This is quite possibly because some of these nulls are of the “difficult to EQ” kind and Oratory preferred to leave them alone (the 3.6kHz or so null for the red trace for example). Some of the residual difference is “on purpose”, in the sense that Oratory’s profiles don’t correct for it. On the other hand the residual differences are still quite important and audible, whether it's because of sample variation, profiles incompletely correcting the FR, HPTF concerns, pads breaking in / wearing in, etc. That’s particularly the case for headphones with a significant departure from the target, even more so the case when it happens in the range where ear simulator are the most accurate to everyone’s experience (such as the headphones with a strong dip at around 1500Hz). So Oratory’s profiles were quite successful to bring the headphones quite a bit closer to each others. I'm also tempted to suggest, given the magnitude of some of the residual differences, that listening tests alone should be enough to attest to some these differences. See the spread at 2kHz after EQ ? Well at that frequency I'm very confident that what I'm actually hearing is similarly spread. The TLDR is : not confident enough that it’s below threshold of audibility, but quite enough that the degree of inaccuracy is a good deal lower than the residual differences observed. How confident I can be in the relative results from that DIY probe is a long boring ass subject that would probably be best left to a new thread but involves criss-crossing the results from different types of mics and compare the "different difference" between their results in the range where they logically are the most relevant (ie, there are good reasons why my DIY probe and my in-concha mics with open ear canals are agreeing somewhat well in the 2-3kHz region, but the blocked ear canal mics don't). Perhaps I'll share my experience in a more complete post with more headphones involved, maybe a new thread, after sending my own samples to Oratory if he's interested, and re-measure them with his generic profiles and profiles tailored to my own samples, at some point in time in 2022.īut for now that's how these eight headphones measure, relatively speaking, without EQ, with my DIY probe mics (averages of five individual traces, right channel only, normalised across one octave centred at 500hz - remember that the absolute values are inaccurate, and that these results are not valid for you, using the exact same tools and methodology on your own head would yield different results): While I rarely measure headphones with EQ presets as, these days, I’m starting to get a fairly decent idea of where I prefer the FR to land anyway, I recently measured 8 headphones with Oratory’s EQ presets and without, with three different types of mics (DIY probe, in-concha microphones, blocked ear canal entrance microphones), in an attempt to get an idea of what the Harman target exactly sounds like. It’s something that I tried to illustrate, as far as my own experience is concerned, in this post from a few months ago, with on head, in situ measurements : There’s nothing new here, and it’s something that’s fully acknowledged by, for example, Oratory : I think that most of us who have EQed headphones with presets based on ear simulator measurements would have noticed that headphones still sound different, occasionally quite jarringly so. To the point where the question of "follows the Harman curve" or even "what does the Harman target actually sounds like ?" is not a straightforward subject to answer IMO. It's just that what's measured on an ear simulator may not be what happens on your own head below threshold of audibility, that's all.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |