Lily Audio Genesis One Headphone Review

[dagot23]

It feels like, with maybe the exception of Dan Clark and Audeze, average sound quality falls off a cliff above $500. Embarrassing how many boutique $1,000+ headphones are dreadful.

The more expensive something is, the higher the chance the average audiophool will buy it since "bigger price=better sound" or "the technicalities though". That's how Abyss operates, for example. Buy comically expensive headphones + "oooh you have to use our $1000 cables and $5000 amp for it to sound good" and some people believe it. Dan measures his stuff and makes sure it sounds good and measures good. And Audeze - a lot of mixing engineers buy their stuff so they have to make sure it doesn't distort and, despite the weird stock FR, is at least easy to correct with EQ. To their credit, they corrected LCD-X's FR not long after Amir reviewed the older model and sent him the new one to measure, so it's clear they care about that stuff. But it's kind of the same way with IEMs. You have lots of these $300+ IEMs that measure far worse than a $20 Chu 2 or a $50 Zero Red, for example. A lot of those "premium" headphone/IEM companies with products that cost a few grand don't care about the measurements, only about milking the audiophool for all they can. It's kind of sad. That part about the cables is true, by the way. I remember they made a series about the headphones they make and one video was about cables and they said to buy their expensive cables for better sound.

 

[tonyng]

Hi Amir and the ASR community. It's Tony from Lily Audio here. First, a sincere thank you for your analysis of the Genesis One. We value detailed objective feedback. I would like to clarify that the measured unit belongs to our very first production run. Since that initial batch, we have made tuning revisions to improve the ear gain response. Specifically, as of the 2nd batch, we added internal damping to the driver to tame the treble peaks (ref: KS update #16). We are supporting all original backers by retrofitting this fix.
Given that the product has evolved, I would love to send you a unit that reflects our current tuning for updated measurements. Thanks!

 

[amirm]

Hello Tony. Welcome to ASR! And thanks for the offer. Would be happy to test a new sample. Please start a "conversation" with me and we can organize the shipment.

 

[tonyng]

I personally think it's wild that they'd ship a unit with 4-12dB channel imbalance between 2-6 Khz, that's basically guaranteed weird sound and brutalized stereo image, if you get any at all.

I totally get why that looks wild. Seeing a 4–12dB variance on a graph in that 2–6 kHz region is definitely alarming, and on a normal headphone, it absolutely would wreck the stereo image. But the Genesis One is a bit of a special case due to its design. It has massive ear cups designed specifically for maximum openness and soundstage. The catch is that huge ear cups almost always measure pretty ugly on graphs. Measuring treble is tricky for us because of all the reflections and cancellations inside a large cup and the measurement rig. If the headphone shifts on the dummy head, the treble response curve changes drastically. This isn't really an issue with smaller headphones.
To get a "true" treble measurement, you have to reposition the headphones multiple times and average the graphs out. Unfortunately doing this for every single unit on the production line is not feasible. Because treble measurements on rigs vary a lot, we focus our strict channel matching on the bass and mids, where the numbers are stable. For the treble, we check the general trend line rather than stressing over individual peaks and dips. Hope that sheds some light on why the graphs look so chaotic compared to how they actually sound!

 

[solderdude]

I would suggest to add a series resistor in one of the wires going to the driver inside the headphone.
The minimum value should be 47ohm but perhaps 100ohm would be better.
This will prevent some amplifiers from oscillating (amplifiers don't like to see a 160nF capacitance directly on a low ohmic output) by destabilizing the feedback loop.
Any amplifier will 'see' the headphone WITH a series resistor as a 'Boucherot cell' instead of a pure capacitor which is a much better load.
Also some audio(phools) like to use a filterless DAC and have high amounts of 44kHz (at which point the impedance is 25ohm but the output current is out of phase.
As a bonus it will tone down the upper treble.
I'm pretty sure an optimal value can be found that brings down the boosted upper treble to 'better' levels.

I would suggest experimenting with resistors.
Also in the specs it might be an idea to list the max. voltage swing is as ceramics have a max. voltage.

