The Miracle of EQ, or "How I Learned to Stop Worrying and Love my LCD2C Again"

audeze
eq

#1

Full Disclosure - I may be biased by my lovely wife recently gifting me this custom-made wood and glass display case for my LCD2C. Still, I think EQ had a positive influence as well!

Background

2 months ago I reviewed the Beyerdynamic DT 1990 Pro in comparison to the Sennheiser HD 600 and the Audeze LCD2C. As a result of that review, I developed some impressions that stuck with me until recently:

  1. I generally prefer the DT 1990 (EQ’d) as a good all-rounder

  2. The LCD2C while fast on the transients has a wetness that hurts overall clarity and makes it suboptimal for orchestral and metal music. That wetness does make it very smooth and adds a euphony to certain music like choral works. Its tonality is warm bordering on dark but still natural sounding because of the smooth frequency response and not oppressively dark because of the bump in the treble.

  3. The HD 600 is a solid set that nevertheless falls behind both the DT 1990 and the LCD2C because of more congested imaging on busy passages and muddier bass. Not mentioned in that review, I’ve since noticed a pervasive grain that makes it sound less smooth than the other two (especially the LCD2C) and despite being overall very neutral it suffers from a bit of extra energy in the low treble that can become sibilant (not extreme, but noticeable).

Well, recently I started getting more serious about EQ using my MiniDSP E.A.R.S. It occurred to me that I was a bit unfair to the LCD2C in that I EQ’ed the DT 1990 to correct their flawed treble response but didn’t really help out the LCD2C with similarly substantial EQ. I also didn’t bother to EQ the HD 600 at all both because it sounded so neutral to my ears and because of its reputation for neutrality.

Findings

As a result of my more intensive EQ efforts, the LCD2C now shares the top spot with the DT 1990 as my favorite pair of headphones and I’ve come to appreciate the following about EQ:

  1. Unsurprisingly, EQ works well for modifying tonality and can be used to make any headphones sound “neutral”. In particular, taking a deliberate approach based on measurements of the specific pair of headphones seems to yield generally pleasing results.

  2. Through conventional wisdom I had come to fear over-equalizing because of its impact on distortion. I couldn’t hear any such impact and my measurements of the LCD2C support what I heard. Please note that these measurements were taken at the average levels at which I listen, which are lower than the usual testing levels.

  3. The objectionable aspects of the HD 600 and DT 1990’s treble response are completely resolved by EQ’ing them down.

  4. What I heard as wetness in the LCD2C was actually a combination of real wetness (i.e. reverb) together with a dark tonality that added to my perception of thickness. Fixing the dark tonality with EQ significantly transformed my perception of this wetness to the point where I no longer think of the LCD2C as “too wet”. With this out of the way, I can properly appreciate its smoothness, speed, extended bass and overall very natural sound.

  5. Congestion seems to be noticeably affected by frequency response. The LCD2C, which already separated quite well, has even better instrument separation when EQ’d. Equalizing the HD 600 helped clear up its congestion quite a bit too, though it still doesn’t surpass the other two in this regard. I suspect there’s just something fundamental to it that holds it back, though perhaps with more focused work on this aspect of their performance, I could make even further improvements with EQ.

  6. The HD 600’s graininess was not particularly affected by EQ. Perhaps this is just fundamental to its sound, or perhaps the DT 1990’s lack of grain is due to a dip in the presence region?

  7. The HD 600’s bass response isn’t particularly elevated and so didn’t need tonal correction, I think it just sounds a little muddier because of higher distortion, which I can’t fix with EQ. I did not try to elevate the sub-bass because I feel like the bass is already too muddy and adding more energy there doesn’t seem like a good thing.

  8. The DT 1990 doesn’t sound good to me when matching the Harman response too closely. I don’t know if that’s due my bias of having known and liked it as a slightly trebly headphone in the past, or if the aggressiveness of the EQ required to make it fully neutral interferes too much with its fundamental behavior.

EQ’d, the LCD2C is simultaneously fast but smooth, detailed without being sharp, with deep, textured and punchy bass, a balanced and natural midrange and enough treble to sparkle and shimmer when required by the music, though without ever approaching fatiguing. It’s no soundstage monster, but since I’m addicted to crossfeed I hardly care. The imaging and separation though is very good (though still a tad behind the DT 1990). Like the DT 1990, the EQ’d LCD2C sounds excellent with well mastered music but more than the DT 1990 it maintains its pre-EQ ability to make poor recordings sound good too. I’m very pleased!

