Headphone EQ Settings & Diagram Generator for oratory1990 Measurements

Hello Kai, thanks for the suggestion.

Would you bind this to the ADI-2's Treble control, or assign it to one of the PEQ bands? I think we still want the Treble control itself sitting at 6.5 kHz and free for the user to adjust by ear, which means we have to put the compensation on a different band.

There is a hard reason against the Treble control anyway: its frequency range on the ADI-2 stops at 3 kHz at the low end, so a compensation meant to start around 2 kHz cannot be dialed in there at all. Band 5 has no such problem; it reaches down to 200 Hz.

But then we need to be careful with handing out bands for special purposes like this. I am already using Band 5 for an optional slow filter compensation at 44.1 kHz. I already had feedback that spending a band on that is wasteful, so it is no longer the default.

That criticism is fair, as the Meze Elite Tungsten needs 8 EQ adjustments below 2 kHz. The software already had to drop the two corrections with the smallest gain; you can see which ones are documented on the graph on the right.

So I suggest using either "slow" or "treble energy" compensation. The config file already has one field per headphone for slow filter compensation. That field could become a single either/or choice:
- "slowcompensation",
- "treblecompensation" or
- "none"

Both are high-shelf filters, so they can share the same slot, and the user never pays more than one band for either. The Treble control then stays untouched at 6.5 kHz and remains available for adjusting treble by taste.

One thing I would like your view on. In your proposal, the shelf frequency is not something we choose in advance.
It falls out of fitting the smoothed error curve, so it will differ from headphone to headphone.
For a headphone whose treble error already starts around 2 kHz, the best fit would put the corner frequency down at roughly 2.5 to 3 kHz.
A shelf sitting that low does not act on treble alone anymore; it reaches right into the presence region around 3 kHz, where the Harman target has its ear gain peak, and which is tonally the most sensitive part of the range.
It is also the region where measurements are least reliable in the first place because of fit and individual ear canal geometry, which is the very reason we stop correcting at 2 kHz.

So my question is: do you consider it essential that the compensation cover the range starting right at 2 kHz, or would it serve the same purpose if it only started above the ear gain peak, say from 4 or 5 kHz upward?

RME's own factory presets appear to lean towards the latter. The HD800S preset carries a high shelf at 8 kHz, Q 2.0, +3 dB. The HD650 and Heddphone presets pull the treble down broadly using the Treble control at 10 kHz, by -5 and -3 dB, respectively. None of the four headphone presets attempts anything of that kind, starting as low as 2 kHz.

BTW, while we were discussing this, I noticed something in my own code: in the Room EQ export, I am writing Band 9, the counterpart of the ADI-2's Treble control, as a Peak filter rather than a shelf. It sits at 6.5 kHz with 0 dB, so it is inert as shipped, but the moment someone turns it up, it would behave differently from the ADI-2 for the same preset. I have since worked out the value the format uses, 1 for Shelving and 0 for Bell, so the fix is a single character.

This will not go into the next release, by the way. I am working through a code review of the tool first and want that finished before adding new features.