I want to apologize for adding to some of the confusion around this topic, based on previous posts I've made. An article in the most recent article of VoiceCoil magazine had an interview with Earl Geddes and his use of some of these terms has led me to re-evaluate the semantics of exactly what we're saying.
First, we must use the terms correctly to understand exactly what we are inferring when we say that. This is what has led to the confusion, at least on my part.
So, when we say "sound quality", what do we really mean? According to many, sound quality means accurately reproducing the source. When we look at the term "sound quality", there is no possible way it can mean this. Reason? For starters, how accurately we can reproduce the source is a completely quantitative practice, meaning we can use science and numbers to determine how close to the source the material is. This is COMPLETELY different from a "qualitative" approach, where we wish to describe the behaviour with adjectives. The term "sound quality" or "SQ" is intended to describe a certain "quality" of the speaker (or system), and not any qualitative analysis. In this sense, SQ is completely up to whomever the listener is.
Where's the problem? Well, the problem is with organizations who hold SQ competitions, pretend that they are checking accuracy to the source, all the while not realizing that they are attempting to use qualitative analysis to describe behaviour that is more easily described using quantitative measurements.
Simply put: SQ is personal preference, as it is the "quality" of sound. Being accurate to the source (which many claim SQ is, but in actuality is not) is not about personal preference, and is completely quantitative.
Further thinking:
1. Yes, I still stand pat in saying that the goal of a speaker system is to reproduce, not to impress. As such, quantitative analysis (ie. with reliable measurement) is a better approach to system design, particularly during R&D.
2. Using a large selection of qualitative data (enough to fairly represent the Earth's population), we can develop a model which will show a correlation between the quantitative data (the measurements) and the qualitative data (the perception). GedLee has been working on this, and it really shakes the core of a lot of audio engineering. The use of the THD and IMD metrics is completely inappropriate for describing the perceived behaviour, but passing the same physical data that creates THD and IMD measurements through a model similar to the GedLee one will make prediction of perception accurate for the number of humans who fall within the reasonable curve of that model.
Regardless, the purpose of a speaker system is to reproduce the signal (btw, for whomever said that there are speakers with low distortion, I submit that this is completely false within any reasonable bandwidth, although obviously we must define what "low" is). Those who choose to remain ignorant of science, at any level, will continue to do so at their own peril.
