Which equipment specs are responsible for separation, soundstage and imaging?
Soundstage and imaging are practically the same thing. An audio system will
image (verb) instruments laid out on a
soundstage (noun) in front of the listener to varying degrees of success. Some people just use "soundstage" as a verb, ie, soundstaging.
Separation is one spec that helps....on speaker systems. Even there it isn't as simple. You can use separate DACs on each channels on the CDP or DAC, and use a dual mono amplifier, physically separating both channels even before they're converted to analogue, but that won't guarantee good imaging since the speakers and the room are also part of the equation. Some speakers have a dispersion patterns and response that kind of suck; for example having a bump in the upper bass sounds good tonally since it gets your feet tapping a lot more to eye popping percussion exploding all over, but the key word there is, "all over," since in such cases there's a tendency to push the percussion forward instead to the back where the drummer is relative to the rest of the band. All rooms have some kind of acoustic issue that can hamper listening: reflections aren't uniform throughout the range, so in some cases higher frequencies can bounce around more (my room was worse with the concrete exterior wall on one side and wood interior wall on the other), so you can end up with the cymbals farther out to the flanks with the guitars and not with the rest of the percussion (my Wharfedale Pi10 has tweeters on swivel mounts, so like in my car, I just gave the tweeters a
lot more toe-in).
You don't have the room as a factor on headphones but that doesn't make it simpler. Your first problem there is that short of binaural recordings each ear only hears one channel, and most recordings were mastered on studio monitors and designed to be heard as such, ie, two ears hearing two speakers in-room. DSPs try to get around this by doing the exact opposite of that, ie, Crossfeed selectively filtering a range of frequencies across both channels. Just like how EQ will boost all frequencies instead of just the one instrument that was screwed up (ex if you listen to power metal, try applying EQ to fix the drums on Nightwish's first album that sounds like all the drums are snares stuffed with pillows; on their ~10th anniversary box set this one was released as a remastered edition), Crossfeed is still imprecise, while binaural recordings are already mastered with all sounds filtered across both channels to simulate both ears hearing both drivers (just without reverb like how virtual surround is done; you get a more consistent soundstage, but not necessarily a gigantic soundstage in front of you that sounds exactly like speakers).
I have an idea that it's maybe a combination of low distortion and high dynamics, but I really have no idea.
Those can help but more in tonality and dynamic range. if these help with imaging, it's secondary.
Lower distortion means you get less of an effect where the upper bass boost of most headphones can be made worse by an amp that boosts all the bass and pushes everything towards the front, so you get less of a sense of having the bass drum at around the same point on the Z-axis as the vocals for example.
High current and ergo high dynamics...or rather,
preserved dynamics...just allows you to hear more of what's in there, so in the same example above, you can hear the bass drum more without screwing with the response, but just giving the headphone enough power to reproduce that sudden kick.
This will not necessarily prevent the cymbals from being imaged to the flanks or giving you a trident shaped soundstage where you get a strong left, weak sounds here (if any), strong center, weak sounds here (if any), strong right.
Crossfeed can do a lot more to make for a more coherent soundstage provided the amp being used isn't absolutely problematic. Otherwise it's very hard to pin down exactly how each measurement translates to soundstage. I mean, some tube amps (speakers or headphones) can have ridiculously high THD and high noise on lower impedance loads, and yet the imaging is usually wider and deeper, with Crossfeed if applied usually doing more of making everything more coherent than the miniscule amount of depth that they add.
The reason it's on my mind is because I just got my first desktop amp rig (Schiit Modi3/Magni3) and I wasn’t really expecting much difference compared to my JDS C5 portable amp. But wow am I pleasantly surprised about just how much bigger the sound, wider the space and greater the presence of fine details. I’m hearing new things in some of my favorite & most familiar music – I never expected this.
My setup for about the last 4 years has been a FiiO X3 --> C5 --> HD600. The first time I really experienced decent space & separation was when I got my HD600's, which were my first open backs and a huge improvement over my portable headphones (for soundstage & separation especially). Now with the Schiit stack, the main thing I notice (and the most thrilling thing) is another step in this direction (not to mention as the extra oomph it gives my HD600's). I think I have determined that the majority of these improvements comes from the amp stage, as I am having a hard time hearing much difference between the X3 and Modi's DACs. I have really loved the C5 -- the clarity, the refinement -- but this Schiit stack has really opened my eyes!
Now, I also have JDS Atom on the way to compare against the Magni 3. I know it measures even lower distortion (we’ll see whether this is an audible difference), so I’m most curious about any improvements I might perceive in instrument separation and the illusion of 3D space.
One other factor there is the listener's perception. Throughout this forum you'll hear people talk about very wide soundstage...and they're talking about Grados, because their idea of wide soundstage is make the drummer look like he's Reed Richards using his stretchy arms to hit cymbals standing where the two guitars are (as opposed to also pushing the guitars farther out to the flanks than the cymbals), and a system with these would tend to have lower distortion and noise (assuming everything else is the same and not using an amp that can't handle 32ohms properly). If this is what you're hearing it isn't wider soundstage, you're just able to go louder without obviously bad distortion, but otherwise the cymbals getting set that far apart is not necessarily "wider soundstage."
