Zitat von live3TA im Beitrag #20
Why doesn't Sweelinq use two impulse responses that are horizontally displaced about the width of the human head? That way, there would be individual sound images for the left and right ear that would better match the room's acoustic when heard in person.
Wouldn't this contribute much more to the "realness" of the room's acoustic than Sweelinq's two closer and farther IRs?
I do not know which exact method Sweelinq used to capture its impulse responses, so I would be cautious about speculating too much on that point.
What you describe, two horizontally displaced pickup positions roughly corresponding to the width of the human head, goes in the direction of techniques such as ORTF, which historically tried to approximate human listening by using spaced and angled microphones. That was an important step, but today binaural techniques go much further because they account not only for left-right spacing but also for the full interaural behaviour of human hearing: timing differences, level differences, and the way the head and outer ears shape the signal.
In my own recent work, I used ambisonic microphones and then rendered the recordings binaurally. In my experience, binaural reproduction remains the most realistic approach when the goal is immersive audio that closely resembles human hearing. The limitation is well known: it works properly only over headphones. Once loudspeakers come into play, the listening room itself becomes part of the result, inevitably altering the spatial impression.
That is why even very sophisticated surround systems still do not provide quite the same level of immersion as a good binaural rendering over headphones. Loudspeaker playback always interacts with the room in which the speakers are placed, so the reproduction is never isolated from that acoustic context.
So if the aim is the most realistic reproduction of a recorded acoustic space, then yes: a binaurally captured impulse response, listened to over headphones, is in my view the most direct route. For loudspeaker playback, one always ends up with a compromise of one kind or another.
As for Sweelinq’s adjustable mix between closer and farther reverberation, I suspect this is also partly a response to what Hauptwerk users have become accustomed to over the years: the culture of freely mixing multiple perspectives. One often sees combinations in demo videos such as 100% close, 50% mid, 100% rear, which, from a strict acoustical point of view, is of course highly artificial, rather like putting ketchup on a Margherita in Naples. But tastes differ, and some users clearly enjoy having that freedom.
From a product perspective, I understand why Sweelinq goes in that direction. It has to compete with Hauptwerk, which means not only offering similar possibilities, but also giving users some additional reasons to choose it (for example, usability and the consistency of the same GUI across all instruments).
So my answer would be: your idea is acoustically interesting, but if one really wants to imitate human hearing as closely as possible, the discussion quickly moves beyond ORTF toward proper binaural capture and binaural playback.
