Harman Curve Poll (Public)

My understanding an “accurate” loudspeaker to them, which they claim on slide 8 the Revel F208 is, has a flat on-axis anechoic frequency and tilted in-room response (slide 11 figure 19 cyan curve) which follows their target curve.

At the listening position the acoustic frequency response of such a speaker does not match that of the electrical signal in the recording, thus I don’t see how it can be deemed “accurate”. And we haven’t even discussed the time/phase response of this speaker.

 

You should probably make the f**k me in the ass poll question and bar graph statement the same. They seem to be saying opposite things.

 

The Revel speakers are the bad kind of "hi-fi" that gets mocked because it's so inaccurate and makes everything sound good and similar. There are so many 6.5" woofer, 1" tweeter speakers that kick any modern Harman speaker's ass for accuracy from 300-5000 dollars. When you move into three ways where the bass doesn't inter-modulate the mids, it's game over for them.

Wait the Harman dealer heading ASR shits on ATCs, who makes the best midrange driver in the world. Amir and Ilkness probably crap on Duntech too. They honestly believe that a class d PA amp feeding phase mangled metal cones and compression drivers is doing to match stuff so clean that it makes most modern music sound awful because it has little pleasurable THD and tons of bad IMD from limiters, clipping, and bad chip pres that all gets played back.

 

I am kinda thinking that many of us (myself included) actually don't know much about details of the SBAF compensation. And feel very ironic we have barely touched about how the curve was derived on Marv's side.

Below is what I remember. Posting this to hope someone or @purr1n may correct if I am missing any.

1. The sbaf-ear compensation was originated from his flat-coupler compensation long ago.
2. And that original compensation was to mimic in-room response of his speakers that time (Fostex BLH I believe. Can't remember if subs included). Have little idea regarding exact steps taken except mild downward room curve (B&K or at least similarly approached) applied for flat line being perceptual neutral. No idea about speakers placement and room conditions, either.
3. In translating the FC compensation to EARS, several different types of headphones used in validating to address overfitting concerns and and possibly to get more generalizable results.

It's pretty interesting to see both Harman and SBAF curves have empirical speakers and real rooms in mind by nature. In contrast, diffusing-field compensation was a little more technical and depended more on a conceptual model -- shortly speaking, it took average of anechoic HATS responses in all directions (which is theoretically equivalent to diffuse field response).

 

Thanks for your correction. I might be misinterpreting or over-synthesizing some of previous clues posted in multiple scattered threads. All from my memory years ago when I was relatively new to this forum (specifically I only vaguely remember about v1 and v2 couplers, can't even name which thread was which). Happy to get a chance to update my brain. lol

 

f**k that bullcrap.

Here's what the F-statistic means: ratio of the variability between groups and the variability within groups. All we know from the ANOVA bullshit test that they carried out, if it rejected the null hypothesis, is that at least one group had a different average score from all the other groups. It does not tell us exactly how good a specific group is, or how much more "statistical power" a group has, which is another term that Sean Olive completely fucked up on. Also doesn't help that the scoring of 4 loudspeakers is a very noisy measurement of listening ability, and when your data consists of noisy measurements, your statistics are meaningless.

All this tells me is that Sean Olive is incompetent with his tools at best, and a malicious liar at worst. In this one blog post he has demonstrated that he has no credibility at all.

 

Yeah, I agree with @Friday 's points.

We should NOT support Sean's hype trains by overinterpreting his claims. It makes even less sense when the full text was neither handily accessible nor thoroughly reviewed. Presentation slides are typically a very biased (naive descriptions, selective filtering, commentaries usually BS) subset of 'what they want to convince' rather than 'what their findings actually imply' -- as we periodically see in asr. Also it's worth noting Sean's harman target papers are ALL convention papers (roughly speaking, it means 'quick preview' in academia) which are not reviewed as seriously as real journal publications are.

I think some of approaches he took could be potentially interesting and not 100% wrong. But never considered them deserve academic credits.

 

I picked strongly disagree. I always love the Harman vs Etymotic debates, Where on ASR/Reddit I've had many try to argue that the ER4SR has no bass. And how the ER2XR with it 11db peak at 20Hz that slopes down to 0.5db at 350Hz is more flat sounding???.

 

I just got the K371 (impressions), and I had the NAD HP50 (I remember Tyll talking about them which were also supposed to target the Harman curve), but from memory (huge caveat), they sound pretty different to my ears. K371 - clean, slightly hot, better control/transients, grainy; HP50 - warm, smooth, slightly boomy.

I'm not sure how close the K371 or HP50 are to the curve, but I prefer the HP50 and wish they were still available. Tonality wise the K371 has just a tad much treble/graininess for me.

But, I think it's worth trying it and getting more ears on the K371.

 

Call me fully insane but the way I hear headphones is much closer to fully uncompensated. I have to

basically every headphone I own. Always have.

Edit:
Wait a minute


That...might explain something. Would you guys say your wives prefer cans that are super dark from 2-5khz? Maybe this is like pharma studies where women respond way different and we've never verified that most headphones sound good to them.

This:

is my current EQ for the HD6xx and it's...almost exactly the difference between male and female in the above chart. This explains why I like the 8323 so much.

Edit:
Then again...



@purr1n you mentioned your wife liked the Verite. Doesn't that have a huge cliff right here?

 

This would explain the Wilson-effect: where the only way to sell $100k+ speakers is to cater to a demographic that A) can afford $100k+ speakers, and B) has significant high frequency hearing loss.

And, perhaps why after age 40-45, most male professional audio reviewers stop reporting that their significant others "asked what they did to make the stereo sound so much better lately" since they keep tipping up the treble response with each change, to match their decreased HF sensitivity.

 

I should make a list of gynecologist-approved cans that should be flatter for women, like the LCD2rev2 and the ESP95x.

When I get my EARS I'll make a comp curve for women so that we can read the measurements too. That would mean that we would need to overhaul the FR displayer page. Rather than posting images or having to post an image for each sex, I bet we could find a way to embed the viewer in posts and have the person pick their comp curve with a drop down instead.

 

I am like you but not as extreme

Both the 2013 and 2018 pinna gain of Harman curves are too much to me, 2dB too much to be specific for the 2018 pinna gain.

Yet I am still baffled by the muted air region of both curves. I thought noobies like exaggerated sound?.
My hearing is solid and linear up to 17khz so it's not that I have fucked up ears, I am still young .

 

Could be something along the lines of loads of pop releases having way too much air, or certain people (cough) involved in designing the curve don't necessarily give a bother about upper treble response cuz... it makes no real difference to em?

I spent a bit more time with the Samsung Galaxy Buds and the JBL 120TWSes (both of which are Harman products, shrug) and one thing I've noticed is that when just listening, I found them not to be as bassy as graphs suggested. This confused me until I realised that I was lowering the volume to keep upper mids and lower treble from causing real pain, haha.