Measurements are useful compared to some reference. It's the same old but important concept of frame of reference for physics. When you see a headphone measurement, in itself it tells you nearly nothing. Worse, it could be highly misleading because different measurement rigs and compensation curves can make the graph look like anything and everything. If we stick to frequency measurements of headphones where you use some headphone, you know how it sounds to you, and you're curious about some new headphone. Try to find some place online that measured both the new headphone and the model you own(not one graph from one place and the other from another place, that's crap, don't do it!), then it becomes very interesting to put those 2 graphs side by side. Because both measurements were made on the same rig by the same guy and because you know one of the headphones subjectively, you possess the right references for some interpretation.
And that's where
@VNandor's suggestion about EQ becomes the translator you need. The graphs tell you the new headphone has +3dB at 100Hz(suggesting more bass), but to get a subjective idea of what it sounds like, you take your headphone and EQ it with 3+dB at 100Hz. That's the ideal situation where your headphone can handle the EQ. In practice, that kind of works most of the time, except for very high and very low frequencies(where the driver may hit a mechanical limit). In practice, that's going to be useful to you anyway as a headphone A vs headphone B approach. It's not 100%, nothing is here. Those graphs aren't what your eardrum would get from those 2 headphones, so we start with missing data from the acoustic impact of your own ears. We also tend to assume that My headphone XXX will have the exact same FR as the same model measured somewhere. That's nearly never true. Having 2 or more dB of variations between headphones is typical, unless the brand specifically spends time and money to have a tighter production line.
We're just trying to get a general idea. I've seen too many people draw nonsensical conclusions because seemingly similar FR graph online didn't agree with the 2 impressions they had from their own pairs. FR is doing most of the legwork when it comes to subjective impressions, unless the distortions are huge or the sealing effect(cutting off outside noises) is big, chances are that FR is in fact causing most of the impression one way or another. We just rarely have access to the FR graph most meaningful to our own listening circumstances. I've done measurements at the entrance of my ear canal and that's already such a change compared to my cheap and mostly DIY measurement solutions.
So keep in mind, FR graphs can be informative, but remember what they are, which is usually, not your particular pair and not the FR at your eardrum. Audiophiles love dealing in absolutes, reality isn't that convenient and measurements(doing and understanding them) make it clear.
Back on track, just to get a better understanding of the relation between FR and your subjective experience, it is a great idea to just fool around with EQ and get the experience of how that 5kHz area can change what you feel. The more you do it, the easier it will be for your brain to associate some feeling with a frequency, which in turn makes it so much easier to communicate(instead of endless flowery lingo that means something different to each reader and has no actual reference you can fall back to). If you have a bunch of tracks with a single instrument playing, that is educational, IMO. Finding the actual range(like how a drum isn't just a sub sound), or isolated voices can really be interesting to find out how much better they might feel to you with a particular tuning.
To learn, I think using big EQ changes works well. Then later on, when you try to actually improve the FR of your headphone to something you'll enjoy more, it becomes important to consider that you're certainly doing too much. It's just what all beginners do with all the digital toys, audio or not.
---------------------------------------------------------------------------------------------------------------
If you're not satisfied with online measurements and wish to do your own, first remember my warning before, even with the best efforts, you're not going to get the exact measure of what you'll hear because that's not what you will measure.
Doing your own FR graphs could be simple or something so involving it would need to become your new job. With the idea that you have 2 headphones in your house and want to measure how different the frequency responses are, almost any mic and method are good enough. Some cardboard with the mic going through it, and you're in business already.
Next will be your own battle to find how to get repeatable results. Drawing circles on the cardboard to place the pad consistently, or any other trick you can think of to steady the all thing while you proceed with the measurement.
That will teach you how easily it is to get variations and what to trust and not trust too much when you see online FR graphs(basically, seal quality= amount of low frequency, while small changes in placement and distances = big changes at high frequency). Your job being to try and control both as best as you can so that consecutive measurements and measurements between headphones have as little variations due to the measurement itself as possible. That's what makes a measurement valuable or not. If 10 measurements give you 10 noticeably different results, you're not measuring, you're throwing dice. Repeatability is important. Again, impossible to get to 100% but still, more consistency means a more reliable graph.
