Harshness with Bartok using headphones

I’m looking for a little help troubleshooting my system. I’m using the Bartok with Meze Empyreans and for some reason the sound has been painfully harsh. I’ve read reviews of this same combination, and nobody mentioned ear pain, so there has to be something more, probably to do with cables or peripherals that is responsible.

This is my setup:
Shunyata Denali 6000 power distributor fed by Sigma XC 20 amp cable
Shunyata Sigma NR v2 power cable from Denali to Bartok
Innuos Zenith Mk3 server connected directly to Bartok via network cable (I started with the stock ethernet, but now use Shunyata Sigma which I like better)
Headphone cables:
Double Helix Prion 4, balanced and single ended
Meze copper upgrade unbalanced cable
Meze stock unbalanced cable
Moon Audio Black Dragon balanced
IsoAcoustic Indigo vibration feet
Butcher Block Acoustics platform, with another set of Indigo vibration feet to the floor

What I’ve noticed so far: Single ended output is on the edge, but balanced is over the edge into painful treble. The Prion 4 and stock Meze cable (the warmest and fuzziest) are easiest to listen to, while the Black Dragon is absolutely horrid. USB tends to be easier listening that ethernet, but I like the Shunyata Sigma ethernet. I use the Mosaic player, and I’m not running Roon.

I’ve tried the various filter and upsampling modes, DSD and filters 1 and 2 being easiest, but still with very little wiggle room.

The Shunyata v2 power cables are a tad bright IMO, but I’ve also tried an older and mellower Alpha NR straight to the wall, and it wasn’t that much better. Basically most tweaks have brought me either just short of too harsh, or way over the line into harshness.

I’m kind of at my wits end. I just bought the Bartok, and I want to keep it, but it just starts hurting my ears at relatively low volume (low enough that the ambience and bass are not fully fleshed). With rock and electronic music I can last a couple minutes.

The only thing I can think of is going in for another spendy headphone cable, the Transparent Ultra headphone cable that is often paired with the Bartok, and sounded reasonably good in the showroom.

Is it the headphone pairing? Is there something I have overlooked?

Just to complete my picture: what is your headphone level output setting of the Bartók? 0 | -10 | -20 | -30 dB?

What is the volume level you use to play your headphones at? How many db?

A couple of things that might be worth investigating:

  1. Galvanic isolation of the Bartok. To do this you could insert an optical cable run in the ethernet link to the Bartok. You would need 2 ethernet-to-optical fibre boxes (e.g. TPLink) used back to back and connected by optical fibre.
  2. Mains regeneration rather than conditioning. The latter may not be sufficient if your supply is badly distorted harmonically. PS Audio regenerators have excellent reviews for solving this issue.

Right now I’m using -10db and typically have volume setting between -25 and -15. Sometimes I use 0db, and turn everything down, but it seems like -10 is a good match. It is a bit unusual to have the volume cranked up so high for me (I have a couple other amps that are usually set between 9:00-12:00), but this seems to line up with what the manual says.

I think in reality this is probably between 80 and 90db, but if my phone app is accurate, I get some problems even down into the 70s. I also know that at 80db you are supposed to be able to listen for 8 hours without hearing loss, so at say 85 I should be able to go more than 15 minutes without my tinnitus flaring up.

That is an interesting idea. I found the Shunyata ethernet to be a subtle, but nice improvement. I might try an Etherregen for this.

I used to have a PS Audio P15, and it sounded less natural and more bright than the Denali, which is fairly transparent. The only real advantage IMO is during the day when the mains sound a bit duller.

The Meze Empyrian has a impedance of 31,6 Ω. The Bartók’s minimum output impedance is 33 Ω with full-scale output level at 1.4W rms, while the Empyrean needs just 1mW to output 100 dB.

It looks like the two don’t match too well, specially if you output it from the Bartók at such high volume levels. But dCS assured this would not be an issue, and in fact the output impedance of the balanced output is a mere 0.1Ω.

However, I am puzzled why you have to output the Bartók at between -25 and -15 dB volume out. The Empyrean should sound harsh then, distorted.

Maybe something is wrong with your input level? Maybe it is way too low, and thus you have to crank up the volume?

I thought the same thing about the minimum impedance, but as you pointed out, DCS does say this headphone is compatible, and a number of positive reviews used it.

I can set the output to 0db, and then I’m running generally -35 to -20db, although this actually sounds a little bit harsher. The manual says volume should be between -30 and -10db. I would expect to not have to have the volume so close to maximum, but I guess this is just how it is with planars. It seems strange to have this range of volume and be expected to only use the top of the range.

I assume that you have tried the 4 voltage output settings on the Bartok ? I am unsure if they affect the headphone as well as the line outputs ( the spec sheet for the Bartok with Headphone amp is unclear on this).

The Bartok has a full output of 1.4W into 33 ohms. The Meze Empyrean has a very close impedance to this, 31.6 ohms, call it 32 ohms. That 1 ohm differnce is insignificant and probably falls within manufacturing tolerances. The 'phones are specified by the manufacturer as having high senstitivity ( efficiency) and will produce an SPL of 100dB @1 mW input. That is over 1000 times less than the Bartok is feeding it given a 0.0dB volume setting with typical programme material; you don’t even need a recording modulated to 0.0dBfs to achieve this in these circumstances.

