So we seem to keep sidestepping this question/problem just a bit. It is mentioned in various degrees of obliqueness in various threads both past and present, sometimes saying it is the first problem to be addressed before any real work to treatment can begin, and sometimes saying it is a minor concern in that the model is "whatever you prefer..."
Well, I don't recall seeing a discussion that really addresses this here on AVS. Some on other forums like gearslutz gets close, but the context is a bit different.
The question/problem: If we have the measuring tools and targeted treatments available to implement a chosen small room acoustic response model in a home listening environment, out of the many "vetted" models that have been developed over the years, which should we choose and why?
I know some assumptions/constraints have to be applied to even approach this question, so allow me to insert my own preferences that I imagine will suffice for many other forum members. Specifically, I'm talking about a multichannel playback environment meant for movies and multichannel sound playback in a modest size home room (say, anything from 15x15 to 30x30, if this must be constrained). Multichannel as in both discrete and extracted via processing such as Logic 7/Q, ProLogic, Neo, Trifield, etc. Multichannel as in minimum 5 or 7 channels, but extended to include 9, 11, and beyond.
From a practical standpoint, if I have a room stripped, isolated, and ready to begin installation, measurement, and targeted treatment... what am I treating to achieve? It is of no use to respond "whatever you prefer" due to the practical issues this raises. How is one to configure, evaluate, and compare various room models given but one room? If the metric is personal preference, I don't think this can be reasonably established without real time access to two identically sized and equipped rooms which follow differing response models. Even then it would be difficult for a single listener to concretely establish his personal preference due to memory persistency issues, etc. More easily accomplished would be sufficiently powered trials to establish "general population" preferences, but then whether that sampling of the population represents the specific user is always in question.
Pieces of this have been addressed by Toole and others. i.e., do listeners prefer on average reflective vs absorptive first specular reflection treatments, etc. But even he touches on some of the issues regarding trained music professionals such as mixing/mastering engineers having different tolerances and differing requirements, listeners preferring a more utopic experience vs more resolving of whatever detail exists, etc. And he comments that as more channels are added, isolated high gain specular reflections are likely (in his opinion?) to become less and less important, contrary to much forum discussion I've seen over the years that assumed as more channels are added the room should become more and more absorbent as it becomes less important to maintain acoustic energy secondary to processing handling ambience, etc. Is that even true? Does the processing from Harmon, Dolby et al available accomplish the same goals with respect to ISD-termination, etc. as the various room models? Discrete multichannel? And the real question... is there any real data on preference for the various models?
Historically, most of the literature and forum discussions for LEDE, RFZ, ambiechoic, AE et. al are addressing recording/mixing environments, often for two channel playback. There may be differences in goals, listener experience/preference, room size, number of playback channels, number of listeners, etc. Can we extrapolate any data on listener preferences from that perspective to a home environment meant for multiple listeners and largely multichannel reproduction?
Again, my problem is that from the perspective of a single listener, how do I choose a room model to use as a goal? If it can't for practical reasons be based on my own personal preferences due to inability to fairly evaluate various room models in the actual target room, and if there is no good data on general listener preference in home multichannel environments, where do we even start the discussion?
If it is on technical grounds comparing the goals of the various models, well established research on how users perceive certain isolated effects, how those effects are approached by the various models, how these models are likely to perform in home multichannel environments, with extrapolation of bits and pieces to listener preference with data where available, ala Toole's work on lateral reflections, so be it. Where do we begin? We, you, I tell people to measure and surgically treat to achieve the desired target response (following a chosen model). How do we help them, and me, choose a model?
Well, I don't recall seeing a discussion that really addresses this here on AVS. Some on other forums like gearslutz gets close, but the context is a bit different.
The question/problem: If we have the measuring tools and targeted treatments available to implement a chosen small room acoustic response model in a home listening environment, out of the many "vetted" models that have been developed over the years, which should we choose and why?
I know some assumptions/constraints have to be applied to even approach this question, so allow me to insert my own preferences that I imagine will suffice for many other forum members. Specifically, I'm talking about a multichannel playback environment meant for movies and multichannel sound playback in a modest size home room (say, anything from 15x15 to 30x30, if this must be constrained). Multichannel as in both discrete and extracted via processing such as Logic 7/Q, ProLogic, Neo, Trifield, etc. Multichannel as in minimum 5 or 7 channels, but extended to include 9, 11, and beyond.
From a practical standpoint, if I have a room stripped, isolated, and ready to begin installation, measurement, and targeted treatment... what am I treating to achieve? It is of no use to respond "whatever you prefer" due to the practical issues this raises. How is one to configure, evaluate, and compare various room models given but one room? If the metric is personal preference, I don't think this can be reasonably established without real time access to two identically sized and equipped rooms which follow differing response models. Even then it would be difficult for a single listener to concretely establish his personal preference due to memory persistency issues, etc. More easily accomplished would be sufficiently powered trials to establish "general population" preferences, but then whether that sampling of the population represents the specific user is always in question.
Pieces of this have been addressed by Toole and others. i.e., do listeners prefer on average reflective vs absorptive first specular reflection treatments, etc. But even he touches on some of the issues regarding trained music professionals such as mixing/mastering engineers having different tolerances and differing requirements, listeners preferring a more utopic experience vs more resolving of whatever detail exists, etc. And he comments that as more channels are added, isolated high gain specular reflections are likely (in his opinion?) to become less and less important, contrary to much forum discussion I've seen over the years that assumed as more channels are added the room should become more and more absorbent as it becomes less important to maintain acoustic energy secondary to processing handling ambience, etc. Is that even true? Does the processing from Harmon, Dolby et al available accomplish the same goals with respect to ISD-termination, etc. as the various room models? Discrete multichannel? And the real question... is there any real data on preference for the various models?
Historically, most of the literature and forum discussions for LEDE, RFZ, ambiechoic, AE et. al are addressing recording/mixing environments, often for two channel playback. There may be differences in goals, listener experience/preference, room size, number of playback channels, number of listeners, etc. Can we extrapolate any data on listener preferences from that perspective to a home environment meant for multiple listeners and largely multichannel reproduction?
Again, my problem is that from the perspective of a single listener, how do I choose a room model to use as a goal? If it can't for practical reasons be based on my own personal preferences due to inability to fairly evaluate various room models in the actual target room, and if there is no good data on general listener preference in home multichannel environments, where do we even start the discussion?
If it is on technical grounds comparing the goals of the various models, well established research on how users perceive certain isolated effects, how those effects are approached by the various models, how these models are likely to perform in home multichannel environments, with extrapolation of bits and pieces to listener preference with data where available, ala Toole's work on lateral reflections, so be it. Where do we begin? We, you, I tell people to measure and surgically treat to achieve the desired target response (following a chosen model). How do we help them, and me, choose a model?