[earmonger]

If I am looking at the 20Hz region on the graph and one line is sloping downwards, let's say to 50dB and the other line is sloping upwards to let's say 65dB, is the 65dB line representing more bass? RVG, definitely the man!

Yes. The higher the decibels, the more bass (or treble or midrange).

 

[castleofargh]

Earmonger, you're the best, thank you. If I am looking at the 20Hz region on the graph and one line is sloping downwards, let's say to 50dB and the other line is sloping upwards to let's say 65dB, is the 65dB line representing more bass? RVG, definitely the man!

20Hz is too low for most people to hear. With a subwoofer, you get the tactile impact of such low frequencies as they literally shake your body, so you wouldn't want a lot of that anyway (it's common for sound engineers to set a high pass of 6 or 12dB somewhere around 30Hz). With headphones, it's more complicated and in general 20Hz is unlikely to be heard or perceived. Some circumstances can cause a change in our experience, like massive SPL, and the headphone itself shaking from such low frequencies. You could get some of that as a mechanical sensation through the pads and headband. Physical vibrations(at any frequency) can feel like more bass at some other frequency where you'll hear the actual bass( Floyd Toole talked about that). Or It could be elevated distortions creating harmonics at 40 60 80... and that's just from the 20Hz tone, with music a lot more would be generated from a headphone with significant low frequency distortions.
It's complicated to just declare that one or more of those things is what we're feeling, I'd still argue that plain frequency response tends to explain most of what people will describe in soundstage, dynamic, detail, or speed terms. And just because you see 2 headphones with similar bass profile, does not mean you will or should experience the same bass. First because our experience of sound is about relative values, Bass alone cannot tell the full subjective experience of bass. The rest of the frequency response does matter(more so when most impacts starting the bass rumble are found at much higher than frequencies). And second, because there is no certainty that you're getting the same seal quality as the dummy head had for both headphones(who does the measurements, how stiff or big are the fake ears, and how much effort he puts in getting the best seal, those are all relevant variables that an out of context graph tell you nothing about).
Loss of seal is basically a massive high pass filter put on your headphone. The actual impact depends on the headphone, a general rule of thumb for me is that the bigger the driver, the less the roll off when the pads don't seal well.

So we can say that measurements are not enough, and even if the specific variables measured in one graph (like FR or THD) are accurate, they're accurate for that dummy head. Your head is different, My head is different from your own. You'll place the headphone for comfort, the dummy head never brings that up ^_^.

Now, all that is true, and there are even more issues that people who have never done those measurements can't even imagine. But to be perfectly clear, when you read the feedback from someone trying a headphone, you get all those same issues(beside the comfort thing that they might share with you), plus the massive amount of data reinvented in subjective translation, plus sometimes massive variations in the rig used with the headphone. Someone who does measurements will not feed the headphone with a 100ohm 4%THD old tube amp. But someone just giving his feedback might. He might even only like that headphone on that amp.

If you can go and listen to all the gear, do that, of course! It's a listening hobby, listening is the way to decide how we like something.
But if like most people you don't have infinite access, then you do end up having to rely on measurements and reviews/feedback. Most of the issues being found with both options, I think it's important to try and stick with one of each. If you work around measurements from one source, you will learn about his measurement method, his dummy head, coupler etc. but more importantly, he will tend to do things consistently, and that in turn will give you more reliable variations between 2 measurements. Which is what you want to work with. Relative variations between 2 or more headphones, where some of those headphones are known to you.

Same thing with a fellow Headfier/reviewer. You find one who seems to share your feelings and taste in headphones, and you stick with that guy. Over time, you learn about his way of describing sound, and what to be wary of. If you rely on 10 people, you'll get 10 views or 10 different people with 10 different headphones, 10 different HRTFs(how real sounds around you are changed by the shape of your head, which you think is how everything sound, but that's true only for you). And to make it worse, they might all have a different meaning for the same words. You find that with soundstage for example. One guy will talk about a big soundstage when there's a lot of bass. Another might just happen to have the headphone's FR match his natural FR for 90° so he'll get some illusion of more distance to the side and say the soundstage is great, even if maybe he gets the center inside his skull(pretty much my experience of the HD800 BTW). Someone might talk about good soundstage because he can easily pinpoint different instruments, something probably due to more treble content(perhaps too much for most people's taste).
And so on and so forth. Same with people telling you that a headphone is neutral. I can't think of a more BS way to describe a headphone. There is no neutral for a headphone anyway.

And I would say that's my main argument in favor of measurements. Other people's impressions are even less reliable(and that's assuming everybody is sincere...). But again, if you can go and listen to headphones, do that! It's the best option and also the best way to learn to correlate your experience with measurements and feedback. You're your own reference. You have to be, as you've never heard anything else.
I do advise people to fool around with EQ a lot. It's a great learning tool, if you start describing sounds with frequency areas instead of BS audiolingo, you'll communicate better. And of course, over time you can learn to EQ your headphones to what you specifically prefer/need. It might be a tiny change, but I have yet to find a headphone and think it couldn't sound better with some EQ. Once you're ok with EQ, it becomes even easier to tell if you're hearing like a certain graph, and if not, where is the difference, if that difference is found on several measurements from that guy. I think that's when you get the most out of a bunch of FR graph.

You didn't ask, but THD tell you almost nothing about how you'll enjoy a headphone. If you get something with above 1%(-40dB) in the midrange, just forget about that piece of crap headphone. Otherwise, it's not important and will not match well with your own preferences in any clear way(It's not me saying it, it's actual research acknowledging the fact). Other distortions are more relevant, but nobody measures them(I, for one, don't know how to).

 

[Ultron 5]

Hi Castleofargh, thank you for such a thoughtful, sincere response. ". . . listening is the way to decide how we like something." I could not agree more with this sentiment. You said a lot here, so I will take some time to digest and respond. Sometimes, I just want to know how to do a thing, so that I know how to do a thing. On that note, I am going to post an FR graph with my interpretation to know that I understand how to read a graph. Then, I feel I will be a better position to sift through the malarkey and pseudo-science.