So , everything else being equal, if the Bartok volume control is set at 0.0dBfs the phones should be receiving more power than required for any realistic listening level and, on paper, they are not a good match for the Bartok, as Ermos has said. The phones could be being driven into distortion.

Given the numbers above, your estimate of SPLs e.g. 80dB for a -10dB setting may be way out. I say it is an estimate as unless you possess artifical ear type measuring equipment I don’t know how you could arrive at a figure. I am concerned about your hearing if both headphones and Bartok are performing as intended especially as you indicate that after only 15 minutes of listening you suffer from tinnitus.

These are only the line level outputs of the Bartók, and they do not affect the headphone output.

Thanks Erno.

1 Like

This is something I will consider. I would be extremely surprised if my listening volume was anywhere near what the math dictates as I usually can’t stand being in a club without earplugs and I haven’t noticed any actual hearing loss day to day. I use the microphone on my phone in the same proximity to the Empyrean driver and a couple different apps to give me an average db reading, which range from 75db to 87db.

When I have the volume set to the minimum I can really do, -10 output and -30 on the control, I can hear outside noises pretty clearly and some cars 100ft from my apartment can actually drown out the headphones. There is still some shrillness in the treble at this volume.

One other thing that surprises me is that my other headphone amps don’t produce this agitation, and require much lower levels. My LTA MZ3 amp for example should have a fraction of the output power for 32 ohms as the Bartok does in balanced mode, and produces comparable volumes at between 30-60 on the control as the Bartok does at 70-90%, with more fleshed out bass, and without the ear pain.

I will investigate further. Unfortunately, my only other headphone is an Audeze LCD-XC which is way out of spec and sounds incredibly bright.

One potential clue is that balanced headphone operating is more fatiguing per the exact same listening volume.

Just checking: are you using any headphone cable adapter into Bartók? And is 1 headphone plugged into Bartók at a time?

Yeah, one headphone at a time. If I use the Prion 4 in single ended then I have a Double Helix balanced to single ended adaptor. I haven’t used an adaptor with my other cables.

One thing I learned is that the volume is meant to be set really high because the headphone amp uses a digital attenuator. Because some resolution is lost the lower you go, DCS intends listeners to stay toward the top of the scale.

Indeed, but the harshness you experience should not occur. It never happened to me.

However, I have stopped using the headphone part of my Bartók, and use a line amp/ headphone amp, with the Bartók at unity gain 0.0 dB, 6V out. Max resolution, better volume control, better definition and bass, also at low level listening.

Hi Ian,

Firstly, I am sorry to hear you haven’t been enjoying the sound from your Bartók so far. The sound being harsh with the Meze Empyrean headphones is surprising; feedback we have gotten with the Mezes is often that the sound is more warm and full bodied with the Bartók, as opposed to on the harsher (or “clinical”) end of the spectrum. What sort of music or content are you listening to with this setup?

The Empyreans are a fairly efficient pair of headphones, so we usually see them run on the -10 or -20 setting. Running them at 0 can potentially overwhelm the headphones given the efficiency, so it sounds like you have the right settings there.

You mention that you only recently purchased the Bartók, was this a brand new unit? If so, did you have the opportunity to listen to one before purchase?
Same with the Mezes, are these a new pair?

As a side-note, the Audeze LDC-XC are what we use as our “reference” closed-back headphones, for use at factory demonstrations, shows and events. I wouldn’t expect t find any issues with using them with the Bartók under normal circumstances; we think they are a remarkable pairing! Are you finding a similar sort of harshness to the Mezes with them?

Hi James,

I want to be clear I’m not trying to put down your product. Ultimately, I’m troubleshooting some things that may be peripheral to the DAC itself such as power cables and digital inputs and system settling.

I did listen to the Bartok before I purchased it, however not on my own system, so some guessing was made. I think I assumed I could get a smoother sound out of it, but also undervalued the Transparent power and headphone cables in the demo.

Some of the DACs and headphone amps I’ve used in the past have had a somewhat rounded or rolled off sound which allowed listening at higher RMS values, and I think I’m also adjusting to proper levels with a dynamic treble.

I’ve gone back to the LCD-XCs, and they don’t seem to be causing a problem at this point. It might have been burn-in related. They were my reference for a long time, although I do find them a bit tilted towards the treble these days.

Hi Ian,

Absolutely understood, and I hope you can appreciate that my surprise with your experience thus far only comes from that the sort of sonic qualities you are hearing are such a swing from what I might expect from these product combinations. Of course when dealing with something as subjective as hearing, YMMV. Was your demo with the Mezes, Audezes or something else?

The burn in of the unit may have some impact on these qualities, as of course will the peripherals you have been exploring. I would be keen to hear how you are getting on with the unit after some time with it, both for it to settle in and your ears to acclimatise.

Thanks. I think I have largely resolved the issue with a simple fix that’s almost too obscure to mention…but it worked. The tonality of my Bartok is smoother and richer now.

I will say though, I think I also just prefer single ended headphone listening.

FYI, I did demo the Bartok with the same set of Mezes.

Hey Ian!

Would love to know what you did to fix it. Im also using Empyrean with my Bartok and suffer from harshness.