How do you wish to test *accuracy*? Measurements at the MLP compared to the source?
Research already exists that tells us listeners tend to prefer neutral ('accurate') loudspeakers. At the MLP research tells us they tend to prefer an acoustic that offers gradually tipped down treble, tipped up bass. Then on top of that, there's the preferences data about reflections. On top of that we have data on professional vs. 'recreational' listening. How would you design an 'accuracy' preference test distinct from all that?
Or have you done this work already? Please point us to it.
You test the audibility and threshold of comb filtering and reflections and how they are perceived. Some researchers does this with head phones. Others with speakers in room with either speech, music material or pink noise. We have numerous studies of this over years. Audibility, threshold and how they perceived will depend on many factors. For example what level, the time delay vs the direct signal, direction of the reflections and possible masking effects of other high gain reflections. If you a lot of early specular reflections, the threshold of invdidual ones goes way down. Eveything is somewhat a blur and you can't detect individuals anymore.
Studies give it different description. Olive and Toole called the result of lateral reflections "image spreadening" or "image shift". If the reflections arrive sufficiently late, they also add to spaciousness. Another study descriped it as following:
"According to most listeners, the source width increased and the source focus decreased in the presence of early reflections. Some listeners (20%) experienced changes in the environment width, timbral differences, and changes in loudness."
Bottom line is the "imaging" suffers but a sense of space can increase. And worse imaging is the same as degradation of clarity, intelligibility, localization and tonality depending on several factors. And this is accuracy as you are not able to hear the recorded signal as well anymore. You basically can't have both great imaging/accuracy and spaciousness in regards to lateral contributions or not. You have to pick. The only way to get some of both is by introducing diffusion. But certainly not Skylines placed on side walls which don't diffuse much and waste energy in wrong directions.
While it's true that headphones have their own issues, generally the imaging is of good quality and much better than what's experienced in untreated rooms. I'm sure many can attest to that.
When it comes to preferences, many variables will come into play. But generally very early reflections are not preferred but late lateral ones can be under certain conditions with certain music genres. Often classical is the genre that benefits from later laterel reflections. None of these studies have done much research in different types of treatment of the room though, and they generally use quite poor acoustic treatment. But much research went into accurate critical listening rooms decades ago.
It's also worth noticing that in a home, people seldom have perfectly reflective walls. There will often be furnitures, windows, drapes, pictures, etc. which give a much complex impedance and thus altering the reflections differently from room to room.
That's why and combined with listening to various music material over time may a give a different experience as a short blind test with one music genre, I recommend people to experience for themselves. While I'm strong believer in blind tests when it comes to distuinguishing differences, I don't necessarily believe they give us answer to our preferences over time in our environments that highly differ. A comment from a study in regards to another possible issue in homes:
If, however, the reflection arrangement had been unsymmetrical, then an image shift would probably have been perceived toward the strongest reflection plane or hemisphere
FIY: I develop speakers with constant directivity. In regards to commercial interests it would far better for me to advocate that customers only need such speakers and no treatment. But I know this is simpy not the case. Your mileage will vary.
Below is a threshold for different instruments from a research. As you can see, there's no argument here that it's audible or not. The question is whether what you prefer in your own listening room.
It is clear from the results that a very high proportion of subjects detected the difference in the sound field of the room with reflections on and off. It was interesting to note that the highest percentage of correct results was for sounds with a pronounced transient nature and with a mainly mid- and high-frequency spectrum; in this case for cornet and trumpet
I encourage people to rear up on different researchers. Some are available as papers that can be bought and others can be "hidden" in books and in certain organizations. But do you own testing as well if you can. It doesn't have to be a blind test when we already know it's audible. There are so many factors that come into play with "preferences", that being conclusive in regards to what's scientific in the matter is a moot IMO.