Perceptual Science for Audiophiles

Audio Science
61

Yes, though a lot of the results (not all) are quite close to how we would have EQ’d those headphones already, particularly for the ones where those adjustments were made by tone gen. I think it’s more that now we can be more confident in advocating for tone gen based methods than we were before.

Effectively getting the lab data for our HRTFs confirmed that the DF condition tracks with perceptual balance using tone gen, which is something we hadn’t known for sure before. Moreover, it also makes the entire process of “doing it right” way easier for us. Because now we have a visualization for it as well.

1 Like
62

All true. However we must also remember that there is no way for any of us to truly share what we are experiencing with another. I think this is core to the problem. This is getting more into philosophy than perceptual science, but really… if you can’t prove to me that you actually exist then how can you prove to me that you experience music at the same level I am experiencing it? For example you say you used EQ to tune an HD600 to sound EXACTLY like an HD800S, but how do I know you are experiencing the music through the HD800S at the same level I do in the first place? Maybe you only really pay attention to overall tonality while I’m experiencing the feeling of actually being in a concert hall with the symphony playing in front of me.

And then there’s the confirmation bias of expecting the HD600 to sound like an HD800S and having no way to blind test or even rapidly A/B switch them.

Note I am not doubting your claim. I agree it is all just FR, mostly because there is only FR, distortion, and noise. And distortion has continually been shown to not be an issue with headphones. And since it is all just FR then it follows that the right EQ should be able to make any sound profile desired within the physical limits. I’m just pointing out (rambling really) where some of the difficulty is for you to prove your claim to others.

2 Likes
63

Yeah unit variation is an issue but one that is difficult to account for as we don’t have the data. And cheaper headphones are bound to have more unit variation.

I may just try to prove to myself that it is possible. I own a few flagship headphones. I have EQ’d them to sound their best. I do believe that any headphone can be improved with EQ. But I’ve never taken a mid-fi headphone and tried to EQ it to summit-fi using tone gen. Maybe a Sony MDR-MDV1 purchase is in my future.

2 Likes
64

Agreed, but it could be demonstrated to the person if they are willing to give it a good faith try. If we’re saying “Hey I used to care about all these technicalities things, but now I’ve learned that it’s all FR, here’s you can confirm that for yourself”, and the person turns around and says “I don’t want to listen to test tones”… at a certain point you can’t really have a further conversation about that.

This we did actually do - at least I also did this to some degree with those specific profiles, since at one point Grif was doing the switching without telling or showing me which was which, and I had to guess. But that’s also a bit less interesting since differentiation was never the point, that’s an easy one.

But yes, full agreement that there are difficulties in proving it to people without putting mics in their ears and doing this in person for those people.

1 Like
65

There’s also the issue of the tools used to apply the EQ. I keep noticing that when I apply EQ, either via Sonarworks or PGGB with Foobar2000, it seems to alter the playback in an adverse condition. For example, the attack/decay of a stringed instrument seems to become altered. This seems to happen a majority of the time with most headphones, so tend to stay away from it.

66

I officially volunteer to give it a good faith try. When do I fly out to the Dyson lab? :stuck_out_tongue_winking_eye:

But seriously you can understand why there is little motivation for an audiophile to purchase a headphone for the explicit purpose of making their existing headphone that they love redundant.

But what you need to be able to do is rapidly and blindly switch between the HD600 tuned to the HD800S and the actual HD800S. Only then will you know if they sound exactly the same. If only you could ABX test it.

1 Like
67

Oh, that we also did - though not actually blind because… obviously the headphone is different. But yes, we were swapping quickly between the HD 600 EQ’d to HD800S and the HD800S, playing the same track at the same spot in the track back and forth. We filmed some of that, so hopefully that part makes the cut for the video next week.

Yeah and I also wouldn’t advocate for that, because I think the part people are missing is that… at least two of us preferred the cheaper headphone! Both Grif and myself had a clear preference for the default HD 600 over the HD 800 S, despite the latter’s advantages in terms of openness and comfort.

So instead I’d advocate making purchase decisions around build and comfort, and then ease of EQ. Then do the tone-gen based EQ to make it perfect.

