FR graphs - neutral, yet below 0 dB?

You're interested in the shape (flatness), not where it's placed in terms of vertical position on the graph.  Shift the entire curve up or down in your head if it makes you feel better.

Actually, I have no idea how he's placing those plots.  I've seen others normalize all their graphs so the reference point of 0 dB is always the level at 1 kHz, but I don't see any convention to what's at InnerFidelity. Maybe I'm missing something?

Quote:

Originally Posted by Siva108 

But isn't it true that the further above/below reference the line is the more over/underemphasized a particular frequency is?

Yes but the zero on the y-axis is arbitrary, one simply has to turn up the signal amplitude to move the curve up.

The exact shape however will likely change slightly (and drastically as you approach the SPL limits, I would expect). Having a convention of providing sufficient input power for a response of x dB (or pascals, whatever) at y Hz would be a good idea.

To put it simply, the vertical positioning of the curve as a whole has no importance whatsoever, what matters is simply how range x compares to range y in the graph, it's ALL ABOUT THE BALANCE of the amplitudes of the frequencies AGAINST each other in the same graph only, not if headphone A has the amplitude something vs headphone B which has another amplitude at the given frequency range. To understand what's concidered recessed or emphasized in a headphone's graph you need to introduce an "average frequency amplitude" for the graphs which is simply the average amplitude value which I'll show an example of in a few mins in the given FR graphs you posted above.

Limegreen line represents the "average frequency amplitude". I don't find any use of taking into account below 20Hz and above 20kHz as it may just skew the results, especially for open headphones which usually lack bass extension versus closed. 

On V6 it has probably among the most flat measuring headphones I've seen with the most neutral "tonality" I've seen, probably why it's found in so many recording studios. Due to the slight dip after 50Hz as well as around 6kHz area (but should be mentioned Tyll's measurement almost always shows a dip in this area so it may just be something limited to this specific equipment/test procedure of his and might even be quite straight in reality which is my guess due to almost all headphones showing this dip) the line should be drawn slightly under the rather straight FR plot in the 50 - 5kHz or so range. Well if you want to be extremely picky you could perhaps make the line possibly ever so slightly upwards tilting due to the bass extension roll-off and 9~10kHz spike but that would be such a small tilted angle I just preferred drawing it straight in this case. Anyway it looks like there's like ever so slight emphasis around midrange and even smaller so in the 50-150Hz or so range but from my experience a 3dB variety in the FR graph isn't really going to stand out and will still sound pretty flat to our ears, might be more closer to 5dB or so variances when it becomes noticable, the peak at 9~10kHz is probably slightly noticable. Anyway this is probably the most flat measuring headphone in its price class by far, with only a slight roll-off below 50Hz and above 10kHz (I'm not convinced enough about Tyll's 6kHz dip as it's so frequently showing up and I've seen some other plots where it wasn't the case either so)

HE272 then again has a slight warm tilt in tonality (line is tilting forwards) versus the V6 with a more significant midrange spike in-between the 1000 - 2500 Hz range which is most likely audible for a more experienced ear probably with certain female vocals for example. It also has a relatively similar spike around 9kHz as V6 (well most headphones have a slight emphasis in this area and this is more the norm than an exception and a 5dB spike here isn't anything unusual at all). But yea it's fairly flat in the bass range overall and has an emphasis in the 1000 - 2500Hz range and recessed/rolled-off slightly around 2500 - 8kHz range and above 10kHz.

Quote:

Originally Posted by Siva108 

WOOT I tried drawing the average frequency amplitude line on the MDR-V6 graph without looking at yours and it came out the same. I GOT IT.

 

I really should have taken math for my final year in school lol. :p

 

Thanks a lot bro. You cleared all my doubts and made it so much easier to compare frequency response.

 

Edit: One other doubt, is there any reason why sites like Innerfidelity don't include this with the datasheets?

Well there's no written in stone rule about this, but since FR is really about balance (you make one part more recessed and the rest of the range will automaticly get louder/more present as a result or vice versa) it makes sense to compare to an average amplitude of some sort, but unsure what's the proper way of doing it calculation wise etc, I just measure roughly with the eyes.

Oh and the starting position is probably a bit awkward on the 2nd graph for the HE272 I showed (should maybe be like 2dB lower or so) but yea at least it gives you an idea.

I'd say that's roughly it as you drew it. 

No, there can be headphones which has FR graphs which looks like a rollercoaster and have like 10dB varieties yet the line could perfectly straight as the dips and peaks could be equally large and therefore averaging out themselves so the result is still a straight line but what I think it does is that it gives a hint of a headphone's "tonality", is it warm or bright sounding headphone. I'd say already rather straight line tend to sound relatively bright such as the V6 to most people's ears already.

downwards = warm

upwards = bright