The Stax Thread III

[dagothur]

It’s either that they felt it wasn’t a good complement in both price point and sonic signature or they have a new version in the works.
Retailers have said the bulk of Stax sales were 007 and x9k , so it was not demand related, but no official reason was announced.

I really hope they make a new version with the X9K's headband and detachable cables. The sound signature is perfect to me.

 

[BenF]

Her are the graphs:

Larger:

Is this a real frequency response of X9000?
Looks terrible, ~20dB drop from 1kHz to 5kHz
009 measures a lot better.

 

[LarsHP]

Is this a real frequency response of X9000?
Looks terrible, ~20dB drop from 1kHz to 5kHz
009 measures a lot better.

Unfortunately, we can't compare different measurement setups. There is only relative value, comparing different headphones or amps using the same setup.

 

[Kiats]

Been enjoying this new pair of 007 Mk1 I picked up from Yahoo JP auctions. Still one of my favourite Stax HPs.

 

[SolarCetacean]

Is this a real frequency response of X9000?
Looks terrible, ~20dB drop from 1kHz to 5kHz
009 measures a lot better.

The measurements in that thread are using a non-standard measurement rig. If you want a GRAS measurement, Oratory1990 has one here.

 

[Svperstar]

Been enjoying this new pair of 007 Mk1 I picked up from Yahoo JP auctions. Still one of my favourite Stax HPs.

How much did you pay if you don't mind me asking?

 

[BenF]

Unfortunately, we can't compare different measurement setups. There is only relative value, comparing different headphones or amps using the same setup.

We should be able to compare compensated measurements (not exactly 100%, but should be close).

The measurements in that thread are using a non-standard measurement rig. If you want a GRAS measurement, Oratory1990 has one here.

Now this one looks much more similar to his measurement of 009
His measurements seem too smooth though, he probably used 1/3dB or 1/6dB smoothing - I use 1/12.
He may need to re-calibrate his measuring head, 009 doesn't have 12-13dB drop in the bass compared to 1kHz..

 

[LarsHP]

We should be able to compare compensated measurements (not exactly 100%, but should be close).


Now this one looks much more similar to his measurement of 009
His measurements seem too smooth though, he probably used 1/3dB or 1/6dB smoothing - I use 1/12.
He may need to re-calibrate his measuring head, 009 doesn't have 12-13dB drop in the bass compared to 1kHz..

The Oratory1990 measurements makes my point quite clear.
Different setups will measure differently, particularly when it's non-standard headphone measurements.

 

[walterboss]

I entered the Stax world a few weeks ago with an L700mk2.

Absolutely love it.

Have an X9000 arriving tomorrow.

Anyone else made a similar jump and have some impressions?

 

[SolarCetacean]

I entered the Stax world a few weeks ago with an L700mk2.

Absolutely love it.

Have an X9000 arriving tomorrow.

Anyone else made a similar jump and have some impressions?

My first Stax was also the L700mk2 and I also upgraded to the X9000. I found the X9000 to be better overall. Bigger soundstage and more spaciousness, less honky midrange tonality, smoother treble, and more neutral sound signature on the X9000. The L700 was noticeably warmer with a larger midbass hump and could have some "glassiness" or "crystalline" timbre in the treble in addition to a more colored midrange tonality. However, the L700 has its unique soundstaging where it has this wedge-shaped soundstage, like you're in the middle of a pie slice and looking towards the point. It has a very defined center image and it pulls elements forward. It's focused and stage-like, though at the cost of creating a gap in the imaging at around 45 degrees from center forward which can mess with some binaural elements.

I haven't listened to my L700 in a few months because I loaned it to a friend. He heard both the X9000 and the L700 and he preferred the L700. He liked how it imaged vocals closer to him and he enjoyed the bass. I dislike overly forward vocals, so I prefer how the X9000 moves them further away and I prefer the flatter bass on the X9000. However, I think that the colorations on the L700 can work very well with some types of music. The midrange honk works well with brass instruments, like it highlights their brass nature, and on some vocals it can produce this very lifelike rendition, as if I was listening to the raw mic feed rather than the processed and produced studio track. The X9000 can still do those things, though maybe not to the same heights, while also having fewer side effects in its tonality and imaging.

The X9000 is also more comfortable to me (the rectangular earpads on the L700 always pressed on my cheekbones the wrong way, which isn't helped by how prominent my cheekbones are) and I think it looks a lot better too.

 

[DougD]

The Oratory1990 measurements makes my point quite clear.
Different setups will measure differently, particularly when it's non-standard headphone measurements.

As a career data guy, I have near-zero trust in FR measurements of headphones. When you look at one, there's just no way to know if what you're seeing is artifacts of the measuring setup plus post-processing, or is the true nature of the h/p. Especially above 8k or below 200. Reproducibility of results across different testers is abyssmal. I understand the DESIRE to have FR measurements on h/ps, but the reality is that today's affordable technology just isn't up to the difficulty of the task.

I would NEVER use FR charts to help me make decisions on headphones.

MUCH safer IMO to lean on the subjective impressions of reviewers whose opinions you've learned to calibrate against your own preferences. Plus of course listening sessions if you can.

My opinion. YMMV.