 

[tonyng]

I would suggest to add a series resistor in one of the wires going to the driver inside the headphone.
The minimum value should be 47ohm but perhaps 100ohm would be better.
This will prevent some amplifiers from oscillating (amplifiers don't like to see a 160nF capacitance directly on a low ohmic output) by destabilizing the feedback loop.
Any amplifier will 'see' the headphone WITH a series resistor as a 'Boucherot cell' instead of a pure capacitor which is a much better load.
Also some audio(phools) like to use a filterless DAC and have high amounts of 44kHz (at which point the impedance is 25ohm but the output current is out of phase.
As a bonus it will tone down the upper treble.
I'm pretty sure an optimal value can be found that brings down the boosted upper treble to 'better' levels.

I would suggest experimenting with resistors.
Also in the specs it might be an idea to find out what the max. voltage swing is as ceramics have a max. voltage.

Thanks so much for the awesome feedback and suggestion. This aligns perfectly with what we have planned. If you check out our Kickstarter Update #14 (link), you'll see we're actually on the exact same page. We're releasing an accessory in about two weeks that lets users swap between different resistor values ranging from 50 to 300 ohms.
Just like you mentioned, it's going to be a great tuning tool to help roll off that upper treble for people who prefer a more natural, gradual roll-off like you get with dynamic drivers. Plus, it gives that extra bit of stability for finicky amps. Really appreciate you sharing these technical insights, it's awesome validation for what we're building!

 

[solderdude]

I would suggest to put 47 ohm IN the headphone and add 1 or 2 'extra' resistors for sound tuning.
This way the headphone when used by itself as is can't destabilize some amps, 160nF is a rather high capacitive load.
There's going to be some folks using it without series resistors and probably drive it with esoteric amps that may get damaged.

 

[shuppatsu]

I really appreciate the developer coming in with the positivity. It would be easy to get their back up and see the negative aspects of the review and the somewhat negative response and write off the whole thing.

This headphone is trying to do something very different. It’s not going to be for everyone and probably shouldn’t be recommended as a general purpose headphone. But I can see the appeal for some and it’s cool so long as people know what they’re getting into. And I’m happy the manufacturer isn’t running away from that.

 

[solderdude]

@tonyng

The schematic above fixes the 2.5kHz peaking and has the adjustable treble selector.
Will fit in your box (just looked at the kickstarter post).

It lowers the impedance above 1kHz to roughly 100ohm.

Inductor must be able to handle 100mA or so, low resistance is not needed at all so can be small in size for instance BOURNS RLB9012-472KL
Cap should be rated > 50V (say... KEMET R82EC3470Z370J)
Resistors R1 and R2 around 1W.

 

[torgeirs]

But it might destroy the distortion spec of the amp. So maybe a special amp compencated for this load is needed.

https://www.ti.com/lit/an/slyt630/slyt630.pdf

For some context on capacity load and stability https://www.ti.com/video/series/precision-labs/ti-precision-labs-op-amps.html
A voltage step with headphones connected is maybe the easiest check for stability of the amp with the headphones connected. So scope mode measurements.
A wide bandwith sweep (maybe up to Mhz) should also reveal instability. A HF peak bigger than 3 dB should discourage use of the amp for those headphones

 

[solderdude]

But it might destroy the distortion spec of the amp. So maybe a special amp compencated for this load is needed.

Not with a series resistor nor with the proposed filter.

 

[torgeirs]

Just a random example og different Riso, OPA1612 and 150nano fahrad load

 

[torgeirs]

From the TI pdf

 

[torgeirs]

But much better with a little distortion than unstability!

 

[solderdude]

exactly my point..
Besides even if the distortion would increase a tiny bit (because of the capacitive driver) that will be dwarfed by the relative substantial distortion of the headphone itself.

 

[torgeirs]

Totally agree. But on this site they measure sub 120dB distortion into a resistive load and put it in a who is best chart for headphone amps

This is some the effect that is causing distortion in a resistor - transducer combination. So the voltage devider, Riso and Ztransduscer is unlinear, that is different at large execursion than at low execursion. And it differs with frequency.

 

[torgeirs]

Just an distortion effect not that many are aware of

 

[theBoomer]

The only purpose for this to exist is so audiophiles can brag about how difficult their prized headphones are to drive when talking about others’ headphone amps. Oh yours could never push MY reference cans to proper listening levels lulz.

Paraphrasing someone who was speaking about standards, the great thing about reference cans is that there are so many you can choose from.

 

[FTB]

Again some 1.2k$ headphones that measure like crap.
What a joke from these brands.

 

[solderdude]

Tough crowd