68

We are coming at this from fundamentally different conceptual realities. I’ve used some form of EQ when needed for more than 50 years. It was mostly crap EQ back then, graphic equalizers. 5 band, 10 band 23 band were common. Recordings were variable and the KOSS Pro 4AA was a good headphone. Many of us had stuff like VERITAS that we bought at Rat Shack for $29 ($129 today).

When I was in college, the scientific audio was learning about speaker design. Butterworth boxes and equivalent electrical circuits. Room treatments. etc. And still recordings were pretty variable.

I listened to a LOT of live music. Lots of classical guitar. Mom taught piano and classical guitar, had a masters in music and attended master classes in guitar at the Kansas City Conservatory. I went to a lot of traditional concerts. When I was older I saw a lot of outdoor performances from Bluegrass to folk and rock. Saw Zappa, Springsteen, Hancock, Bill Evans, Rush, Kiss, Blue Oyster Cult, Jethro Tull, Brubeck, Ramatam, Mayall, NRPS, Exuma, Jarrett, Preservation Hall Jazz Band, Poco, Journey, Def Leppard, Santana, Toto. Lots more many of them several times. So I know what stuff sounds like live, even if they played too loudly or my seats weren’t perfect.

My first decent headphones were Audio Technicas from the 1970s. In the 80s I found STAX, and still have my SR5n STAX. In the 90s I bought the Sennheiser HD-580 when it was new, and a friend had the high end Grados in 2000 as well as STAX “earbuds” that were sort of headphones.

Before headphones.com, there was Head-Fi. and teeny little pocket battery powered amplifiers. When headphones.com came along, I got the invites (all of us headfi people did) and after maybe 6 months joined and caught the upgrade bug.

My old instincts wanted simple. The Schiit Lokius (and to some extent the Loki before that) gave me access to a few knobs I could twirl and improve most of what was wrong when some tin-eared engineer, or producer that wanted things to sound good on car radios had done the mix. Yes PEQ was clearly better in many respects, but it’s not easy to adjust on the fly for one song.

I use PEQ in my car (via ROON ARC) to correct some of the usual car stuff. It’s frustrating that the car audio doesn’t include PEQ equalization as most aftermarket autosound head units do. That’s an appropriate use of EQ - especially since I upgraded the factory bass driver. And no, I don’t have much choice, it’s a Cadillac LYRIQ and the INFOTAINMENT system is a pain in the ass to interrupt to apply something like DSP.

Having something like the LOKIUS confirms to me the limtations on bass of headphones like the HD6xx, the HD580 and the STAX SR5n. The HD-580, perfectly competent in the midrange can’t be pushed with the Lokius - and so far its got the same limitations with PEQ when it comes to bass and extended treble. Certainly I can make it sound better than stock to my ears with PEQ. But no matter what I do, I don’t think there is magic that can make it as resolving as the ROSSON RAD-0 or the Hive Nectar eSTAT. or my ZMF Auteur Classics. And why should it? Even given the esthetics of fancy wood or composite materials, the ZMF, HIVE, and ROSSON are 25 years newer, have more advanced materials, higher end materials. The engineering of the HD580 was pretty good for it’s time.

Just as I hear more SOMETHING - depth, clarity, definition with the Schiit BiFrost 2/64 than I hear when I use an Apple Dongle DAC/AMP, I also hear more of that SOMETHING - depth clarity definition when I listen with any of the 3 new headphones compared to the old Senn HD580.

So is it possible to make the HD580 sound as good to my ears as the ZMF? I don’t think so. I’ll bet Zach doesn’t think so either.

3 Likes
69

I’d love to be a subject that gets to listen to the test tones. I just want someone to do it for me and bring it to me. Bring the stuff to my office, and I’ll make you a killer cup of coffee. I can also show you the value prop of using a professional financial advisor.

And I get some perfect EQ that’ll make me as happy with the Senns as I am with the ZMF and the EgglestonWorks Nico Evo speakers in my office, it’s a win-win.

70

If we ever get a chance to be in the same place I’ll bring some in-ear mics and maybe I can demonstrate it for ya.

71

Then, no offense, you are doing it wrong. No really, bear with me. If the EQ makes a change for the worse then it is the wrong EQ for you. But don’t throw the baby out with the bathwater.

First, you should eventually EQ by ear rather than by measurement. This is very important. Because we all hear a different FR than the measurement rig you can never get a truly accurate EQ without doing it for your own anatomy. You can start with a measurement based EQ or use the measured FR as a guide, but using tools like Owliophile.com should be the ultimate goal. IMHO no matter how you EQ by ear, it is more of an art than a science (unless you have your own HRTF measured and proper in ear microphones like those lucky Headphones.com guys). There is no simple to follow, one size fits all recipe.

Second, you should EQ to your taste rather than a target. You should already like the sound of the headphone and just want to improve it or fix a few issues. If one of the filters in your EQ removes some aspect of the headphone that you enjoy, turn it off. It’s not an all or nothing thing. It doesn’t matter if it makes the FR “less neutral” or “target compliant” or “correct”. Keep iterating enough and maybe you end up without any filters at all turned on. Or maybe you end up with just adding some bass. But if you are truly EQing by ear you will very likely still have a few changes to the treble.

I had an experience recently with the Hifiman Mini Shangri-La where I had an EQ filter at 3.5khz -4db or something. I think I got it from trying oratory1990’s preset. Regardless you’d think “OMG these are way too loud at the ear gain peak! That must surely be an important filter to have on so that they sound ‘correct’!” When in actuality, to my ears at least, turning that filter off just gave the drum snap aka transient attack back that I’d been missing. The rest of the EQ preset was fine without it. Instead of concluding that EQ or a specific EQ preset ruined the headphones, I simply turned off one filter and was content - for a while anyway. And I learned that I could boost that same region a bit in other headphone’s EQ to add that same technicality to them.

1 Like
72

Oh, I don’t plan to give up just yet. The EQ settings I found for the Z1R are excellent. I’ve had some luck with EQ on the HD 820 as well. I’ve had no luck with the rest of the headphones.

I’ll start with the ADX-5000 and see if can make it better. It’s already a headphone I really like. Just needs a little more midrange warmth.

1 Like
73

Sounds perfect

I just wanted to add that EQing - especially when doing it by ear - is an iterative process of usually small improvements. And similar to @Klute’s experience with multiple headphones, comparing multiple EQ presets is tiring. Do a little bit and then enjoy the results for the rest of the day, is my advice.

1 Like
74

That’s great - and I’ll bring a silk purse that you can turn into a sow’s ear. I figure that the procedure is reversible.

75

100%. That’s the real magic trick.

76

So now that we’ve got that out of the way, why not talk about the need to EQ the music, not the headphones. If you read my longish post above, I’ve got pretty decent headphones, and can listen to them fine without EQ.

It’s the damn stuff that someone did to the friggin’ music that’s the problem. They put the mic in the wrong place. Or they thought they should mix several mics they way they shouldn’t. Or they don’t know an oboe from a trombone. Or maybe their girlfriend sounds shrill and that’s they way they think all female singers are supposed to sound.

The opportunities for EQ are endless.

77

Yeah unless you’re recording your own music, it’s all being EQ’d anyway, and sometimes that process isn’t done so well. It’s rather annoying.

78

Oh man, don’t get me started! When it was still called iTunes, you had the ability to assign custom graphic EQ to individual songs. You still can in Apple Music, but it doesn’t do anything!

I consider this the biggest feature gap in Roon and it would be pretty easy to do. If not per song PEQ, then at least per album graphic EQ or somewhere in between please. This makes me want a Loki of my own.

1 Like
79

The Lokius has 6 bands. It’s great. Great enough that I have thought about the Lokius Maximus F.
The idea that it has presets you can customize is really tempting.
Schiit Audio: Lokius

Schiit Audio: Lokius Maximus F

2 Likes
80

Just to take another opportunity to push my favorite music player app :grin:, Neutron player on Android can assign EQ presets per-folder, it’s had this for years. Myself, I never used this because in the few cases where I found I absolutely had to correct a song or album here and there, I felt changing the files themselves in Audacity and re-encoding them was acceptable (no bit-perfect idolatry here), and a more permanent and reliable solution.