You guys think a flat in-room response (well, flat other than a couple of bass bumps) is a good response?
If the more or less flat in-room response is the result of a linear anechoic response, neutral early reflections and tonally balanced reverb at the same time... well, yes, that is pretty likely to be a good response! Other way ´round, a flat response alone in a room is not sufficient to ensure excellent tonal balance. So, ´it depends´ is the correct answer.
I did not notice anyone claiming to prove a balanced response solely based on this single in-room curve made under unknown circumstances. What we can see, though, is a more or less balanced behavior of a speaker which frequently receives hostile feedback for its anechoic measurements, which certainly don't meet technical ideals. So it is proof enough that in a room things can be turning out differently, not always to the worse, that's all we can say.
a flat in-room response tends to sound bright, so this is not good. You typically want a downward sloping in room response.
That is not the case. An in-room response tilted downwards all over the frequency bands for which our ears are sensitive, will always sound dull, lacking treble/brillance, heavy on bass/lower midrange, this or that way. We should note that a minor downward slope in the highest treble region, i.e. 7K+, is to be expected and not necessarily contributing to dull tonality (as our brain is used to a certain degree of absorbtion/dissipation for shorter wavelengths in huge rooms and greater distances).
A flat in-room response will not sound bright, if both direct and indirect sound are tonally well-balanced. Cases in which a balanced in-room response is reported to sound overly bright, are in all cases known to me, a result of increasing level of direct sound towards higher frequencies, while direct and indirect sound show a pretty different tonality. Well, yes, that is expected to be bright, but it is also indicative of two massive flaws at the same time.
In-room response doesn't really paint the full picture of a speaker sounds, in fact it is a very unreliable source of information.
I would argue that anechoic curves alone are as unreliable, telling only a fraction of the picture, particularly if the indirect sound in a given listening room is likely to dominate.
IMHO it really makes sense to look at the whole spinorama data, particularly anechoic response, early reflections, early vertical reflections separated, and overall d.i. or sound power. Not that this all would allow a precise and sufficient prediction how it exactly will sound in a given room, but it helps understanding what is going on and predicting if a desired combination of loudspeaker and room might work, or rather not.