LCD2C EQ Specifics

The EQ settings for the LCD2C ended up being very simple, just two peak/dip filters.

  • -2.4dB at 766Hz with Q=1
  • +6.4dB at 4045Hz with Q=1

The notch filter at 766Hz turns down the fullness of the LCD2C’s sound a bit which I generally like, but when I want a little fuller sound I can just toggle this filter off.

The peak filter at 4kHz fills in the upper mids and lower treble even out the overall response, which gets rid of the darkness and brings the LCD2C in line with the HD 600’s neutrality.


Sennheiser HD58X vs Focal Elex Impressions
Audeze LCD-3 over-ear planar magnetic Headphones - Official Thread
#2

crazy guy talking to himself

So, I just picked up Chris Thile’s “Deceiver” and was struck by just how much more I enjoyed the instrumental numbers on the brighter DT 1990 than the eq’d LCD2C, but enjoyed the vocals more on the LCD2C. While the EQ’d LCD2C is plenty bright for me most of the time, on this album Thile’s Mandolin riffs just sounded kind of lifeless, with neither bite nor sparkle.

Generally I’d be fine with picking the right headphones for the right job, but it’s pretty infuriating to feel the need to change headphones from one song to the next on the same album. So, on a whim, I decided to see if I could boost the LCD2C’s highs with about 5 dB of high-shelf filter coming in at around 7300 hZ. Voila! All of a sudden, the mandolin sounds like I expect a mandolin to sound, but when I’m on a vocal number I can just switch back to the more neutral EQ profile.

I draw two conclusions from this experience:

  1. The LCD2C is a chameleon. It takes EQ easily and sounds great doing it.

  2. Hearing music as “the artist intended” doesn’t necessarily mean an uncolored reproduction. I can’t imagine that Chris Thile (a mandolinist) would want the mandolin part to almost disappear in his songs, yet when I listen with a pair of neutrally tuned headphones, that’s exactly what happens. Only when I add some coloration do I finally hear the mandolin in its glory and properly connect with the music on an emotional level.


#3

Something that @Ishcabible mentioned on Reddit planted a seed of doubt about whether I’d properly sorted out the LCD2C’s sound. He said that they sound “plasticky”. I never really thought this before, but whether through the power of auto-suggestion or something else, they started sounding plasticky to me too! So, back to the EQ drawing board I went.

I ended up with the below EQ settings. So far they’ve excised the “plastic” demons and sound very natural and balanced to me. If anyone tries these settings and still finds the sound plasticky, please don’t tell me as I really don’t want to tinker with it anymore :wink:


#4

I just can’t stop equalizing! Something that the original picture missed due to the scale is that there’s a rise from bass into the middle mids. I had been ignoring this and focusing instead on filling in the hole in the high mids and low treble.

My new EQ settings don’t just fill in the hole in the high mids but also flatten out the high bass and low mids. The result? A totally different sounding headphone. Like, it doesn’t sound like an EQ’d LCD2C, it just sounds like its own thing. What’s blowing my mind right now is

  1. How much the frequency response affects the overall sound. Wet? I don’t think so. Imaging? Pretty incredible so far.

  2. How much difference the low to middle mids make to the overall sound signature.

A comparison with my old DT 1990 is instructive.

  • Red - LCD2C
  • Blue - LCD2C Equalized
  • Green - DT 1990

This uses a somewhat compressed scale that exaggerates differences, but with smoothing to keep it from looking too crazy. Before EQ, the LDC2C had similar bass to the DT 1990, much more forward signature around 1kHz, a bit less presence, and considerably darker across the treble.

With EQ, the LCD2C has significantly more bass (with a nice mid-bass hump), isn’t quite as forward as before, has a bit more presence than the DT 1990 and pretty nicely extended treble (without the DT 1990’s peakiness).

After letting the new sound sink in, I think I’m going to re-review the LCD2C with the original song list and see what comes out.


#5

Interesting progression!

Have you compared your EQ curves, or at least the audible results of them, to the results of the Audeze “Reveal” plug-in?

The “Reveal” adjustments are based on averages across multiple headphones, where as your own will be specific to the unit you actually have - as well as having your own “taste” incorporated, but I’d be interested to see how they compared for you.


