For me, a minimum for experiencing a large "soundstage" is for the headphones to have large earpads such that absolutely nothing is touching your ears. The entry level for some of the largest pads out there is the HiFiMan Edition XS, though that would run you around 400 Euros; higher models mainly afford better build, maybe nicer aesthetics, better sub-bass extension, and likely better distortion performance and driver matching. The rest regarding "soundstage" would to me constitute artifacts of frequency response, a dip around 1 kHz to 3 kHz tending to be interpreted by many as "width" or "space", while for me, that just dulls the sound of some instruments. My preferred method of attaining more "spaciousness" in headphones is to listen to sine sweeps and pink noise and use EQ to smooth out peaks or tame any frequency bands that seem to be getting too loud; I use Equalizer APO for this. Treble peaks for me tend to amplify noise or present a hiss one may not realize is there, detracting from the spaciousness. Midrange bloat can also detract from clarity and spaciousness. Otherwise, I have never really noticed any actual differences in
image size or span between headphones, the sound sources seeming to occupy the same space, the headphone earpads and drivers simply presenting differently sized windows for "viewing" that space.
I also recommend crossfeed for creating a better forward image or taming sound sources that seem to be panned too far to the left or right. I personally use
foobar2000 and the
bs2b crossfeed plugin while using
VB-Cable and
foo_record to feed my PC's audio through that crossfeed (intimidating at first, but it was quite worth it). One thing I have found is that folks probably hype soundstage "width" too much when rather in live concerts, most of the sonic image is perceived right in front of you within the span of the orchestra or musicians, the only real "width" to be considered being the sensation of a
void of sound (other than some reflected sound) to the sides. For me, that "void" can be achieved with the help of bs2b as well as EQing down the treble peaks. I've personally found that basic crossfeed filters struggle with properly bringing higher-frequency content forward. I had gone as far as to purchase speakers and a kit from
Earfish to measure my HRTF (basically the frequency response and timing relations your ears receive from sound sources from every direction) with in-ear microphones, my plugging that into the SPARTA AmbiRoomSim and AmbiBIN or Binauraliser NF VST plugins in the Reaper DAW; generic HRTF files could work, though some imaging or tonal changes may be wrong for your ears. After
EQing my headphones to match my measured responses for an equilateral stereo triangle (post #61) (speakers 30 degrees to the left and right), I can now simulate anechoic free-field (aka. "clean" and nigh perfectly neutral) listening, the head-tracking sensor allowing me to rotate my head toward individual sound sources in the stereo field laid out in front of me between the two virtual sound sources. With this, all sound sources and frequency content are
coherently imaged before you without anything being panned beyond the bounds of the virtual stereo field (unless the mixing engineer used some special trickery). These plugins also support fairly convincing simulation of surround sound. I used Voicemeeter and Rearoute to route up to 8 channels from my PC audio into appropriate settings within the aforementioned SPARTA plugins.
As for EQing each channel to implement better matching of the frequency responses and hence more coherent imaging, I would rather listen to sine sweeps and conduct EQ adjustments to ensure that the pure tones always sound centered, whereby any wandering left and right means that one channel is playing that frequency more loudly or quietly than the other; Equalizer APO's variable-band graphic EQ would allow for easier point-to-point EQing of individual frequencies. MiniDSP EARS represents only one pair of perfectly symmetrical ears, whereby I would rather
acquire a pair of in-ear microphones (post #5,152) and cross-check that against what I actually hear in sine sweeps. As I have
found in practice (post #4,665), EQing headphones based on measurements from the same test rig does not guarantee that those headphones will come to have the same frequency response on your own head and ears.