, and the sound quality is superb. It’s currently configured as 3:3 stereo. Like stereo came from the cinema, 20 years before it reached home audio, Dolby Atmos will likely move from the cinema to home systems over time. There are already quite a few Dolby Atmos recordings by the world’s leading artists and I’ve listened to a few (on Amazon HD).

            Of course the bigger market is for home cinema, which I also have installed with the same sysem. This works well by Airplay via AppleTV. It’s pretty poor using Bluetooth direct from a TV – the sound quality is not good enough, but that will improve with the new Qualcom Snapdragon chips.

            So I suspect that a combination of home cinema, better Bluetooth, Airplay and better small active wireless speakers will drive an immersive sound market which, as I said above, is like headphones without the headphones. Crossfeed will then be a thing of the past.

        2. Steven,
          Unless you & I have different definitions of ‘pinpoint imaging’ I find it interesting that you say that Alan Shaw “does not like” that because
          when I owned my Harbeths back in the mid-nineties that’s exactly what
          I got from the home-audio set-up that included the HL Compact’s.
          The musicians & singer held their positions within the holographic soundstage & the drummer could be heard beating his drum-kit
          behind all of them…that to me is ‘pinpoint imaging’.

  2. Try as I might, over the decades, I just can’t enjoy listening to canned music through cans.
    I’ve even built a bass engine chair to add in all of the feelable bass that extends the frequency range down to 8Hz but still, after 4 or 5 songs into an album, I start to lose interest in the whole listening experience & I have to unplug & go back to listening through my loudspeakers, which is not always possible at 02:30 hours.
    I’ve always suspected that there’s some sort of aural claustrophobia going on with me & cans…maybe it is, or maybe, for me, it’s an unnaturalness in the sound-field brought about by Paul’s cross-feed, or lack of it, explanation above.

    Album of the day:- Dead Can Dance – ‘In Concert’ (2013)

  3. If only „most recordings were made with a single microphone“! I don’t think so to be honest.

    I think we repeatedly get surprised how theoretical considerations and seemingly best practices or best technical solutions don’t lead to the best sonic result (yet). Although I have nothing against solid state electronics and digital (and certainly use it, too), those examples are always in our mind when we get surprised which vintage concepts still seem to beat them in their final outcome in several regards (at least for many).

    I think playback (headphones or speakers) in general follows different rules than the theoretically optimal, it follows the rules of the experienced optimal. It is so imo, because the preconditions within the whole recording and media production as well as playback processes and gear are too suboptimal and unperfect to assume theoretically perfect follow up processes and gear would lead to a most realistic result.

    Means: if double crossfeed sounds better, use it, even if it’s suboptimal theoretically.

    IMO theoretically optimal processes within the playback chain will start to provide best overall results only as soon as everything before (e.g. recording) and after (speakers or headphones) are near perfect. I guess everyone will have an opinion if this precondition will ever be fulfilled or not 😉

    Those who insist in going the theoretically best path, do great research work for the whole community and the technology they support, but imo during that time, will rarely have best sonic results yet. But sometimes they anyway have to sell it like that to promote it.

  4. Todays topic reminds me of a YouTube video by Kem Ishiwata, I have looked this morning to many to look through so you’ll have to take my word.
    In the video Ken said headphones are a left and right mono channels and as Paul said above crossfeed is the way we hear everything, why should stereo from two loudspeakers be any different.
    I’m sticking with what I have two channel with crossfeed.

  5. I think it depends on the genre. Popular is skewed to headphones because most listeners use their phones for listening. Jazz/Orchestral
    /Classical skewed towards speakers.

  6. Microphones may well pick up all that spatial information, which I believe is a good thing, but they don’t have two ears, they are a point receiver, so is it really crossfeed, or am I missing something?

    On the other hand, if double crossfeed is a problem why don’t they eliminate it in the recording process?

  7. I have been a headphones person having owned more than 50 models of headphones and more than a dozen dedicated headphone amps. At the moment, I’m looking at 5 pairs and a Violectric amp I am using.
    It is much cheaper and easier to achieve a high fidelity playback on cans and there is no substitute for privacy in listening, especially on business trips.
    Owning several Meier models, I had a chance to listen and compare crossfeed with ordinary stereo on good headphones.
    I tend to agree with Paul. Crossfeed is unnatural though sometimes better sounding. It does artificially expose a listener to sound non-existing on the master recording that is rarely if ever monitored on crossfeed.
    It is a nice to have feature but by no means essential. I would say, as many other things in life – an acquired taste.

  8. I listen to headphones a lot more than my speaker system and they sound great. Chord Hugo2 > ZMF Pendant amp (made by Ampsandsound), ZMF Verite Closed headphones and HifiMan HE1000V2 Planar headphones.

    I don’t feel it’s distortion, rather, it’s recreating what you would have heard with speakers. Signal is phase reversed, slightly delayed and lowered in volume, then fed to the opposite channel.

    Crossfeed on the Hugo is normally off. If it’s on when you don’t need it, the stage width is a little narrower. However, on some recordings where something is hard panned to one channel, it can feel very un natural, and even a bit uncomfortable, like someone vacuumed all the sound out of one side. I wish I could think of an example to give you.

    Basically, when you need crossfeed, you need it. Otherwise it’s off.

  9. In my alleged mind adding cross-feed to cans is just an effort to make them sound more like we’re used to when listening to speakers. I remember some headphone amps having a switch that would add cross-feed by throwing a switch.

    I listen to cans (Sennheiser HD6XX) occasionally and the fist thing I notice is the complete absence of my listening room. I gain more detail but it sounds a bit off.