#6

So, I measured using Roon -> Hifiberry -> Magni 3 (lo gain).

First plain (no EQ) vs reveal:

  • Blue - Plain
  • Dark Purple - Reveal 100% Wet

As suspected, the effects of Reveal are very subtle.

Here including Reveal 100% wet and 50% wet. Not surprisingly, 50% wet splits the difference between plain and 100% wet.

  • Blue - Plain
  • Dark Purple - Reveal 100% Wet
  • Light Purple - Reveal 50% Wet

Now just my EQs (this time matched at 1 kHz instead of 300 Hz)

  • Blue - Plain
  • Green - Forward (my 2nd to last EQ attempt)
  • Red - Clear (my most recent EQ attempt)

The names “Forward” and “Clear” come from the fact that the forward settings sound really nice to me with vocals but perhaps a tad more congested, whereas the clear settings seem good on instrumental stuff like symphonies, but not as good for folk and rock.

Looking at the graph, both sets of EQ’d settings are more similar to each other than to the plain LCD2C. Both have more extended bass, both are a bit less forward (around 800 Hz), both have a touch more clarity (around 2 kHz), both have more presence (3-6 kHz) and both have a bit more mid treble energy (8-10 kHz). The “Clear” EQ has less fullness around 300 Hz and perhaps a tad more presence around 3 kHz, and its mid treble peak is 1 kHz earlier. Once we get to the higher treble all 3 look basically the same. I suspect that the dip at 300 Hz accounts for the main differences in terms of congestion and vocal sweetness.

I think my next step is to try and get rid of that nasty peak at 4.4 kHz.


JPS Labs Abyss (original) and Audeze LCD-4 Comparison
#7

Latest EQ settings here.

  • Red - Original
  • Blue - EQ’d

Here’s the theory behind this tuning.

  1. Better extended bass, because bass is always good.
  2. A bit more forward around 1 kHz to help out voices
  3. More clarity/detail around 2 kHz to help out instruments (I think this might help separation too(?), we’ll see)
  4. More presence in 3 - 5 kHz to even out that region with the midrange (should provide better vocal balance), but less presence than “neutral” because too much presence can sound congested and grainy to me.
  5. A little bit more treble around 6 kHz to improve percussive sounds
  6. Nothing extra around 8 kHz because I’m sensitive to sibilance here
  7. Extra treble starting at around 10 kHz improve attack on percussive instruments and add air, plus I’m older and my high treble hearing is weak anyway :slight_smile:

Put more succinctly, this aims to mostly follow the Harman response with just a smidge more clarity, a bit less presence and a bit more high treble.

So far it’s sounding nice.


#8

One thing I don’t like about the most recent EQ is that it’s complex (11 bands) which, amongst other things, makes me unable to use it in UAPP on Android. So, I took a shot at making a simplified version with only 6 bands.

  514 Hz   +4.00   q  0.30
  750 Hz   +2.50   q  4.00
 4400 Hz   +2.00   q  1.00
 4420 Hz   -5.00   q 16.00
 7250 Hz   -6.00   q 10.00      
14000 Hz   +3.00   q  2.00

Here’s how the simplified filters compare with the complex ones:

It’s pretty similar, with flatter response from bass through most of the mids and a bit more energy in the clarity area at 2 kHz and through the treble (except around 8 kHz to which I’m overly sensitive anyway).


#9

Okay, this is sounding pretty fantastic, but on some stuff like Fleetwood Mac’s Rumours, there just enough treble energy to push things slightly into sibilant territory. Moving the 14 kHz filter to 15 kHz seems to solve it but still maintain air.

  514 Hz   +4.00   q  0.30
  750 Hz   +2.50   q  4.00
 4400 Hz   +2.00   q  1.00
 4420 Hz   -5.00   q 16.00
 7250 Hz   -6.00   q 10.00      
15000 Hz   +3.00   q  2.00


#10

Well, it turns out that too much bass seems to hurt imaging! I like the rumble and all that, but compared to my old DT 1990’s it was just too thick. I’ve reduced the bass level a little bit and imaging/separation seem much better. My benchmark for testing separation is a Yuja Wang recording of Rachmaninov’s 3rd Piano Concerto–there’s a crescendo starting at around 7:30 into the first movement that peaks at around 8:30, with furious piano playing, blaring horns, soaring strings and banging timpani that easily turns into a wall of sound. The DT 1990 always handle it like a champ and while the stock LCD2C do pretty well here, they’re no match for teh DT 1990. With the current EQ settings, they do much better.