 

[LarsHP]

As a career data guy, I have near-zero trust in FR measurements of headphones. When you look at one, there's just no way to know if what you're seeing is artifacts of the measuring setup plus post-processing, or is the true nature of the h/p. Especially above 8k or below 200. Reproducibility of results across different testers is abyssmal. I understand the DESIRE to have FR measurements on h/ps, but the reality is that today's affordable technology just isn't up to the difficulty of the task.

I would NEVER use FR charts to help me make decisions on headphones.

MUCH safer IMO to lean on the subjective impressions of reviewers whose opinions you've learned to calibrate against your own preferences. Plus of course listening sessions if you can.

My opinion. YMMV.

I'd say I agree with your point about FR curves in the treble (above 4 or 6kHz) and bass are generally unreliable.

I remember Tyll Hertsens did several measurements, repositioning the headphones between each in order to find an average in the treble, which typically changes drastically depending on the position, where the spikes and sharp dips "move" because the resonances and cancelings changes.

Regarding bass, the SR-007 Mk2 I just bought shows a rolled off sub bass below 60Hz while the HE-6 measures dead flat down to about 30Hz. However, when I put the headphones on and compare them subjectively, it's the HE-6 which has less sub bass. This could be because the HE-6 has a tighter bass (which it has IMO).

 

[Svperstar]

Especially above 8k or below 200.

Interesting you mention below 200hz. One of the most economical ways to get into estats is get a used Koss/Drop E95X. An estat and an amp that sold originally for $499. That setup is notorious for lack of bass slam. If you look at measurements there is a very steep drop as soon as you hit 200hz and below. You can EQ it up a little bit but eventually the driver will distort. When the E950 was originally released I guess they assumed most people into it would be listening to Classical or Classic Rock, not modern day bass heavy music so the bass is dropped so the driver can focus on the rest of the music. At least that is one theory. I think Kevin Gilmore posted it is due to the design of the amp it simply can't produce low bass.

Anyway, if you get an adapter cable to use the headset on a different amp you will be shocked how good the bass is on the Koss 95X. If you just went by measurements you wouldn't think it was capable of that kind of slam.

 

[Svperstar]

Oh yeah today is exactly 20 years since I signed up for Head-Fi. Took me almost exactly 20 years to really get into estats.

 

[DougD]

I remember Tyll Hertsens did several measurements, repositioning the headphones between each in order to find an average in the treble, which typically changes drastically depending on the position, where the spikes and sharp dips "move" because the resonances and cancelings changes.

Yes Tyll did average multiple measurements, and that was one of the reasons I liked his approach better than most.

I assume ... Tyll probably explained, but I don't recall precisely ... that many of the spikes & dips & resonances & cancelings are largely produced by the measuring gear & set-up. Hence my use of the term "artifacts." What that tells us is that for the measuring gear, the placement of the headphones on the dummy's head can have a significant effect.

What is unclear ... to me, and I confess I've never bothered to look for research on this ... is whether the placement of headphones on a human's ear is equally significant. In general, I don't think it is. We frequently read things like "we all hear differently, everybody's inner ear is shaped differently" etc. But it's rare that I perceive much difference if I shift the position of a specific model of headphones around my ear.

Better evidence is provided if we listen w/o headphones to a sound signal. Like a mono speaker. If I am facing directly at the speaker, I hear a certain tonal range etc. If I angle my face 5 degrees left, I don't perceive any difference. 10, 20, 40, 70 degrees, left, right, up, down ... very little difference in what I am hearing. When I rotate my head in various directions while listening, I do not hear any bumps/dips/irregularities/discontinuities at certain orientations. And yet there can be no question that the sound waves are arriving from different directions when I do that, and the patterns of resonances & cancelings inside the outer and inner ear therefore MUST be different.

If that must be happening, why can't I perceive it?

One (partial) answer could be that I am an untrained and therefore poor listener with lousy discrimination. Guilty as charged. Probably true for many of us. (Some day I will bore you with my theory about why being an untrained listener is a good strategy for the thrifty audiophile.)

My more general theory is it's because we have a two-part hearing systems. Our ears are the external sensors that monitor our local environment for soundwaves. Each sensor has its technical limits, which generally degrade with age. But there is a semi-magical processing unit also involved, called the brain. The autonomous part of each brain learns to synthesize its sensory inputs and produce what I'll call "the angle-adjusted (normalized) sound spectrum", which it then passes onto the evaluative part of the brain. This autonomous processing will drop stuff, fill in holes, do smoothing, who knows .. it's magic.

And so per my theory, part of the reason why when we listen to headphones they don't sound spikey & dippy like the charts is because our brains are applying advanced filtering and smoothing. Including the brain's expectation (memory) of what the sound SHOULD sound like. Trained listeners can overcome some of that, but most of us can't/don't detect relatively small effects.

A corollary of this theory helps explain why "everybody's inner ear is different" is technically true but largely irrelevant for most people, i.e., those with "normal hearing." My brain tells me "normal sound spectrum, or not normal sound spectrum", given my body's microphones and my brain's lifetime of synthesizing/normalizing those inputs. Your brain is telling you "normal sound spectrum, or not normal sound spectrum", given your body's microphones and your brain's lifetime of synthesizing/normalizing those inputs. The two brains are generally going to arrive at the same interpretations.

In conclusion ... I'm going to pretend there was some point to this speculative ramble ... another reason why h/p frequency charts aren't generally helpful is because humans don't process sounds like machines do, and vice versa.