.

[USER]

.

 

[Sparky]

Good stuff....

 

[Alexanderc]

This is great! Nigh unusable on my phone though. I’ll have to do a thorough exploration on my laptop later.

 

[Sparky]

This is great! Nigh unusable on my phone though. I’ll have to do a thorough exploration on my laptop later.

I just click on "page settings" then "desktop version" and it works fine.

 

[bigredrobotninja]

What am I looking for?

EDIT: oh I see it...

 

[Doodski]

Need a sticky on this web page. Great work.

 

[waynel]

Just in case: its not the (glorious) site itself (MZKM's valuable work, I believe), but the TBA.

I saw the TBA - poor performance but it's a $650 speaker, would be much more interesting if it were an expensive speaker. Any guesses?

 

[Wes]

Thx for this - is there a way to select out bookshelf models?

 

[waynel]

My working guess is:

View attachment 83593

Yeah, that’s a better guess than mine

 

[richard12511]

Just in case: it's not the (glorious) site itself (MZKM's valuable work, I believe), but the TBA. Things are about to get rowdy here at ASR.

What's the TBA?

 

[RayDunzl]

The TBA on Preference Score made me break out laughing...

 

[sweetchaos]

You've found the easter egg.

 

[bunkbail]

It surely is the LRS and it is totally expected anyways. What I want to know is if it can score much higher if the response is corrected full-range by Dirac or any room correction softwares.

 

[richard12511]

It surely is the LRS and it is totally expected anyways. What I want to know is if it can score much higher if the response is corrected full-range by Dirac or any room correction softwares.

Yeah, I figured it would score low. I'm more interested in Amir's subjective impression tbh. Does it sound as bad as it measures? If so, that goes a long way to validating such measurements. If it measures bad and sounds good, that's also interesting, as it means we have work to do in terms of measurement interpretations.

 

[Sancus]

preference score != measurement quality

 

[Snoochers]

preference score != measurement quality

What do you mean? Preference score is almost exactly measurement quality.

 

[Sancus]

What do you mean? Preference score is almost exactly measurement quality.

No it's not, there are a lot of things not included like distortion. The way it includes directivity width is questionable and controversial as to how definitive it is. The model was also developed on a limited set of speakers mostly of traditional design, there's not too much evidence that it would even be valid when applied to weird things like Magnepans unless they were included in the original data set, which so far as I know, they weren't.

There are also speakers with good preference scores that appear to have quite poor measurements like the Harbeths. There are various explanations for this, but it's also controversial.

Preference score is an INTERPRETATION of measurements. It is one that is supported by pretty good math and a decent data set, but to claim that it's equivalent to measurement quality is simply not correct. Many measurements of speakers on this site have been interpreted differently than the preference score by various knowledgeable people including Amir.

 

[Snoochers]

No it's not, there are a lot of things not included like distortion. The way it includes directivity width is questionable and controversial as to how definitive it is. The model was also developed on a limited set of speakers mostly of traditional design, there's not too much evidence that it would even be valid when applied to weird things like Magnepans unless they were included in the original data set, which so far as I know, they weren't.

There are also speakers with good preference scores that appear to have quite poor measurements like the Harbeths. There are various explanations for this, but it's also controversial.

Preference score is an INTERPRETATION of measurements. It is one that is supported by pretty good math and a decent data set, but to claim that it's equivalent to measurement quality is simply not correct. Many measurements of speakers on this site have been interpreted differently than the preference score by various knowledgeable people including Amir.

It isn't perfect, but it correlates very good with speaker quality. That is like saying a car that has a big engine, low weight, good breaks, and great aerodynamics isn't "fast", it is good lap times that are "fast". Sure, that is true, but damn the engine, weight, breaks, and aerodynamics are correlated very very highly with lap times!

 

[Sancus]

It isn't perfect, but it correlates very good with speaker quality. That is like saying a car that has a big engine, low weight, good breaks, and great aerodynamics isn't "fast", it is good lap times that are "fast". Sure, that is true, but damn the engine, weight, breaks, and aerodynamics are correlated very very highly with lap times!

The preference score can only work for speaker types which were modeled in the original study, this is literally a limitation mentioned by Olive in the study. A generic analogy doesn't change that, sorry.

 

[Snoochers]

The preference score can only work for speaker types which were modeled in the original study, this is literally a limitation mentioned by Olive in the study. A generic analogy doesn't change that, sorry.

What proportion of speakers that are manufactured and purchased does the preference score apply reasonably well to?