  10. I do 90% of my listening with streamed Apple Music (now lossless) through my iPad Pro & AirPods Max over the ear headphones. While Bluetooth has some loss to it, it’s easily the most satisfying headphone experience I have had. Equally satisfying is the AirPods Max’s Spacial Audio presentation of streaming Atmos movie sound tracks through the latest AppleTV.

    Previously, I would have been in the purist camp of no crosstalk. The AirPods Max experience has changed things. Stereos are, in effect, processing sound to recreate an experience. We just aren’t used to the possibility of software doing the same thing.

  11. As far as I’m concerned headphones need something. I just don’t appreciate headphones the way I do a two channel system. I want to feel the music in the room. Headphones serve one purpose. They prevent the music from bothering anyone else. Good late at night when people are trying to sleep.

  12. I do enjoy Speakers and also my HEADPHONES.
    Talking about cross-feed??
    The Speakers are 4 wired systems….two for Left and Two for Right.
    The Headphones are 3 wired systems.
    ONE + common for LEFT and
    ONE + common for RIGHT.
    Does this not introduce CROSSFEED or Distortion??
    Do clarify.

  13. Headphones for me are for travelling, I can enjoy my Fiio and AKGs for many hours on a plane. Listening at home is all loudspeakers, all the time.
    From what I have read, crossfeed is for mixing and mastering on headpones, so you can hear what it would sound like played over speakers. The idea is to emulate the sound of speakers; helpfull when creating the music, so you know how it will sound reproduced. But listening to recorded music as playback with crossfeed, yes you are making it sound more like it’s being played back on a loudspeaker system rather than hearing the music unadultered. We could be so used to hearing reproduced music rather than live music that reproduced sounds more natural! Maybe kind of the opposite to listening to binaural music on speakers?

  14. Thanks Paul. Like it when you address controversial subjects. I wonder if
    The poor crosstalk performance of phono cartridge compared to digital audio does not contribute to the difference in perception between vinyl and cd.

  15. I had the PSB headphones designed to sound like speakers in the room. They were OK but still not like speakers. Open ear headphones might be the best way to duplicate room sound with headphones. I like Beyer Dynamic DT990 and Sennheiser HD600 and 650 open ear headphones. Also B&W P7 closed ear headphones sound really good. Lifelike and musical. Some of the vintage Sennheiser’s from the 70’s are worth buying. HD414 and 424. They have a unique seductive sound.

  16. In real life we do not live in an anechoic chamber, so in addition to direct crossfeed from our speakers and other sound sources we get indirect crosstalk reflected off room surfaces all around us. Too little crosstalk sounds unnatural. Too much destroys the stereoscopic image. With headphones we do not get room reflections. Our brains expect to hear a certain amount of crossfeed because that’s how we typically hear sounds in rooms and even outdoors nature.

  17. I find the crossfeed feature to be similar to Carver’s sonic holography button. In both cases I get the ‘hmm, that’s kind of cool’ listening experience, but when I turn it off after even a short sampling, I get the ‘oh that’s better’. My headphone DAC has variable levels of crossfeed, it’s fun to play with once or twice, but it’s certainly not a listening tool. I’ve always felt that processors like crossfeed, sonic holography, subharmonic synthesizers, reverb, dynamic range processors, noise suppressors, impulse noise eliminators … (& I’ve had em all!) can certainly bandaid a lower quality or poorly set up system but I find they just degrade & mud up any high quality system. With the exception of digital EQ. Way too many poorly mastered recordings out there that I need to doctor up.
    Where’s the kick drum?!? Why are the cymbals & hi-hat so weak & muffled? BAD engineer – NO cookie!
    And back when vinyl my only available hifi source, my DBX 3BX & transient tick and pop eliminator were absolutely necessary. Y’know, to make that 1800’s technology music source listenable….as I turned on my subsonic filter and delicately tiptoed about the house and tried not to DROP anything on the floor that would subsequently grenade a woofer across the room…

    1. The intent of the sonic hologram function was to mask crosstalk but the effect can be heard on headphones. I will admit to using it with my basement speakers when listening to FM radio.

  18. I thought the idea of crossfeed for headphones was to make the sound seem less ‘in the head’ and therefore more natural. Is that not the case?

    In any event, like many posters here, I prefer to listen to music reproduction via speakers as a) I like to feel the bass and b) I find headphones oppressive after a while.

    But, different strokes for different folks. If you enjoy it, then great!

  19. I decided to upgrade my Bifrost/Asgard rig with the RME ADI-2 Pro FS R. It offered amazing new capabilities including A-to-D conversion, DSD recording & playback, DSP parametric equalization, digital format conversion, balanced headphone outputs and headphone crossfeed. A year later most of those new functions are rarely used and the Bifrost multibit DAC is still serving in my system (and I am still thinking of upgrading it) but I always turn to the RME for the crossfeed when listening to headphones. I find listening with some crossfeed applied is more natural and more relaxing. I actually tolerate listening via phones for a longer period of time when crossfeed is applied.

    I reason that most recordings are engineered monitoring with speakers and for speakers, and that crossfeed restores the imaging to more closely resemble what the engineer intended.

    1. I’ve the ADI-2 FS non-pro version. (fed to a Audio-GD Master 9 or Liquid Platinum.) You’ve convinced me to revisit the crossfeed feature for more experimentation.
      This is what I love about this forum.
      As far as headphones vs speaker audio, I enjoy both. However long headphone sessions nearly always graduate into the soundroom…
      I’ve got a few sets, but for me the Arya is unbeatable.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram