BTW why is Sean so fascinated about predicted score? I definitely don't like it. If any I prefer binary scoring system. Sound like shit or sound acceptable.
Likely because of the
rather robust correlation with listener preference, if I were to hazard a guess. Sean's worked hard on those models, not surprising he'd like to see them put to good use!
As promised I attempted to read about Chris Struck's approach using that link you gave (
https://www.aes.org/e-lib/browse.cfm?elib=17041 ), but that was just to an Abstract of the paper, not viewable without purchase.
However I did Google around the subject, and came across a mostly unrelated set of slides from him that are nonetheless interesting & informative so I thought I'd link them, I've skimmed them at this point rather than interpreted all of it.....don't look at this next link in detail (or at all), it's just a by-the-by point of interest mostly unrelated but sufficiently interesting to include: http://www.cjs-labs.com/sitebuildercontent/sitebuilderfiles/HeadphoneMeasurements-CJSLabs.pdf
But anyway, going back to the Chris Struck approach that you linked, from the small bits I've read is this essentially the same as the Harman Headphone Approach for calculating a Headphone Frequency Response Target except it's theoretically calculated rather than measured? And the only measurement it uses is the anechoic data of the speaker, all with a view to creating a Frequency Response Target for headphones that is to mimic any specific model (eg JBL 305p Mkii for example) of speaker that has had it's frequency response anechoically measured.....and then you add in theoretical room variables of your choice (to describe the room you wish to theoretically listen in) to combine with the supplied Diffuse & Free Field HRTF files from the HATS you are using to measure headphones......and that's how the Frequency Response Target is created? I'm making a lot of assumptions here because there's not a lot to read on this subject and I've not spent hours on it. I'm assuming we could also add in things like Low Shelf Bass Boosts on top of that to account for the Harman preference research. Do we have everything at our disposal to try the Chris Struck approach with the B&K 5128? We'd need access to the formulas though I imagine, in order to combine all of this to create the Target Frequency Response?
EDIT: as an aside, isn't this related to the Smyth Realiser system, albeit Chris Struck approach is using more theoretical measurements (dummy head HRTF & theoretical room & 'theoretical' anechoically measured speakers) rather than actual measurements?
I got to see Chris give that slideshow presentation once, it was really good! Although I'm not as sold as he is on swept tone IMD...
FWIW, I'd strongly recommend an associate AES membership for any audio science inclined folks - it's only about a hundred bucks a year, and the back catalog is worth every penny.
You've got the thrust of it - Chris's model allows you to get an approximation of the combined direct + indirect sound power at the eardrum of a head, in a room, for a speaker, given that you have the reverberation time data of the room, the directivity of the speaker, and a set of characterization HRTFs for the head. This would be essentially equivalent to a higher-resolution version of what Olive and Welti did in
2013, where they measured an ear in the Harman room with their Revel speaker pair - a fair amount of smoothing was required for a smooth response due to the in-room condition there, but with the anechoic data, we can be much more high-res
This, I'll note, isn't approximate to the Harman target itself, but rather the "flat in-room baseline" which the Harman target was created from. From there, a large body of listeners were allowed to adjust two shelf filters (12dB/oct at 105~ and 2500~hz, I believe) to preference for both the speakers and headphones, yielding these adjustments in the 2013 study
Subsequently small adjustments were made in the 2015 and 2017 papers, but in premise I would argue that applying the adjustment shelves to a Struck-derived "in-room HRTF" would be appropriate, yes.
Re: the Realizer, I'd kind of argue it's almost the opposite, although the practice of measurement is similar - the Realizer does away entirely with target curve theory and simple says "for a given pair of ears, measured at the canal entrance, what transfer function is required to make circumstance A (headphone listening with a given pair of headphones) match circumstance B (speaker listening with a given speaker set and room)?" - it rather neatly "cuts the knot" of what something "should" sound like.
That's why I mentioned that anything above 5kHz is rig dependent and one should be careful relying on that data to provide accurate EQ.
Below 5kHz there should be not much difference unless the pinnae differ too much. As you say the coupling is direct, yet the pinna does also have an amplitude changing effect above 1kHz so that would indicate that above 1kHz the coupling already isn't direct anymore.
