oluvsgadgets
patreon
Oluv's neutral headphone target
🕑 Added 2023-07-14 14:20:32 +0000 UTCI am starting a completely new post on this topic instead of updating the existing ones, but as I started with a competely new setup this time, which also led to different results than my last approaches, I think this needs to be celebrated with a dedicated post.
First I tried setting up a kind of listening setup in my office room at my desk, but the acoustics was not suitable for that task and the KH80 despite some equalisation didn't sound correct to me with that placement Thankfully I was allowed to build up the KH80 in our living room again for more than a week and create a representative and consistent listening situation without interferring too much with our daily life. As we spend the hot seasin mostly outdoors I got the permission from my wife to waste my time a bit more with something which obviously never comes to and end. During this period I just had to relocate the left unit every day but as I had measured all the distances from the stands to the sofa I hopefully managed a consistent setup with some possible slight variance added for maybe even better spreading. Without having a constant reference it's close to impossible to create something faithfull because the human ear adjusts immediately to any change in sound, thus you need to switch all the time between your original original source and your target to get something correct.
This time I placed the speakers even closer at exactly 1m between me and the speakers in order to minimise room influence even more. I didn't want to eliminate room reflections completely, as you always listen in some kind of room, but due to the nearfield placement those reflections are rather kept at a minimum and should represent a realistic in-room-response. With the given perceived spectrum sitting inside the sweetspot of the stereo triangle I was again trying to approximate the same sound through the HD58X as close as possible. I spent some time to achieve this adding lots of breaks between different listening sessions to reset my ears and brain every time I started a new listening session. This helped me a lot to perceive some additional nuances which tend to get lost during longer listening periods because the brain immediately adjusts to the given sound. But the effect of switching from the speakers to the headphone should be as similar as possible, although I am still not quite sure about some particular frequency ranges in detail, I think the overall character I am hearing from the speakers is represented really faithfully through the headphone as well, although not 100% the same, it sounds quite alike. Less with pink noise than with music as you can hear from my video. I think the sound sample is captured really well this time. I tried doing some EQ matchings of the measured responses but the only way to get both sound similar was by endless listening hours adjusting single filters bit by bit just to notice after a while that one previous version was already better than the current one, having to get back to that one and start from here again. Thus, nothing new so far, it's just the same procedure like all the times before.
Here again some measurements of the recent listening situation, KH80 equalized in black and the measured response with my ear sitting exactly in the sweet spot without any stronger movement and no averaging to other angles:
The biggest issue I had to deal with was the huge change in sound from the speaekers by some slight position changes only, while I can just recreate a single static spectrum over the headphone which again will change a lot but to a different degree influenced by headphone positioning, shape of pads etc...
Below you can see additional measurements of the KH80 captured with my ears by just moving some cm to the front and back on top of the measured sweet spot in orange:
Instead of concentrating on performing lots of averaging this time, which means taking lots of position samples, also re-placing my microphones for every single measurement and results in much more time needed, I took some short and simple but most important measurements and only used them as a rough guide for my adjstments, because what may have looked as perfect headphone response matching that of the speaekers very closely, may have looked again wrong compared to another sample measurement. Overall I think despite the deviations between the responses for the KH80s and the HD58X, the perceived sound I am hearing is quite similar and mostly deviates through some frequency ranges which I simply couldn't reproduce by the headphone properly for whatever reason.
Below you can see the measured KH80 together with the final response for the HD58X and also unequalised in blue:
You should notice that I follow the response of the speakers with the headphone pretty well, except deviating in some areas, thus I didn't fill up the 3khz range that much otherwise it sounded wrong to me, also the 8khz dip despite raised a bit compared to stock it would sound just too resonant if matched to the speaker's response, also that peak around 12khz measured from the speakers didn't sound correct when I tried to simulate it over the headphone. But with this response it sounds quite close to what I hear from the speakers and equally smooth at the same time, just missing that weird phantom frequency which I couldn't reproduce properly like already mentioned before. That phantom frequency disappears when I move my head to the side, like I demonstrated in the video, but as soon as I listen frontally I get some Hhhhh kind of sound, which I just cannot simulate by the headphone. It is hearable from the recording, thus the headphone can reproduce it, but I cannot manage to hear it with pure pink noise, therefore it must be some side effect of the phantom-center, really no idea.
I let my wife and son listen to my EQ comparing that to the speakers and both found my EQ to sound closer and more realistic than the HD58X unequalised. To perform further experiments I modified my ear and did additional measurements to check how much the response of the speakers and headphone would deviate in all cases.
Below you can see my ear mods by putting some putty inside:
The corresponding responses of all 3 mods for the KH80, the HD58X with EQ and unequalized in blue:
Although the deviation of my EQ becomes stronger with the mods, it still sounded closer to the speakers than stock. Thus it should translate quite well over to others' ears as well, although probably with less precision and less refinement.
I let Oratory1990 measure my HD58X unit on his GRAS system, which may be interesting to compare to a measurement done with my own ears. As I measure at the ear entrance point omitting the ear canal, there is of course some stronger difference, GRAS in blue, my ears in red:
Finally I can show you how the HD58X with my EQ would measure on a proper system, which you can see below together with the Harman target:
Even with my additional filters simulating a natural in-room response, the bass is way lower with my EQ, I personally find Harman way too boomy, not even talking about the out-of-balance rest:
Apparently I am way closer to the free-field-target, which you can see below:
The biggest difference to the classic free-field is that I didn't measure in an anechoic chamber but inside a room albeit nearfield, and I measure 2 sound sources with incoming sound at 30° to the sides, while free-field measures a single frontal sound source.
Some stronger smoothing applied you can also see my target compared to other targets like diffuse field in green, Sonarworks in red and RTings in blue:
The closest would seem to be the Sonarworks target, if I apply my in-room-response simulation it would look like this:
As mine sounds more or less correct to me, the balance of all frequencies must be slightly off at Sonarworks with too much upper bass, too much response at 1.5khz, too little in the ear gain region and too much at 7khz, which may sound more correct to other's ears though.
I would welcome all to try out my EQ for the Sennheiser HD58X and tell me your opinion how you perceive the sound. I am not sure how well this may sound to others' ears and you may need some additional corrections for perfect results. I offer this EQ currently for the enthusiast tier. Keep in mind, this is just to get as much feedback as possible from different users. If the response will be positive I would like to offer EQs following my target for other headphones as well, starting with the HD6XX and later also HD600 with new pads choosing another headphone each month. I am not sure I can offer this for so cheap, as the effort is pretty high to achieve a high level already for me, not being even sure if you will perceive it the same way. Therefore I need your honest feedback to know how to proceed.
Comments
Dragan Petronijević
You can use HD 58x CSV from AutoEQ github as an target emulation & than download + apply Oluv's target for HD 58x on top of the HD 58x emulation convolver. (I use EasyEffects, If you use Windows, you will likely use parametric EQ for EqualizerAPO/PeaceGUI) That is what I did for my Anker Soundcore Q30 to get the approximate Oluv's target for it & it sounds amazing! Harman sounds so lifeless, cramped & robotic when I listen to it now.
gavin xq
I use autoeq.app to apply target curve to measurement form oratory1990 to generate eq setting. And is there a csv file of Oluv target for me to download? Currently using Harman target from 2018. Thanks.
Back and stronger than ever
