Free field as a baseline for closed-back headphones

[qrayzie]

The amp’s impedance is noticeable when it comes to transients, as they can sometimes be fairly sharp with my headphone. I can also get these fairly loud and still be listenable, it’s just that because of the impedance of the drivers the distortion starts to muddy up/smear the sound. Still perfectly listenable however.

I’m also not sure that the headphone should be the point of discussion… My main point is that using the free-field as a baseline to form a target curve for closed-back headphones is closer to ideal than the diffuse field. I have my headphones EQ’d to my target curve, and so the default sound signature characteristics are not present as far as I am aware.

Constant A/Bing and tweaking of both free field and diffuse field baselines and targets have yet to convince me that the diffuse field is a better use case, especially when free field sounds much more accurate to real life to my ears. I don’t believe my headphones should have a strong bearing on this, especially when I know fit/seal is not an issue and that they are very responsive, reliable, and predictable when using EQ.

[ADU]

I appreciate all your thoughts on this qrayzie. And can certainly understand where you’re comin from. As I mentioned above though, it’s difficult to talk intelligently about this subject without some actual in-ear measurements of speakers in a room for comparison. This is still, for me, the biggest missing piece in the puzzle.

Free field was tested (a couple times, I think) in the Harman research, and was one of the least preferred in-ear responses they tested. A “flat” DF curve also didn’t do well. Most found it excessively bright. I don’t think Harman tested DF (or FF) with a slope though. At least not until much later.

Many of their early tests were also done on a flat plate GRAS rig that did not have an actual head or torso that could be accurately measured in a diffuse or free sound field. So that is something else to bear in mind.

If you’ve read or listened to some of Floyd Toole’s talks, then you’ll know that speakers also do not really behave like a free field source, with sound emanating from a single point or direction. The direct sound dominates in the higher frequencies. But they radiate sound in all directions at lower frequencies. And the room will also reflect and diffuse the sound to some degree at all frequencies. This is why I think DF, with a slope to account for the broader radiation in lower frequencies, is probably a somewhat better model or approximation than a free field (esp. with no slope). The sound of speakers in a room is really a blend of all the above though.

Your own listening experiences can also be an important factor in this. If you’ve only used speakers primarily in a near field, or in very well damped rooms, then a free field might very well sound more like what you’re used to.

I’m still working on some EQ curves for the M50X btw. It’s just taking awhile because I’m looking for more and better samples for my target curve. It’s interesting to compare all the neutral samples to the M50X’s response though. And helps to put the headphone’s FR into somewhat better context imho. This is something you can do to some extent with Listener’s squiglink btw. You just select all the headphones you want to compare to the M50X so they appear on the same graph (with the + after the name). And use the M50X as the compensation curve for the other headphones to see the difference.

Instead of giving you a single EQ curve, this will give you more of a range of possible adjustments to try on the headphones.

[qrayzie]

Hey there, just chiming back in to say I’ve since upgraded to a new pair of headphones (Sennheiser HD 490 Pro), and I’ve found a target that I am very happy with. It appears your suspicions about the M50x and/or my HRTF were correct. Until I got this new pair there was a nasty 9kHz peak that was nowhere to be seen on any measurement rigs, I assume because it is a closed-back headphone. I can safely say to disregard this topic, and if anything, it’s a testament to why open-backs are generally the way to go. There was no telling of where the harsh treble was coming from on any measurements, so I was, as you say, just compensating for that inaccuracy and interaction with my personal HRTF (which is sibilance sensitive, in person). I appreciate your input on this topic.

Cheers

[SSN757]

I think the Sennheiser HD 490 may be one of the best overall buys in the headphone realm. Excellent choice.

[Nuance]

I quite enjoy them after some EQ. Having two set of pads to provide two different sound signatures is something I wish more manufacturers would do.

[DJ_Hastings]

I’ve had exactly the same problems as you OP. That’s what, for better or for worse, got me into this hobby in the first place. See, I had these old pair of headphones that were fantastic, but then they started to break down so I went to buy new ones, with the M50x being recommended. I found the M50x very harsh/sibilant and the music sounded like it was in a tiny, stuffy, recording studio. So I was like ??? Then I tried Oppo PM3, AKG K553 and K712, Neumann NDH20, Shure SRH1540, Blue Sadie, Sennheiser HD 660S etc. And even with EQ and lots of critical listening, I could never get that sibilant band to tame down enough while still maintaining correct tone. They were all unlistenable for my complete music library.

Since then, I’ve tried nearly every headphone in existence, and I’ve basically heard all of them this way, and had the same, continuing, issues with eq etc. Although I will say that proper DF measurements for a rig helped a lot, as especially gras rigs have way less energy in that mid treble than say the Harman target. So, frustrated, I decided to just re-buy a unit of my old “gateway” headphones from ebay, and they still sound fantastic, like a less sibilant Maxwell, and almost perfectly nail how I hear treble. So, I went and checked their KB50xx measurement against the Gras KB50xx 0’ 0’ free field with a -1db tilt and, wouldn’t you know it, but the treble above 5k matches exceptionally well. They are the Allen and Heath Xone XD-53 DJ headphones. Here’s what they look like against the aforementioned free field.

I think the 600Hz and 1.5 kHz features are for DJ monitoring purposes in super noisy environments.

So yeah, in general, your hypothesis seems to agree with my hearing, that, for me, most headphones need a very dramatic cut just before 10k for me to start hearing them as neutral and natural. The more I think about this, the more I’m like, well when you go to a concert in an open field, can’t this be somewhat approximated by a tilted free field, and those can sound great? idk. I’m going to hopefully do a bunch more listening tests with eq on other headphones to see if this pans out.

[ADU]

I appreciate hearing your thoughts on this above, DJ. I think different perspectives on this are interesting, and potentially educational.

One possible explanation is that your ears are just different, and you’re hearing some sibilance that others don’t. That’s not an explanation I favor. But it’s one that can’t be ruled out.

I’ve also noticed some sibilance on the M50X though. And believe it is due to some narrow band peaks in the treble, or thereabouts. Qrazie asked above how it’s possible for peaks like this to get lost on an FR graph. And the answer to that isn’t simple. Smoothing is one way. But they can also hide in the notches caused by the concha around the 9-10k range on some measurement rigs, such as the GRAS.

Smooth target curves, like the Harman response curve, may also give the false impression that the treble should be more smooth and orderly that it really ought to be. Resonant peaks and dips are normal in the higher frequencies, when measuring inside the ear though. The hard part is trying to sort out which ones are ok and normal ear resonances. And which are not ok, and the likely result of unwanted resonances in the headphones themselves. This is sometimes difficult for even more experienced graphers and graph readers to decipher.

There are also alot of misconceptions about headphones and their intended uses. Many studio headphones are brighter or more sibilant in the treble, for example, because it helps audio engineers (esp. older ones who’ve lost alot of their higher frequency hearing) to more easily pick up unwanted pops, clicks, and snaps when editing recordings. This is why most audiophiles DO NOT use them for pleasure listening. Because they’re too fatiguing on the ears for comfortable longer term listening, and tend to reveal too many of the warts in their recordings.

The Sony MDR-7506 is a very popular studio headphone, for example. Alot of instrumentalists like it because it has pretty decent bass and mids for a ~$100 headphone. Things start to get a little more unpleasant though above 2k on the 7506.

Many audiophiles are also older, and have reduced high frequency hearing. And may be trying to compensate for some of that with brighter headphones. This is not something I recommend. But prior to the Harman curve, most “audiophile” headphones were indeed on the brighter side, and a little closer to DF. And were often praised for their greater detail because of this.

Ear burn-in could be another possible factor in some of this. If you’ve listened to one type of sound signature for a very long time, it’s possible that your ears may be more comfortable or accustomed (“tuned”, you could say) to that type of response. And less sensitive to some of the resonant peaks in its response versus headphones with a different type of response.

Anyway, these are just a few more things to think about on this.

I know I’d find the DJ headphone shown above rather unbearable btw, because I’m most sensitive to sounds in the upper mids. And I like mid-treble that is more present than FF. And also good extension in the lower frequencies for bassdrops. I haven’t done any DJ-ing btw, but I think most DJ’s prefer a V-shaped response with alot of bass. Your XD-53 has V’s, but not in places that would be that helpful imo. I’m sorry, but I think the mids would just stand out like a sore thumb on that puppy to my ears.