Hey, don’t tell me, tell them!
1 Like
mumble mumble mumble…preaching to the choir…falling on deaf ears…pearls before swine…spitting in the wind…spitting in the ocean…mumble mumble mumble
Gearheads gonna gearhead. It’s what they do.
1 Like
Blockquote
Simply put, the HpTF variation is going to be far too strong for any of these programatic methods of doing EQ to be effective beyond making truly horrible headphones sound less horrible. It’s a misunderstanding of both how headphones behave and what the measured performance actually indicates.
Not to mention the lack of measurements of the outcome, which always struck me as odd. The curves accompanying rococo EQ profiles showing achievement of the target are just calculations rather than measurements themselves, at least as far as I could tell.
1 Like
TRUE! You genuinely don’t know what you’ve EQ’d your unit to after the fact. I empathize with folks trying to get into this stuff without measurement gear to verify it. We can cheat!
1 Like
I remember an oratory1990 post from a few years ago where he said he stopped verifying the effect of his EQ through measurements quite early on because it always came out exactly where predicted. I’ve always had my doubts about that. Is that not your experience?
1 Like
I could see that holding true for something like a bass shelf – it should come out as the calculations predict. For narrow filters used to try to smooth out treble peaks, “graininess,” what have you, I find that implausible.
For the EQ values sure, I’m the same way. Like at a certain point you know what your adjustments are going to do, even if you just estimate the change. But we’re the ones doing the initial measurement of the headphone. A random person can’t know how their unit necessarily measures. But more importantly, they don’t know what the in situ FR is going to be.
1 Like
Ah! So you’re saying because of unit variation a consumer cannot know what the measured EQ’d FR will be without measuring the before and after themselves? I understand that we are not talking about in situ FR here.
So essentially if you have a measurement rig, take a measurement of a headphone, apply EQ, and remeasure that same headphone on the same rig then the measured FR should match the predicted EQ’d FR exactly, just as oratory1990 stated.
Yes, at least when measured on that same rig. But yeah there are all kinds of sources of potential mismatch when someone else is using those profiles like unit variation and pad wear.
1 Like
Wanted to post an update about EQ. The latest software version of Sonarworks has been released. Since loading the most recent version, I’ve had better results with using the Sonarworks EQ curves with Foobar2000. I’ve tried it with the Sennheiser HD 600, 650, and the Focal Lensys. All the headphones sound improved compared to no EQ applied.
After contacting the developer of PGGB, he provided guidance regarding optimum EQ convolution settings for use with the Sennheiser HD 820 with PGGB-RT. This definitely makes the HD 820 sound much improved.
1 Like
Very interesting. I’m a little surprised that there is that level of predictability with higher frequencies. But this may be because I don’t know how headphones are tuned. When I see FR curves that show dips and daggers, I usually guess that those were accidental, so to speak – that there was some kind of cancellation owing to reflections – and that that might mean that if you try to add some juice to that frequency, you won’t get a 1:1 result.
One of the more interesting and illuminating things on this subject I’ve seen on the Headphone Show vids was when you went through the curve generated by the rig for the ATH7000 and said out loud whether the peaks on your head were where the rig indicated they were, and they were often close, but not in exactly the same places.
Yes, you can expect the headphone behavior variation to be substantial for those features - it depends a bit on the type of headphone but generally it’s to be expected that you won’t hear the dips and daggers the same way, at least above certain frequencies. When we see those on the graph we have to think “that head and ears hears this headphone with those dips and daggers in those spots, they may be similar for me, but not guaranteed”.
I think the question to do with predictability though was with reference to how a headphone would measure on a rig after EQ, and that’s sufficiently predictable that you wouldn’t even need to measure it after the fact. But, also no reason not to.
Yeah I sometimes find this to be the case, particularly with open back undamped designs. Like maybe a feature exists but not in the same spot. But there are also times when a measured peak just isn’t there on my head for example.
1 Like
I would like a new video on EQing as well, EQ FR and all of the science is way over my head for some reason, and I’m a smart person just not on this subject.
I was thinking that when I get my Tungsten I would challenge myself to make them sound like a KPH30 or if the form factor makes that impossible, like a Fidelio X3. Something that is very unique compared to anything else I’ve heard before. I don’t want to come off as a skeptic, just someone that doesn’t understand. Limitations I come across, the process of doing it etc. Would a planar magnetic have the ability to sound like a dynamic driver? The tungsten, though I don’t have it would be far more capable than the two I mentioned, but different technology. Does that just mean that you’d need an infinite number of EQ adjustments to make it work right? Maybe make them sound like another planar, para II, XS and t50rp are all I have left of planars.
Also what equipment or software is needed would be useful, I have a miniDSP that could be connected to a headphone amp if that gives better adjustments. I also have a Centaraus and a PC
We definitely need to make an updated video on EQ with some new process guides. Re your question, yes you could theoretically make a planar sound like a dynamic or vice versa with enough time or patience. So… for example, every time I mention to Oratory “I struggled to get this one to sound like this other headphone”, he’d say to me “yeah… skill issue”. In that instance, it had to do with coupling and acoustic impedance. But he’s right, for that kind of ultra precise goal of getting two headphones to genuinely sound the same, you have to use in-ear mics of some kind and do a lot of it subjectively. And yeah you need a lot of filters. But the point is really that you shouldn’t be doing this anyway. You should instead be trying to get them to sound better than anything currently on the market, not like anything else.
3 Likes
I’ve been listening to the FR of a number of other headphones on my DT770’s recently. It is certainly not a perfect process, due to some of the aforementioned reasons. But still interesting and educational.
I think Super* Review explains how to do this in one of his squiglink EQ videos. There are couple different ways to accomplish it though. And I still prefer to use GEQs for this, rather than PEQs.
One method I’ve used is to add the FR of my headphone (DT770) and a target headphone into squiglink. Add some smoothing, And use my headphone’s FR to compensate the target curve. Then use UsyTrace to convert the resulting difference curve into a GEQ that can be loaded into EAPO.
If you already have smoothed CSVs of the two headphones’ FR curves, then you can save a couple steps by simply loading them both directly in EAPO’s Configuration Editor, and inverting your headphone’s GEQ to compensate the target headphone’s GEQ curve. If done correctly, this should give the same result.
As mentioned earlier, I also use averages of several good-sounding headphones for my target FR curves with this type of approach. And it seems to work pretty well. I think it’s probably useful/helpful to listen to the individual headphones in a sampling though, using virtual techniques like the above, to see if they really have the kind of sound sig you desire. Or something close to it.
I listen to music. 
1 Like
It’s the universal language.
Absolutely.
Getting back to EQ, recently pulled out a Sony MDR-Z7 out of retirement. This was one of the very first headphones I bought when I got back into the hobby way back when. I’ve kept them primarily for sentimental reasons, but also as a reference against other headphones. They were fun, low distortion, warm, but suffered issues associated with closed headphones. Attempts at EQ were a disaster.
Fast forward to now, tried the Z7 with the Sonarworks EQ app. Pleasantly surprised at the improvement the Sonarworks EQ provided to the Z7. The sound from the Z7 with the Sonarworks EQ applied makes the headphone quite enjoyable.
So now, I’ve two closed back headphones that sound quite good with the right EQ applied. There are some advantages to closed back listening IF one can get the right EQ applied.
Caveat Emptor.
1 Like
I don’t think even 31 band fixed freq EQ can do many headphones certainly 3-7 bands cannot. However a 6 band PEQ and for sure a 10 band PEQ can.
I don’t like the Harman - too low bass heavy and too much in the upper mids/lower treble, but its a good guide on where to start. I also add minor changes
for planars that have issues with CSD. I am getting a used pr of HFM HEK Stealth which need that 6-8k range cut, 1.7-2.7k and 20-60 hz raised, and it
becomes an excellent can for $525. I also use it for my HE-500 and HE-6 SE and Senn HD-600, but not my speakers,
CSD is likely a red herring. Even when it looks super raised with long tail elements, that’s just an FR peak. It sometimes looks like there’s something to worry about there but they go away when the FR peak is fixed, and since headphones are generally minimum phase, there’s no reason to even consider CSD most of the time. Maybe considered excess group delay first to see if there’s anything there.