Effect of Loudspeaker Directivity Compared with In-room Measurements

[dc655321]

Just a comment: it’s much easier to see differences if the data is presented as a difference. YMMV.

 

[Kvalsvoll]

Just a comment: it’s much easier to see differences if the data is presented as a difference. YMMV.

And that is possible to extract from the measurements.

Coming from a background in simulations and dynamic systems, I tend to look for patterns in the data. Trying to simplify the data too much, even trying to end up with a single number, can be misleading, and will obscure the behavior of the system. Sound is more about patterns, than single numbers.

 

[Duke]

Coming from a background in simulations and dynamic systems, I tend to look for patterns in the data. Trying to simplify the data too much, even trying to end up with a single number, can be misleading, and will obscure the behavior of the system. Sound is more about patterns, than single numbers.

I lack your background, but agree with your priority.

One of the areas where I think John Atkinson could improve on Stereophile's loudspeaker measurements would be if he did NOT normalize the off-axis curves. Doing so DOES throw a spotlight on the difference between the on-axis and off-axis response, BUT also makes it more difficult to see what the off-axis response actually IS because in order to do so you have to mentally modify the off-axis curves by the inverse of the on-axis curve.

I would much rather see all of the data and have to squint a bit to discern the difference, than see only the difference and not know what the "big picture" is.

Imo your text pointing out what to look for and explaining the implications is extremely helpful and fills in the gaps.

 

[dc655321]

And that is possible to extract from the measurements.

Coming from a background in simulations and dynamic systems, I tend to look for patterns in the data. Trying to simplify the data too much, even trying to end up with a single number, can be misleading, and will obscure the behavior of the system. Sound is more about patterns, than single numbers.

My suggestion is not mutually exclusive to your initial effort.

And nowhere did I suggest dumbing things down. Not sure where you pulled that from, or why…

 

[Kvalsvoll]

One of the areas where I think John Atkinson could improve on Stereophile's loudspeaker measurements would be if he did NOT normalize the off-axis curves. Doing so DOES throw a spotlight on the difference between the on-axis and off-axis response, BUT also makes it more difficult to see what the off-axis response actually IS because in order to do so you have to mentally modify the off-axis curves by the inverse of the on-axis curve.

Funny to learn you also noticed this, but then there is also a huge problem, once the data now are already presented in the world's largest and most complete database for loudspeaker measurements - if the presentation changes, it will be harder to compare against all those older speakers.

In a publication like Stereophile, there will always be a limited space available for such technical details, so they have to choose some measurements, that they find most useful for their audience. In hindsight, it is easy to be wiser, but here, I can understand why they ended up with normalized graphs, and why they choose not to change this.

 

[Kvalsvoll]

My suggestion is not mutually exclusive to your initial effort.

And nowhere did I suggest dumbing things down. Not sure where you pulled that from, or why…

I fully agree that this can be presented in a way that shows the differences better, but it will require more explanations, the data will be even more unfamiliar.

I show this using tools that many of the more tech-oriented audiophiles are already familiar with. Most people who do measurements use REW, everyone has the tool and can experiment with own measurements and ways to visualize things.

I observe that many tend obsess about simple ways to look at the data, typically the frequency response, without considering what happens in the time domain. I see people measure their rooms, and then show a frequency response chart, implied that this is representative of the sound that is heard in this room, which it clearly is not. But viewing the data in different ways, can indeed reveal a lot about what is going on, and relates much better to what we actually hear.

Dumbing it down would be to present some sort of number, and then say this represents the sound of a speaker.