I greatly prefer this presentation to the compensated version. It's much more familiar-looking to anyone who's been reading graphs for the last five years.
I mean this as politely as I can, but people are going to misread/misinterpret graphs no matter how they are presented. I think withholding uncompensated graphs out of a desire to protect the community from itself is misguided at best and kind of gatekeeper-y at worst.
You are, of course, free to prefer what you like.
That being said, I think as the people publishing data, we are all curating in some respect; no matter how objective anyone may
think the data is, it is always presented subjectively. We all make decisions with how we present—raw vs. compensated, where we normalize, what target we show, etc.—and as such there are always going to be preferential differences between reviewers.
However, the benefits of the compensated view are numerous enough that I'm almost positive most of the detractors who actually care to make things easier and more comparable for their viewers are likely going to switch to compensated (or calibrated, as we're trying to call it now) once they've acclimated. Remember, we as a community have already made this change before, when we moved from calibrated—specifically Tyll's "Independent of Direction" calibrated—measurements, to the "raw + target" paradigm we've had for the past few years thanks to people like Crin.
I'd like to briefly talk about one of my favorite things about this view:
comparability.
By calibrating measurements to the rig-relevant Diffuse Field Head-Related Transfer Function (DF HRTF), we can much more easily compare measurements on different rigs.
Let's say you have a 711 measurement of an IEM, like the Mega5EST mentioned prior, but you want to compare it to an HD 650 (which of course, uses a different rig) and another well tuned IEM like the Truthear Hexa, of which you have a great measurement from the B&K Type 5128 thanks to Crin. What are you supposed to do? Overlay them raw?
In my opinion, this view tells you basically nothing about how any of these actually sound in comparison to one another... which is a bummer.
However, if you calibrate each measurement to its rig-relevant DF HRTF, you are basically performing the equivalent of mic calibration—subtracting the response of the measurement microphone to—ideally—only show the response of the headphone or IEM on its own (plus whatever unique interaction it has with the rig, of course). Example shown below.
The above is a more perceptually-relevant result, in my opinion. The major gaps in bass and treble are closed to be much more accurate to my perception, enough so that I'd be perfectly okay showing this graph to a friend if I needed a quick visual aid to assist in explaining why I find the Mega5EST to have similar treble issues to Hexa above 10kHz, while finding them both somewhat similar in the midrange.
Another example!
Even if you just had the same headphone on 3 different rigs, you'd find that the difference might be exaggerated or untrue to perception. Below are three HD 650s measured on three different (but totally viable, though the HMSii.3 is kinda suspect at times) rigs. Is the pink HD 650
really 6dB brighter at ~7kHz than the blue one?
Well... no, it's an HD 650 and they should all measure rather similarly. DF calibration of these measurements (below) closes the gap in the advertised simulation range (up to 8-10kHz) to be much closer overall, narrowing that 6dB gap to be closer to 3dB.
Of course, rigs—just like different human heads—are gonna have some funky interactions with the headphone itself that won't be calibrated away with any speaker-based HRTF baseline... but you'd actually
want to see some of these differences to see how different people's heads might affect these headphones, right?
The "truth" of a headphone's performance might not be any one measurement, but it is
likely floating around somewhere in the realm of the multiple results we can measure on different rigs (heads) with DF HRTF calibration. And with more heads, we get more data to refine our understanding of how a headphone might perform in a variety of circumstances.
This is
just one of the many,
many reasons we have for compensated viewing being superior. We have people on our writing staff using a variety of measurement rigs, all of which provide useful data, and we want this data to be as comparable as possible for those reading our reviews because more comparable measurements means we can get more close to the "truth" of how headphones and IEMs may perform for a variety of users.
While of course, plenty of people are now used to the "raw" paradigm, they will get used to this too. Yes, because we're making them, but you gotta crack a few eggs to make an omelet. I hope this helped explain at least part of our viewpoint and why we're insistent on getting people acclimated to this view as soon as possible. Thanks for reading.