Do you know at what frequency the coupling starts to become less 'direct' and what would be the reason for this. I mean when coupling is direct why would a pinna affect the incoming sound ?
As with a room's transition frequencies, it is variable by the case - some texts I have seen estimate as high as 1-2khz for the transition, but these I think were from an older age of thinner earpads
However, consider where one goes from the "room gain" band in a listening room - it isn't just directly to sound propagating as rays, is it?
Not sure what you are getting at. I hear the mono part inside my head and on the better headphones the sounds that were panned right or left (or binaurally recorded) outside of my head. Just not to the front.
With things like headtracking the signal applied to the headphones is altered in amplitude and phase response to mimic the effect the pinna has because sounds from the front of us have different properties than sounds from the sides or as you put it are 'direct coupled'.
What's interesting about this (you will probably know this and is why I ask) is that headtracking thus should work equally well with IEMs.
I hate to ram things in my ear canal so never use IEMS.
The direct coupling of higher freq bypassing the influence the pinna has thus could/should result in better spatial sound when the correct HRTF of an individual would be mimic-ed. As not all HRTF are the same that may well be the reason it works better for some than others.
For instance I have never experienced sound coming from the front or around me using headphones. To me it is always L-R and inbetween regardless of the trickery/binaural recordings I use. Never tried headtracking though. I am fine with L-R, it is a compromise which I learned to accept over time to me
I believe we're having a disconnect of phrasing here - the latter bolded part is precisely what I mean regarding in-head localization. We obviously do hear stereo panning - moreso with headphones for lack of crosstalk, if anything - but "panned left" is not the same as "localized in the room to my left", and the fact that a mono signal sits inside of our heads rather than inside of our room when played on headphones further shows the situation: we just aren't hearing these sounds "in space". It's particularly stark with personalized binaural recordings as a contrast, because headphones
can faithfully recreate out of head localization with the correct sound cues, and it sounds entirely different.
Aside from the rather unnatural bass boost (driven by the desire to boost bass, lets forget the reasons for this for the moment) and the downward sloping FR response which deviates just slightly from other room corrections that is added to the 'corrected' flat response, what makes the target response of Harman very different from 'flat' with other room responses. Is this the slightly different slope only ?
To be clear, when I refer to the Harman target, I'm referring to the body of curves like these
which are inclusive of both the shelf filters and the in-room eardrum response of the ear simulator. When I refer to just the EQ shelves, I usually call them as such or synonyms, and other than having an atypically well statistically validated correlation with preference, I don't think there's anything particularly special about them.
ndeed... because of my technical background I thus 'pull apart' the correction for the measurement mic and the room correction. The advantage of looking at it this way is that one can (electrically) experiment with various types of room correction where the end result for the HATS + room correction becomes the total correction which some folks thing is 'the Harman' correction applicable to all HATS.
A lot of confusion is generated by the nontechnical minded folks about the differences between the Harman correction plot (for their HATS) and the Harman target.
I have bones to pick with considering the eardrum response of a HATS equivalent to measurement mic compensation, but the general premise here - that it is interesting to differentiate the preferential/room curve element and the acoustic/HRTF element so that they can be separately experimented with - is something I believe we're in agreement on. Have you had a look at Chris's paper I linked above? It might be of interest to you, as it's basically about deriving a "baseline" HRTF target based on known elements of room and speaker behavior.
A question you are likely to know. Is there a difference between the actual final correction correction (Harman in this case) between the overly smoothed 'target' shown in virtually all publications, as an example the one Robbo99999 is referencing to in the Oratory plot and the actual one used per HATS
For GRAS equipment, to the best of my knowledge, no, the approximately 1/3-1/2oct smoothed in-room measurement of the 43AG-in-a-head is the data that's used, including for things like Sean's experiments in equalizing headphones and the correlation with preference ratings there.
I am unsure of how Sean has gone about making targets for other ears - I would have assumed something as you described, with the HATS placed in the Harman room to derive an in-room HRTF with the flattened-at-the-listening-position-with-an-omni-mic Revel pair, but I don't know this for a fact. Perhaps we should ask him, he's quite responsive.