Sorry but Precedence effect states:
“The precedence effect appears if the subsequent wave fronts arrive between 2 ms and about 50 ms later than the first wave front. This range is signal dependent. For speech the precedence effect disappears for delays above 50 ms, but for music the precedence effect can also appear for delays of some 100 ms.
[8]” wikipedia.
https://en.m.wikipedia.org/wiki/Precedence_effect
i would suggest at least 9 feet or 3meters or 9ms far away if you want to keep a reasonable intelligibility and clarity and do not screw up everything.
If 9 feet of path-length-induced delay were
needed to suppress directional cues from reflections, we'd virtually never have decent imaging (nor apparently clarity and intelligibility) in home audio. Not saying 9 feet is a bad idea, just imo not
necessary.
What I'm looking at is the first onset of the precedence effect. We need enough path length for this to occur. We are looking for when "fusion" - the
perception of hearing a single sound from a single direction - FIRST occurs, and that is given as 1 millisecond:
"fusion occurred when the lag between the two sounds was in the range 1 to 5 ms for clicks"
"humans localize sound sources in the direction of the first arriving sound despite the presence of a single reflection from a different direction. A single auditory event is perceived. A reflection arriving later than 1 ms after the direct sound increases the perceived level and spaciousness (more precisely the perceived width of the sound source)."
I'm not sure where 2 milliseconds comes from, except perhaps that by then the precedence effect is very strong, and can suppress the directional cues of reflections which are LOUDER than the first-arrival sound. If you prefer 2 milliseconds, that corresponds to a path length difference of about 27 inches, which is much easier to achieve than 9 feet. Not that 9 feet isn't preferable, but it may not be practical, and is imo unnecessary for trying out the Hafler circuit. My experience (decades ago) was that if the secondary speakers are closer than the main ones, some sounds are localized at the secondary speakers. Adjusting the path lengths so that the sound from the main speakers arrives first fixed that.
Subsequent readings of AES papers (some by James M. Kates as I recall) led to my current understanding that the ear derives directional cues primarily from the first approximately .68 milliseconds of a sound event, and due to lack of precision this was reported by early researchers as 1 millisecond. This approximately .68 millisecond interval is the time it takes for sound to travel around the head from one ear to the other, and corresponds to a path length of about 9 inches. Up to .68 milliseconds, a repetition of the original signal, or reflection, will be interpreted as a (false) azimuth cue . This is why reflections and diffraction on the face of a loudspeaker can be detrimental to imaging precision, and also why narrow-baffle speakers generally have more precise imaging than wide-baffle speakers (the shorter very-early-reflection paths corresponding to smaller azimuth errors). After .68 milliseconds the precedence effect kicks in and starts to suppresses directional cues from reflections.
(The precedence effect does not suppress loudness and spaciousness cues, therefore spaciousness is enhanced by reflections arriving while it is in effect, as demonstrated by the increase in apparent source width [ASW], or image broadening, which can occur in home audio and in concert halls due to sidewall reflections. Timbre is also affected by reflections which occur while the precedence effect is active, as predicted by loudness cues being accepted.)
Of course I could be wrong.