Geddes on Waveguides

[AllenB]

Earl, I recall from one of your white papers you simulated diffraction/reflections by mixing a second source to be fed via headphones when demonstrating level dependent perception. What about these artifacts when they've been included on a recorded song? Will these on their own be enough to cause this recording to sound worse at higher levels?

[gedlee]

I can only guess, but yes, I would say that diffraction that is not part of the instrument would be a degradation. But keep in mind that things that are part of the original recording are all good if the producer wants them there. Its adding or subtracting things on playback that were not in the original that are bad. If the producer was happy with the sound the so be it, if not then they should fix it.

Clearly producers don't always know when something is "good" or how to fix it. That is more than evident from some of the junk in the marketplace. I have seen seriously clipped - audibly so - songs on first rate musicians CDs. How does this happen? Things do see to be slightly out of control IMO.

[AllenB]

I can't say I have much faith in the consistency of recordings. Mostly I dislike what gets done intentionally more than what happens by 'accident'.

The reason I asked is because I'm trying to judge a speaker by listening in mono for changes that happen with level, but when some do sound worse I'm not sure whether it is a clear sign of speaker problems or whether the recording is contributing to it.

[gedlee]

That is an interesting test, I've thought about doing that before. Let me know how it comes out. One test that we did once was to record the speakers in an anechoic chamber at different levels and then play them back at a constant level. There was not as much difference as one would suspect. The thing is that the diffraction detection depends on the playback level NOT the recording level. That makes it very difficult to test.

[keyser]

Originally Posted by gedlee

That is an interesting test, I've thought about doing that before. Let me know how it comes out. One test that we did once was to record the speakers in an anechoic chamber at different levels and then play them back at a constant level. There was not as much difference as one would suspect. The thing is that the diffraction detection depends on the playback level NOT the recording level. That makes it very difficult to test.

Jumping theads here (Orion beaten by Behringer): wouldn't the same methodology work well to detect audible distortion and/or dynamic compression in loudspeakers?

[gedlee]

Yes, I believe that it would. But it requires some fairly sophisticated techniques that are hard to implement.

[mefistofelez]

Dr. Geddes,

in your paper [1], you are discussing radiation impedance for the waveguide, which shows rapid loss of an efficiency below certain value of ka.

Applying Eq. (22) to a waveguide with a=0.5 inch, and theta=45 deg, which you incidentally use in your product, yields cutoff frequency of about 2.4 kHz. Yet, if I remember correctly, you are using the waveguide down to about 900 Hz.

The paper does mention that "in practice it has been found that they [the efficiency restriction] can be relaxed a great deal", but does not go to any details.

Could you comment on the theoretical vs. practical difference? What would be the disadvantage of going to a larger a as Eq. (22) would imply?

Kindest regards,

M

[1] Earl R Geddes, Acoustic Waveguide Theroy, J. Audio Eng. Soc., Vol. 37, No. 7/8, 1989, July/August

[gedlee]

I think the first thing that you should remember is that a direct radiating loudspeaker operates almost its entire bandwidth on the upward sloping portion of the radiation impedance, so its not really a "cutoff" effect. That said, I have seen no real cutoff effect from the waveguide, it is dominated by the cutoff from the driver itself (below resonance). Perhaps the efficiency is down below 2 kHz, but the drivers resonance is way up in efficiency in that region.

[bohdan1232000]

Hi Earl,

My apologies for not reading 5738 posts first, so perhaps my question was answered already before.

Are you in the position to disclose phase response, step response and impulse response of the Summa, Abbey, Nathan and Harper loudspeaker systems?.


Best Regards,
Bohdan

[gedlee]

Posting impulse responses is no problem and step responses are just the integral of the impulse response so they become obvious from the impulse response. I don't plot out phase anywhere, although all of my measurements and plotting is done complex so the phase is there, I just don't have any use for looking at the phase as a stand alone parameter.

What do you want to do with this data? That will help me to decide what to get you.

I could post the HolmImpulse raw data files which will contain the phase if desired, but they can only be viewed from within Holm.