- Joined
- Jan 23, 2020
- Messages
- 4,484
- Likes
- 6,986
I am (good-naturedly, I assure you) accumulating a lot of bones to pick with this converstaion. I think we are getting a little detached from reality.
The Harman research tends to show that the litmus test is DBT of at least three speakers repeated enough times to be statistically reliable to see what the preference for people with good hearing is. I think there is here truly a lot of over-interpretation of graphs and imagined correlation between what a graph shows and what a person hears in their own listening space. Harman trained listeners, even speaker designers, give you garbage biased results on comparative sighted listening. We should take a lesson from that. The real deal is the DBT, no graphs in hand. The graphs are shorthand to roughly predict what would be preferred on a DBT. Or so the story goes.
I also like to leave some nooks and crannies in my brain for something other than the Harman / Toole / Olive school of thought.
I also note that people with good hearing tend to exercise the same preferences on DBT, no matter their training level or background. Harman trained listeners just tend to get to the same conclusions faster and more efficiently during blind testing. The outliers on preference tend to have defective hearing or be older.
This is all in Toole’s book, backed up by research, etc., etc.
I like to have good speakers because they are good speakers.Kind of like a blank canvas on which to project the aural image created by the recording artists and recording engineers. That’s my motivation. Get into the upper 7s and 8s and 1) the preference scores tend to under-rate the performance of speakers on DBT and 2) you’re likely to have a statistical tie as to preferences. The exciting truth is that good speakers can be inexpensive! And the data presented by @amirm is much more than sufficient to achieve the otherwise nearly impossible task of separating the wheat from the chaff. We now know tremendously more about these speakers than what you would know about a tire by reading the specifications on the tire wall (to paraphrase F. Toole).
Life is good! But for any variety of reasons this truth, that good speakers can be inexpensive, is often shunned or greeted with skepticism or even hostility. This is not a good thing from the standpoint of promoting the hobby by expanding the awareness of the consumer population.
And yet, even so, good speakers are really not needed. I can listen to a Sonos One at my bedside at 32-38 dB at 2 in the morning and greatly enjoy it. Why is that? Cause I’m nuts? Or because I really didn’t need Hifi to enjoy the music? Or it can fill up my whole bedroom. HiFi, nope. Enjoyable, absolutely. On the downside, this presents a lot of unchecked opportunity for corporate cynicism, sloppiness and deceit when it comes to motivations for manufacturers to provide consumers a product not commensurate in performance with dollars (or your currency of choice) spent.
Also the Harman school of thought on loudspeakers is ultimately founded on subjective preferences, objectively ascertained. Objectivist versus subjectivist is a false dichotomy in the context of speakers. It’s just really fortunate that such an admirable body of knowledge has been developed as to what people with good hearing will prefer, and how to make good (but not perfect by any means) predictions as to what people with good hearing will prefer based on measurements. This really sprang from Toole’s intuition at a young age that people would prefer a neutral speaker, and then was refined based on empirical data. If someone has hearing that is problematic to one degree or another, no problem, it’s just that their subjective preferences won’t translate reliably over to others. But their subjective impressions and preferences are certainly valid for them and their enjoyment of music is no less valid and may even be more informed (in the case of musicians with occupational hearing damage, etc.).
End of mild-mannered rant.![]()
I agree that large scale blind test results should probably trump measurements, at least until we can get much better at interpreting those measurements; I don't think the current Preference Rating is there yet. I'm a huge advocate of blind tests, in many aspects of life. With no measurements available, I bought my current mains almost entirely based on how they did in blind tests.
Problem with blind tests is that they're hard(and expensive) to do, and it's unlikely you'll find test data between the speakers you're looking at. If we had a several hundred sample size blind level matched (with subs) test between these 2 Elacs, and that test strongly favored the Adante, then I would trust the results of that test over my(or the formula's) interpretation of these measurements.
However, with only measurements and a few subjective - sighted - impressions to go by, I lean more towards going by the measurements. Imo, measurements are more generalizable to others than solitary sighted impressions. Solitary sighted impressions are heavily influenced by both expectation bias and individual preference. Perhaps the person is one of the outliers Floyd talks about?