Measured monopole and dipole room responses

[mondogenerator]

Originally Posted by Elias

That's just the key! One needs to get rid of the traditional idea of having looking the room in linear time. For a human linear time have little perceived importance at different frequencies.

For example, take a 10ms reflection: At 20Hz it would be very hard to detect by ear as one period of 20Hz is already 50ms. But at 20kHz (one period is only 50us) 10ms reflection should be easily detectable by the ear.

Thus it is better, in my opinion, to scale the time axis to the frequency in question one is making the observation.

Keep the end user in mind!

- Elias

What IS the frequency in question?

I can follow the cycle scale, but it looks to me that any bass periods would not be shown.
Since this is a 'bass' measurement, a scale of 20 cycles at 50hz is far too course. Even in my small room the longest side is shorter than wavelength of 50hz, so what exactly is this really showing? Is it showing the point where the full cycle can form, i.e. Over 50hz, or what? Im not sure this shows quite what you think, or at least not a pure result.

[john k...]

For what its worth, years ago I designed a woofer system called the CRAW (Controlled Radiation Active Woofer). The woofer could be operated in 3 modes: Monopole, Dipole and Cardioid. Equalization was applied to each mode to ensure that the free space, axial response was identical. I then measured the in-room response of each mode at various positions around the room, and with the woofer in various positions. Here is one screen shot of the measurements at one position (about 3 M from the source).

Dipole is red; Monopole is yellow; Cardioid is white. As can bee seen, below about 150 Hz the cardioid has the smoothes response. Below the room fundamental (around 28 Hz) the dipole falls off rapidly due to the inability to pressurize the room.

[markus76]

Thanks John.

Did you also look at source locations very close to the listener?

[john k...]

Originally Posted by markus76

Thanks John.

Did you also look at source locations very close to the listener?

Yes I did. As the source gets closer to the listener the response approached the near field response of the source, which is not unexpected. In such a case it makes sense to go with a monopole because gradient woofers require distance to form the radiation pattern. For example, if you place a dipole so that the front source is close to the listener what you hear and measure is dominated by the near field response of the front source. So you hear boosted bass. Of course if you position the dipole so that listener is 1/2 way between the sources you are still in the null of the response.

I did some room simulation (for a monopole woofer) of this as well:

Very close to the listen;




Woofer positioned directly behind the listening position:



Note that in both cases the direct sound is dominate over the modal response and the response is also minimum phase, given that the source is MP.

I have suggested placing subs close to the listener with level and delay adjusted for some time. At these low frequencies the delay doesn't seem too important. A steep slope LP filter must be used or else the midrange "leakage" gives the position of the woofer away. My listening experience and measurements were consistent with the sims.

[a_tewinkel]

Originally Posted by john k...

A steep slope LP filter must be used or else the midrange "leakage" gives the position of the woofer away.

What would a steep slope be, 24 dB/octave? And what's the maximum cut-off frequency from your experience?

[puppet]

Nice to see what I've stumbled on by trail and error born out scientifically by john k.
Even though a steep slope (4th order) will allow a higher XO frequency, I keep my transition down around 116hz. Almost all midrange content is masked at this range .. even if your ear is right in the driver.

[markus76]

Originally Posted by john k...

Note that in both cases the direct sound is dominate over the modal response and the response is also minimum phase, given that the source is MP.

That's what I've found too but it's not necessarily minimum phase only. The problem with such a configuration is that the level quickly drops with distance. Here's a monopole directly behind the listening position (red) and 4 other points about 2' around it:



There are ways to improve the distance-dependency and the decay. I hope to have something to show soon.

Originally Posted by john k...

I have suggested placing subs close to the listener with level and delay adjusted for some time. At these low frequencies the delay doesn't seem too important. A steep slope LP filter must be used or else the midrange "leakage" gives the position of the woofer away. My listening experience and measurements were consistent with the sims.

Very easy to achieve with any AVR. I' wondering why not more people use it. Any other method of reducing modal decay at low frequencies is more complex and expensive.
I believe that reducing modal decay is even more important than flat magnitude response although the latter is a good indicator of the former.

[john k...]

Originally Posted by markus76

That's what I've found too but it's not necessarily minimum phase only.

Agreed. It depends on how far the source is from the listening position. As the distance gets greater and the modal response becomes more dominate, the MP characteristic is lost. That loss of MP behavior typically starts at higher frequencies where the modal density is greater.

I also agree that with 4th order a cut off of around 80 to 100 Hz seems optimal.

[weltersys]

Originally Posted by weltersys:
Subs equalized outdoors for the same target response through the passband of 40-120 Hz using a DBX Driverack PA, Smaart Magnitude response, one a small bass reflex, the other a bandpass box.

Originally Posted by markus76

And then you put them in a room and listened A/B? Where did you place them in the room and where was the listening position (was it always exactly the same)? Did you also take in-room measurements at the listening position?

Markus,

As you wrote in post #115,

“Two systems with very different impulse responses can have exactly the same magnitude response. “

I spent quite some time getting the “Resopump” bandpass sub working well as far as magnitude response. It looked quite promising, but I was so disappointed in the sound quality when compared to any sealed or bass reflex cabinets that the cabinet was recycled and the raw speaker sold. It never made it in to any listening system.

The raw speaker in the Resopump was an Eminence HL-10C, I compared it directly to a ported Eminence Lab 12, and also to a ported Eminence 4015LF, and an old Acoustic Research sealed 10”.

Although the 10” bandpass sub had more sensitivity than all but the 4015LF, the slowly decaying impulse response made music sound like mush, not something I’d recommend to any listener unless they only listened to music with droning bass.

Art Welter

[markus76]

Art

I lost track what you did compare A/B and what not. You would need to describe the test more thoroughly. Let's revisit your original post:

In an A/B listening test with two systems equalized for identical response, a sub with poor transient response, as described above, will subjectively sound “tubby” or “slow”. Percussive notes tend to blend together, bass lines become less recognizable.

There really is no transient response at low frequencies in acoustically small rooms. What one hears is the room - the modal decay and the magnitude response aberrations. This makes bass sound “tubby” or “slow”, it's not the source. Of course there's always the chance that a sub didn't work as intended but this should have been obvious from free field measurements.