Ideal target curve for virtual speaker/room emulations

[Volutrik]

Hi! For contextualization, Waves NX plugins now use the Harman target as the standard target curve to make headphones "flat". It makes little sense, here's why: The Harman target is supposed to emulate how flat speakers sound in a well treated room. So, by applying this equalization to headphones and then using a software that reproduces the sound of another studio, you're basically not experencing how that studio truly sounds. It's almost like you're doubling the effect, even though it's not really like it. (I've attached a tweet from Dr. Sean Olive to prevent missinterpretation. Yes, that's what the target curve is meant to do).

That's one of the main reasons why I'm slowly starting to believe that room emulations aren't a good idea, no matter how good and advanced the software is. The point they're trying to sell is to give you reliable monitoring through headphones. But, because of this, you're basically going back to the problem that made this emulation idea a thing: Bad frequency response.
Some plugins even state that they're delivering "perfect room acoustics"! Now tell me, how is that perfect if the headphone coloration is affecting the sound? You're better off listening to speakers in a normal room, you can at least apply correction with Sonarworks or REW.
Ok, to not be so rude, using in-ear microphones to EQ your headphones maybe would give you good results.

But, to summarize, if there isn't a better approach to headphone equalization in order to use these room emulation softwares, then this whole idea is just invalidating itself.

What are your opinions about this?

 

[Zensō]

My opinion (others may disagree) is that room emulations are a gimmick that color the output and make mixing decisions more difficult, not less. A better approach for me has been to use either Sonarworks or Oratory1990’s Harman EQ (the targets are similar) combined with a small amount of crossfeed (I use CanOpener). YMMV.

 

[NTK]

The goal of the research by Dr Olive's and colleagues was to find out how to make headphones with the highest listener preference rating (averaged among the listeners). It wasn't intended to emulate any room.

 

[Volutrik]

My opinion (others may disagree) is that room emulations are a gimmick that color the output and make mixing decisions more difficult, not less. A better approach for me has been to use either Sonarworks or Oratory1990’s Harman EQ (the targets are similar) combined with a small amount of crossfeed (I use CanOpener). YMMV.

I mean, it's not totally a gimmick because you can in fact emulate yourself listening to speakers through headphones with the use of your HRTF. There is Impulcifier and Smyth A16 that does this really, really well, for example. But they're just not 100% there because of the headphone's impact in frequency response. If this frequency response aspect is solved, then the emulation is completely nailed. It's more a matter of execution than wether it's valid or not.

My suggestion to solve this problem would be to get the Ear Related Transfer Function from the GRAS or 5128 rig, for example, and then exchange it with your own ERTF. Then measurements would better translate and give more accurate results, because they would output like if they were measured with your own ears.