Trust your ear if you know your resonance... (Patreon)

Hey guys, I wanted to give you a short update of what currently is happening as I was quite active lately and didn't have that much time to do anything else, although I also managed to shoot a video inbetween, but will need to edit it first. The tuning progress so far is really promising. My theory of dedicated resonance-based tunings seems to work like I could experience on my own today.

I was at the Otologist yesterday where I also shot another video covering this topic more in detail, i hope to be able to post it within the next days. I wanted to experience the sound of my tuning before and after ear cleaning conciously, which was quite shocking. I had the Earfun Free Pro with my latest tuning with me which was based on my completely new created target that took several attempts to approach, as it is really a matter of finding the correct response curve. This is only possible if you know exactly "how" you are actually hearing and "what" you are actually measuring. It's a synergy of both to really get a curve which seems to be valid over all resonance frequencies. No surprise all published targets are just off, even the one Axel Grell seems to have used for the TWS/1 is probably based on his own studies, although even his is really off, it just doesn't sound correct. All this is not a 1-hour work, the whole process took some days. The ears need some rest after longer listening sessions, there is a point where you simply cannot judge the sound correctly anymore. Have you ever tried analysing a sine sweep and then playing music right afterwards? This is how the very first Bluetooth codecs sounded...

But actually the result I got from my new target curve was quite stunning, it is different from any IEM targets that are usually used and hyped by most reviewers, let alone taken as reference for creating EQs etc. This makes the whole IEM measuring madness appear just absurd. Please understand that I won't post my target curve public as I simply don't want my hard work to be taken and copied by others. It's actually just a simple curve, but getting there was a real pain in the ass, thank godness it payed off as the timbre I finally managed to achieve is pretty close to a neutral speaker playing in front of me or an overear headphone equalized to flat (e.g. with Oratory1990's Optimum Hifi EQs to be found here: https://www.reddit.com/r/oratory1990/wiki/index/list_of_presets)  

I started the work on this target with my "old" ears  And I was enjoying the result really a lot, the effect of switching from a speaker to the Earfun with this tuning was amazing, it just sounded "correct" But I knew this ear state won't last forever and decided to let my ears clean already yesterday and not wait until next week.

As suspected my "best" tuning for the Earfun Free Pro based on this new target sounded quite different after my ear cleaning. Not really bad sounding like most IEMs, but just not really great. At home I decided to try all main tunings I did so far and my brain got completely confused, none of them sounded really great, my original OE tuning was maybe the worst of all. I thnk the biggest problem is getting used to a particular sound signature without having a "correct" reference in mind and my brain obviously still had some shock having to deal with a ahifted ear-canal resonce from one minute to the other. I think an average joe won't notice, but if you deal with sound the whole day, any change can be drastic. But the human auditory system can be a mystery as it can basically take any frequency response for granted, after a bit of time which is needed to get used to it. Experts call this “burn-in”. This works well as long as the response is not full of peaks or other anomalies, therefore judging the sound for me with again different ear resonances was quite a funny experience, nothing really sounded Ok to me, not even the speaker playing in front of me, so I decided to sleep over it and try the "real work" the next day. 

I wanted to prove my theory if the previously created target which I achieved with "wrong sounding" ears would still be valid with my current ear state, therefore I tested my current resonance frequency and got about 6250Hz right now, not 6500 like I always suspected. Either the resonance frequency tends to vary continuously, therfore you shouldn't be suprised that your heapdhone can sound worse one day or the other, or I simply didn't verify it precisely enough back then, but right now I could track it down to 6250-6300.

I created a simple EQ-preset to be used externally (you can also use it in the Airoha app but it's not as comfortable as from an external EQ-app where you can adjust the frequency with some slider in real time and immediately hear the result). This preset should allow you to concentrate on the whistling kind of sound the peak will create. Just activate the 2nd filter and move the frequency as long as the spectrum somehow starts appearing completely even, the preset will filter high and low frequencies so that you can concentrate on just this single peak:

Filter: ON HP Fc 3500 Hz

Filter: ON PK Fc 6250 Hz Gain -12 dB Q 4

Filter: ON LP Fc 10000 Hz

Some suggested to use a sine sweep for this, I can highly recommend this online tool, just be careful not to stress your ears too much and keep levels really low to avoid any tinnitus as pure sine notes at higher frequencies can screw up your hearing completely. You won't be able to hear "correctly" for quite a bit of time, after such an "analysis", but finding the peak precisely is crucial to get the correct tuning for you as otherwise you may burn in your brain with the wrong response if you don’t choose the right one.

After having found my new ear canal resonance I made a measurement at the given frequency and just tailored the sound to my new target curve. I still had to fine tune it a bit by ear, but the result is very close to what I heard before ear cleaning if not even better. I am currently enjoying the sound I created as I think it approaches a neutral timbre as good as possible with the given limits of having only 10-filters. Tuning can be a bit frustrating like that and some slight errors will still remain as I won't be able to correct for them, but they are far less obvious than not having a smooth response and a plausible overall timbre.  Music will sound as realistic like it never did. You will just really just need to select the right tuning for your given ear canal resonance. At first I made the mistake not having verified my resonance frequency and was still assuming I am at 6500Hz, but the EQ based on the measurement I did at 6500Hz sounded just off and I already suspected my theory would not work in practice, but actually after correcting the frequendy down to 6250Hz and creating an EQ again, the result was much better to my ear. Each tuning must really be based on a particular resonance frequency, because the sound will change just too much with different insertion or ear canal lenght etc. especially in the crucial region around 5-10khz, that most headphones just struggle to render correctly but which is so important to get this "realism" 

I will soon post my first EQfor 6250Hz resonance, based on the new created target, I just need to sleep over it and give my ears a bit of rest, but after about 1 hour of listening with this tuning now it sounds really amazing. You need to give your brain a "reset" provided that you are really using the correct tuning for your own resonance, as in the beginning it may appear "weird" sounding, the brain needs some time to get used to that sound.

 I will then continue tailorying the sound to further frequencies, having a real constant target this work should be much easier than before, where everything was just an estimate. Shall I boost here or there, but particularly the area between 3-10khz seems to be most important to get the sound "realistic". You can basically get used to any kind of response if it is not too far off, but once you get it "correct" you won't be able to get back to anything else. Most other headphones differing from neutral will sound just wrong. The effect of listening to a speaker and then switching to my current tuning is quite revealing. It is a completely new experience. With my current work I tried to close the infamous circle of confusion. The music should sound more or less the same over a speaker, like over a headphone and an IEM. This should also be valid for my own binaural recordings as they are also based on the flat EQ profile for the HD600. Of course I cannot guarantee that you will really get exactly the same result as me and although I tried to be and listen as precisely as possible, there can be still some slight differences between a neutral speaker, a neutral headphone and my "neutral" IEM. Bass amount is still a bit vague i wanted it be rather towards the neutral side although from measurements it would rather appear bass heavy with my setting, you can always add a bit more or less bass to your liking; my EQs should be also a playing ground for your own experiments; but changing one single filter can have huge effect on the overall response. It can all be a bit frustrating, that’s why it took my so long to come up with something usable.

Please give my some time to prepare further tunings, my latest 5850Hz one seems to have got some positive feedback from at least one of you. It is really just a matter of finding the corresponcing tuning which goes along with your own resonance frequency. 

I will also need to check a bit for sample variation, because I noticed that my new EQ can sound a bit brightish on one of my units, while my reference unit sounds more balanced. There is about 3dB difference in the upper frequenices and I still need to figure out if I should create an average, or just ignore the spikes. I will need to apply my new EQ to all units I own and just try measuring them precisely with the same resonance frequency. As soon as I have some of the EQs I will also finalze the giveaway unit which I simply didn't want to send out half ready.

So stay tuned for more...