if summed stereo does sometimes change the tonal characteristics of the recording, so what?
If such a signal would be used in a mono listening test of speakers in order to judge their tonal balance, it would lead to misjudgments. Speakers which compensate for this phenomenon inherent to stereo recordings this or that way, would be more likely to win this test, while speakers which deliver best tonal balance results in stereo tests would more likely to be perceived as colored or imbalanced.
I would not find such a test to be useful, but seemingly it is happening.
Tonal balance varies so wildly from recording to recording yet we are still able to use a variety of recordings to form an opinion about a speaker.
That is the point. You cannot foresee the variations which different mixing engineers have applied on their microphone signals for millions of tracks. I think it is a good idea to stay within the range of a standard studio setup and follow their standards as much as it is practically possible in a living room. You dramatically increase the chances that a majority of recordings which you might be listening to in future, will sound satisfying and not make you want to jump up and correct them with DSP.
Recording techniques, styles and taste of the mixing engineers do vary, but the number of recordings which are completely off in terms of tonality, is astonishingly low.
If you choose a speaker instead which sounds somehow more balanced with a downmixed stereo recording compensating for the inherent tonal imbalances of this method, you are much more likely to have a higher number of recordings being far off average and sounding annoyingly colored, boring, delivering dull ambience, reducing perceived proximity or any other of the side effects of such reversed compensation.
Reference tracks for evaluation is not the same as listening for pleasure, they are tools and are defined and limited.
It's the same kind of consistency, one uses music who knows well.
Having done such evaluation professionally, I am pretty well aware of that. As mentioned, I have auditioned several 100,000s of tracks if they are meaningful for listening tests, and have compiled playlists and sampler albums containing roughly 2,000 to 3,000 tracks which I know well enough in order to choose the one I would consider to be meaningful in a specific listening environment for the given task.
your hypothesis would predict that the highest rated speakers in Harman's listening tests (mono) should show lower DI at 1-2 kHz ("pumping too much energy...into the room") and higher DI at 2-5+ kHz ("attenuating") ... and I interpret "favors" as meaning that Harman's trained listeners would express preference for such.
I cannot predict what a specific group of listener would express preference for, but, yes, that was pretty accurately the result of almost every listening test I have done myself or taken part in. For most of rooms, a constant directivity in that region (1-5K, if not overly deviating from the band one octave lower) is sufficient. If the indirect sound in the room might become dominant, some people who share the same idea even propagate a higher directivity from 0.5-2K compared to a lower one from 2-5K roughly.
Your statements also remind me of Blauert bands and the so-called BBC or Gundry dip
Blauert´s preferred bands were one starting point of this theory, the other being the equivalent loudness for direct vs. diffuse sound as described by Zwicker/Fastl in one of their standard books on psychoaoucstics.
The main idea is: If a sound event like a reverb pattern shows an emphasis in the 0.8-2k band, our brain tends to localize it as coming from the rear/sides. If the 2-5K bands dominate over the 0.8-2k, we get the perception ´coming from frontal angles´. For a mono listening setup, the former is presumably less annoying as it mildens the annoying directness of a mono real source and spreading the angles at which reverb is subjectively coming it, making the mono speaker ´disappear behind the screen´, as Dr. Toole has put it. The latter is in favor of a stereo setup in an overly lively room as it directs the additional reverb to the frontal listening window, ´hiding it´ behind the direct sound and making it more likely to blend with the uncolored reverb from the recording.
The latter is also the reason why some loudspeaker manufacturers do this intentionally.
hese graphs from Toole's second edition of Sound Reproduction suggest relatively smoothly increasing DI up to a few kHz and then relatively flat or decreasing up to around 8 kHz for the Revel Salon 2 and JBL Array 1400, two of the four "kings of the hill"
The question would be which constant directivity or decreasing d.i. in the 1-5K bands loudspeaker without major flaws they have tested as a competitor, and if this has been done with judging tonality and imaging in stereo as well.
I am not aware of such test, as the number of loudspeakers being specifically advertised and achieving CD over a broad band, only appeared after 2010. Around this time I had some eye-opening listening events and tests, some controlled and some not, which brought me to really being sensitive to loudspeakers with increasing directivity index.
I don't perceive any definite evidence favoring your hypothesis over Toole's "relatively constant, or at least smoothly changing directivity.
As mentioned, this is a relatively young concept, as most of the technology to achieve it without major tradeoffs, did not exist before the era of DSP-controlled active speakers.
I encourage everyone to do a comparison between a true constant directivity speaker and one with ´smoothly increasing directivity´ in the 0.5-8K band. I do not see a hard threshold between the two, and there are some speakers with a slight increase in d.i. which one could EQ to satisfaction. I would say a constant plateau between 0.8-5K is the single most important thing with the neighboring bands not making a step up in d.i. But the moment the index is increasing over several octave-broad bands, particularly if the 3-5K band is already of higher d.i., makes me pretty cautious. I have had too many moments of disappointment, as the dull, rear-heavy reverb is also detereorating other aspects of sound quality in my understanding.
my father was a sound engineer at SR (Swedish public service radio ) and later at SVT ( Public service TV ) and then had his own company .he later became an video editor.