Here are the latest settings.

   20 Hz   -3.00   q  0.30
  514 Hz   +4.00   q  0.30
  750 Hz   +2.50   q  4.00
 4400 Hz   +2.00   q  1.00
 4420 Hz   -5.00   q 16.00
 7250 Hz   -6.00   q 10.00      
15000 Hz   +3.00   q  2.00

I’m listening to Tool’s Lateralus with these and enjoying it immensely. I hadn’t realized how intricate some of the songs on this album are–they’ve got lot’s of interesting little instrumental flourishes and atmospheric vocals, including some interesting panning effects a la Dark Side of the Moon. Even with the somewhat reduced bass, there’s enough weight to do justice to the “heavy” in “heavy metal”.

The only bummer is that this requires a 7th EQ band. I can leave out the 15 kHz correction without missing it too much, but at some point I might want to see if I can get close to this with fewer bands. Or just switch to Neutron on Android (despite the lousy UI).


#11

Your inspiring me to start getting to grips with more equalising. I use Neutron on my Android Dap for that very reason but I haven’t gotten round to it yet. Just as an afterthought have you thought about trying the new hardware equaliser from JDS I think it is? Don’t know if it will be a bit basic for you but I think it looks good for anybody not wanting to delve too deeply into equalising.

-Paul-


#12

Not very seriously. I’m kind of addicted to parametric EQ at this point, as it allows me to implement very specific corrections for my various headphones. General tone controls / procedural EQ’s seem like they could be useful for mood/music-dependent tweaks, but I’m still chasing that one “all-rounder” EQ profile for my headphones :slight_smile:

On the hardware side of things, I find the MiniDSP products like the miniDSP and HA-DSP to be more interesting given their greater flexibility, but I have a hard time spending money on hardware EQ when good software EQ is available for free or cheap.


#13

Yes I too think that unless you spend vast amounts of cash your far better off with a good parametric equaliser. I can get a little daunting and complex to the casual user though.

-Paul-


#14

Having a real-time spectrum analyzer can be illuminating. On my Mac, I’ve started using TDR Nova through MenuBus. It’s a nice parametric equalizer that also includes spectrum display. One of my favorite test pieces is Yuja Wang’s performance of Rachmaninov’s 3rd piano concerto. In one of the busiest passages, this is the spectrum display:

Most of the time, nothing shows up below 50 Hz and very little above 5-6 kHz. It seems people spend a fair amount of time equalizing sub-pass and high treble, but for some things like this flavor of classical music, it just doesn’t matter much.

For comparison, here is a screenshot of Daft Punk’s Giorgio by Moroder:

The frequencies spread out much further, and the densest clustering is lower down in the frequency spectrum.

And lastly here’s some acoustic folk, Danny Schmidt’s This Too Shall Pass:

It’s similar to the classical in that it has a lot of acoustic stuff creating tone density in “the middle”, but because it has a steady drum beat we see more energy down low (like with Daft Punk). The energy above 10 kHz is a tambourine, which when you watch the live feed is clearly evident from the rhythm it’s playing.

I think it’s interesting to keep these kinds of differences in mind when comparing headphones that are known for being better with certain genres vs others.


#15

@pwjazz What an awesome thread. I am pretty much a novice, so seeing the graphs is really helpful. Do you think there would be interest in sharing EQ settings for different headphones? In particular EQ settings that would create a neutral sound signature à la Project Kameleon. The thread could become a resource depository for all. Maybe something like this already exists, but I never seem to find anything when I feel motivated enough to do a quick search. @Torq Maybe the mighty Torq knows? :face_with_monocle:


#16

If you’re looking to EQ to “neutral”, check out the Discord as described here.


#17

@pwjazz Sweet. Will do.


#18

@pwjazz This is pretty slick. Have you found a way to convert the output to parametric within EqualizerAPO so it can be used to configure other equalizers on different devices that use fewer bands?


#19

No, hence I’ve never actually used it myself. If you have a supported headphone, something like Toneboosters Morphit might give you good results. There’s a trial available.


#20

@pwjazz https://www.reddit.com/r/headphones/comments/92b08s/eq_settings_for_700_headphones/?utm_source=reddit-android

You see what this person put together? Lots of work went into it.