Going down that road is interesting only if you have or are going to have a bunch of headphones to play with. Getting into all that for 3 headphones, it's not worth it IMO.
And while we're discussing the limitations:
All along I've been talking about relative differences, with and without EQ, headphone A vs headphone B. If you're thinking of making the headphone "neutral" or seeing when the FR is just right objectively, give up! The FR you measure is not the one someone else with another rig will measure, and is also not the FR you will get at your eardrum, which itself is going to be different from the FR at my eardrum when wearing your headphone. So objectively there is no concept of universal neutral for headphones. The reference literally changes with any hearing/measuring system. Some accepted references are the top most expensive measurement rigs used by most professionals. But even that is only relevant as a useful reference. It won't magically tell you what is flat for your ears, only you have access to those ears.
If you plan to measure THD, you will need better everything. The FR is measured as the peak(or RMS) value at a given frequency, so you send a loud clear signal, easy to pick up, easy to isolate from noises and distortions. Easy!
While you start looking at distortions, on a good headphone you'll get most THD below 40dB under the signal that would be measured for FR. You now need to be 40dB more accurate, more sensitive, and of course you need to deal with all the noises that might be just as loud or louder like typical noises in a room. Car down the street, computer fans, people in the house, anything can try and ruin your measurements.
As practical options:
-A cellphone into a cardboard box, so the mic is on the side where the headphone driver is. Any free RTA(real time analyzer) app, you play some noise into the headphone(pink, brown, whatever looks best to you for the type of analyzer setting) and take screenshots or whatever. Maybe just to try things, that's fine and educational.
-Any cheap mic and again some cardboard or anything you come up with to "simulate" a human head. Here, the biggest challenge might be to have the right plug. The old cheap ones with a jack might have the rings arranged differently depending on what they're intended to be plugged into and as a rule of thumb, You probably shouldn't trust the details on an Amazon page.
Anyway, those mics plugged into a phone with the jack-to-usb adapter the phone should have, could in some cases be easier to use than the cellphone alone. If it's one with a cable you plug in a computer, the longer the cable the more noise you'll get(because those usually aren't balanced), but again, for FR that's going to be very fine, just play the headphone a little louder. I suggest you verify the loudness yourself, if it feels too loud to your ears, you're probably close to the limit of what the headphone should play. Nobody wants to ruin a good headphone for a measurement.
-A "real" measurement mic, with an XLR plug and an ADC. Or one of the USB mics where the ADC is in the mic. Again, for headphone FR graph, both are great if not already overkill. For more advanced measurements(stuff about phase/delays, the all-in-one USB options could be an issue sometimes).
There are apps for cellphones. I don't remember which ones I tried, but Google is your friend. Even on Headfi I remember some discussions about that(probably for IEM measurements as that's what I feel into myself, but FR graphs are FR graphs). I'm using mostly REW(Room EQ Wizard) on my PC. It does most of what I can and want to test. It's a speaker measurement tool, but again, it's still FR measurement. For free, it's an amazing software. It also has some EQ oriented options, Not exactly intuitive, but they exist and the difference between predicted and measured FR changes are impressively accurate.
- actual professional stuff with ear simulator.
If at some point you get serious about measurements, they'll hate me for it, but contacting the reviewers who provide such measurements is probably your best source of all around knowledge and tips. Not everybody will become your best friend, but someone will probably be willing to help you a little.
@oratory1990 has been spreading his knowledge free of charge, here, on reddit, and just about anywhere else. looking for those posts would be a good way to learn about a bunch of things relevant to FR, measurements and just acoustic in general. And if he didn't get fed up of doing it yet, if you send him your headphone, he'll measure it.
Personally, for headphone measurements, I've used a mic into some cardboard once or twice and I own a Minidsp E.A.R.S(killed the capsul on one side since), which is simple enough to use but is absolutely not a convincing ear simulator!
Also I kind of got a HD650 a billion years ago and stuck with it ever since, making my own interest and knowledge in headphones nearly non existent. Apprently that doesn't stop me from making a 50 pages long comment anyway. oops.
