Soundstage is not about sound "quality" (i.e. frequency response), it's exactly about 3D audio.
In real life, when there's a sound coming from the left, you hear it with both ears. You just hear it with the left ear first and with higher amplitude (heads are good at dampening sound).
With headphones, the sound coming from the left driver never reaches the right ear, as it's never projected with enough intensity to reach the opposite ear.
This means the 3D audio system, when you're using headphones, needs to
fake your head/year dampening to reproduce sounds coming from the left on the right speaker, or it'll sound fake and take away immersion.
This is what HRTF (head-related transfer function) does in a game engine, and it's usually done using standardized virtual heads that never correspond entirely to each person's head/ears. That's why Cerny mentioned taking pictures of our heads and ears to further optimize the 3D sound and customize a HRTF map for each person. However, even with pictures of our heads they still wouldn't know how the interior of our heads in arranged (e.g. bone density of the skull, geometry of the brain, etc). They'd need to place high-end microfones inside our ears and do several listening sessions to reach something like what they achieved for Cerny:
With external surround speakers, you don't need to fake sounds through HRTF. A sound coming from the left speaker is always going to reach your right ear, and with the "perfect processing" because you're listening it from a "real" source.
HRTF can still happen to emulate an exact placement of a sound source between two speakers, but with much lighter differences. Height is also something that needs HRTF on surround speakers, but not when you have height speakers (though notice how Cerny's HRTF maps aren't even considering height at all).
In summary,
@London Geezer 's setup with 1 center + 2 front + 2 side + 2 back speakers + 2 height speakers + subwoofer is always going to be worlds better for 3D audio than headphones. It's not even a fair comparison. It wouldn't be a fair comparison even with 5.1 vs. headphones.
The article you referenced is comparing stereo speakers with headphones, not a full 7.1.2 setup.