Quote:
As the term crossfeed is used in audio, it can mean any of a number of fundamentally different things so it is not appropriate to lump them all together.
First you have simple "blend", i.e. mixing the left and right channels so that there is less stereo separation. The extreme setting of a blend control is single channel monaural sound. I had a pre-amp that alllowed this and it was of some use on a few old stereo recordings that had extreme separation, eg. voice in one cahnnel, instruments in another. Generally however I found it rarely useful.
The other meanings of crossfeed are based on the proprietary techniques of the designers and do other kinds of black magic, including frequency response alteration as well as some blending. I suspect that much of the appeal of these systems has more to do with the black magic than the blending.
One common claim which I totally disagree with is that by making phones sound more speaker like they will be more realistic. No - they will be more speaker like and that is far from giving a realistic spatial image.
All conventional speakers suffer from inadvertent cross-feed which is simply an artifact of the speaker presentation and which causes each channel to feed both ears with the same signals. In effect your brain is getting hit from 4 signals rather than the 2 in the source. The 2 crossfeed signals are extraneous to the original 2 channel signals and simply degrade the sound.
Headphones by comparison give only one channel to each ear and produce a more accurate spatial image. What they don't do is produce a sense of externalization, rather you get the in-the-ear effect that some complain about. But in other respects the headphone image is much clearer. Accordingly efforts to give phones speaker-like cross-feed are simply wrong in principle, just a way of buggering up the sound.
However I doubt that many commercial crossfeed systems really do provide speaker-like cross-feed which also requires time delays of the cross-fed signal. Most are I guess simply blend, plus frequency tweaking, plus some other voodoo.
From time to time efforts to get rid of speaker cross-feed are tried. Polk made its SDA speakers some years ago. I bought them and still have them because they do a pretty good job of giving a much more precise stereo image. Unfortunately they are no longer made.
http://www.polkaudio.com/forums/showthread.php?t=45468
Here are some other discussions of this issue.
http://news.cnet.com/8301-13645_3-20022412-47.html?tag=mncol;title
http://www.princeton.edu/3D3A/
http://www.freepatentsonline.com/6009178.html
http://kom.aau.dk/group/02gr960/docs/lspkpos02.pdf
http://www.isvr.soton.ac.uk/fdag/vap/html/xtalk.html
I think the problem is that music used to be entirely mixed for listening from speakers, and to this day that remains a primary objective.
Another attempt at nulling out the unwanted signal from the opposite speaker (in addition to the SDA speakers) is Bob Carver's Sonic Holography - Professor Choueiri's system looks like it does the exact same thing, except with software instead of circuitry, and probably with a whole lot more customizability. Anyway, you can get a Sonic Holography processor or preamp pretty cheap these days - about $60 for the C-9 processor or $100-$200 for a C-1 or C-11 preamp - and they work with any speakers.
The basic concept behind all of these systems is to introduce a phase-inverted copy of the opposite channel's signal, delayed by about 0.2 ms (the difference in time it takes for the sound to go to your other ear instead). The goal is to cancel out the unwanted signal from the other speaker. Of course, because the other ear
also hears the cancellation signal, it's not a perfect solution. I'd be willing to bet that neither is Choueiri's.
Oh, and for the record, this type of processing could be put in the recording, but it isn't because listening to it through headphones is entirely unnatural.
The effect is interesting, to say the least.
It isn't always an improvement - while the soundstage becomes absolutely huge in quite a few recordings, and it's crazy to hear sounds around and behind you (I tricked a friend into thinking I had a 5 channel SACD setup) - it's not always very natural sounding. For example, the female backup vocalists on Clapton's
Lay Down Sally from
Slowhand sound like they're behind you... It sounds cool, but it's a novelty that wears off quickly.
It also has a tendency to diffuse the soundstage quite a bit - vocalists and other precisely located instruments become slightly more diffuse in location.
But, it remains that sound coming from headphones, a source so close to your ear, don't fully mimic our ears' and brains' accustomation to hearing far away sound sources in 3D space. Of course, speakers don't either... They both have their failings in this respect, and since they fail in different manners, it's impossible to account for both in the recording.
So yes, I do use crossfeed much of the time in Winamp. The "HeadPlug MKII" plugin actually does a really good job, with adjustable amounts of crossfeed, delay, treble control, and more.
No, it's not perfect - some recordings don't sound nearly as dynamic with it on. But for listening to early stereo stuff it's a godsend - I can't listen to Cream without it! For more modern mixes, it does do a good job of providing a sense of distance and bringing the sounds more in front of you, like at a live show.
But like I said, sometimes I turn it off as it doesn't always sound better.