Received the fixed Erco 2 on Saturday. Fast 2-day turnaround with the distributor (I sent it back out on Thursday), so props to them on that. Listened to it yesterday and my impressions are positive.
My listening chain was the Erco 2 acting as a DAC with XLR line output feeding a CCS-modded Stax SRM-006tS driving the Stax SR-X9000. I'm using the Erco's stock PSU, but with the Shunyata Delta v2 NR power cable connecting the PSU to the wall. Now, I think the Delta makes a difference to the sound (more than the Hypsos), but I also used the Delta on the Erco before I upgraded to the Gen 2, so the cable should be irrelevant to any differences I hear between gen 1 and gen 2.
The two most significant changes vs the Gen 1 are the reduction in "hash" or "haze" in the treble and more precise imaging and localization of sounds. The two combine to produce a more pronounced sense of detail, but I don't think it has more absolute detail, just that it's easier to pick up on details on the Erco 2.
The reduction of hash in the treble has the dual effects of reducing perceived overall treble quantity as well as increasing the perception of mids and bass. To be clear, there is no reduction of treble quality or presence of notes. It's not like a shelving EQ which reduces quantity without changing its texture or quality. The decrease in hash comes from the improved antialiasing filter (HQ Apodizing) which has full attenuation before the Nyquist frequency. Thus, there's no aliasing to contribute to the "hash" in the upper treble. For me, this improvement is most obvious with hi-hats and cymbals. On many DACs, including the FiiO K9 Pro I used for comparison (since I don't have the Erco gen 1 for direct comparison), there's a slight smearing of hi-hat hits on either side of the transient, so instead of a clean transient like "TSH" it sounds more like "sTSHss"; there's this slight sizzle before and after the hit. This gives the hi-hats a "shimmering" character which can sound softer or sifted (think of sifting sugar or salt and that soft "shshshsh" sound it makes), but it also smears the sound. The Erco 2 does not do that. Cymbals and hi-hats are crisp and clear in attack and decay. Because there's no extra shimmer, the upper treble sounds less busy, so other parts of the sound seem easier to hear at the same time, but the treble is no less detailed and it is often easier to separate different notes from each other.
This is really helpful for the EDM that I enjoy because there's often a layer of noise in the track that adds energy to the sound and the Erco 2 lets the noise add the energy without letting it pollute the rest of the track. Songs I used for this were "Worst Day" by Illenium and MAX, "Love Is A Highway" by Nurko and NERIAH, and "Rest of My Life" by Culture Code and Medyk. All of these tracks have that layer of white-ish noise during the drop alongside synths, hi-hats, electronic snares, and more, and the cleaner treble on the Erco 2 allows elements of the music to remain crisp and clear during the busy portions. In particular, the soft background vocal echoes in the drop of "Love Is A Highway" are distinct on the Erco 2 vs other DACs where they're harder to hear. The cleaner treble also improves mids and bass (or more accurately, notes/instruments that lean on those frequency bands) because better treble cleans up the transients for those notes. So it seems like mids and bass are more prominent, but they aren't boosted, they're just cleaner and unimpeded by treble hash and smeared transients.
This reduction in treble hash is something that you can get with apodizing filters in HQPlayer. I experimented with HQPlayer for a bit prior to getting the Erco Gen 2 and I heard there a reduction in hash and a perceived increase in bass prominence as well as better perceived detail. Since the Erco 2's (and Wandla's) HQ Apodizing filter was developed by the guy who makes HQPlayer, it makes sense that this filter does similar things.
The more precise imaging is also related to that absence of haze and smearing. This time, it's like the lack of smear in the spatial rendering of a note. Think of looking at a light through frosted glass; the glass will diffuse the light around it and the edges becomes hazy while the light source appears larger than it is. If you look at the light without the glass, the edges of the light source are better defined. The increased spatial definition makes it easier for me to "lock on" to sounds, so it increases the perception of detail. Now, if I were to A/B compare vs another DAC, I can hear all of those same details, but on the other DAC, the sound isn't as well defined in space, more like a blob of sound, so it is harder for me to lock on to it unless I already knew it was there.
Fortunately, the Erco 2 retains the punchy and dynamic sound of the original Erco. The slight emphasis to the leading edges of notes makes music "pop", like each pluck of a string or smack of a drum stick is momentarily highlighted vs the trailing decay. I found this quality to be very engaging on the original Erco and I'm glad to see that the Gen 2 refinements haven't removed this aspect of the sound. I suppose it's not as punch-emphasized as the Gen 1, but mainly because you can hear better texture in bass due to the cleanup of the treble, so that improved texture competes with the punch for your attention. The punch didn't get worse, everything else just got better.
Listening to the X9000 with the Erco 2 was probably the best I've heard the X9000 at home. I did not really enjoy the X9000 with the original Erco; the original was too forward and when combined with the midrange emphasis of the X9000 it sounded like the mouths of vocalists were stretched out between my eyeballs - they were rendered too wide and too forward for my tastes. And that resulted in "soundstage collapse" where it sounded like the music had no frontal depth, just a flat sheet between my ears. Even though the K9 Pro was less dynamic, I often preferred the X9000 on that DAC since it wasn't so forward that the soundstage collapsed. The more precise imaging of the Erco 2 fixes that problem. The midrange is still pretty forward on the X9000, but vocalists are now better defined in space - not as wide - and I can get the spatial layering that I heard when I demoed the X9000 on the Chord Hugo TT2. The difference in treble quality from the Erco 2 was also easily apparent on the X9000.
Anyways, I like the Erco Gen 2; I think it's a noticeable and real upgrade (rather than just a different flavor) from the Gen 1. I can't wait to listen more.