A lot of us seem to be in agreement with this (I think John K still disagrees, though) . . . but . . . why?
Why, if all the rest of the recording chain is kept "flat", do we need to roll off the highs at the speaker?
You misunderstood "in-situ". I mean in its enclosure, but in a free field of course. So there is no room involved, just the enclosure and its diffraction which already has any "baffle step" included.
Yes, I would say that is absolutely wrong to call that "baffle step". Its much too high in frequency. It is correct to say that you "corrected" a bad crossover.
I design for flat direct field response (on the listening axis) AND flat (or slightly falling) reverberant field response (or what you are calling the "power" response). You have to have both right, not one OR the other.
In part because, with rare exception, modern audio production practices have little regard for accuracy.
Everyone is biased. It's all subjective, this.
Everyone is biased, correct, but only the subjective is subjective, the physics is not.
Yes, I understand that, but how does that translate to "constant directivity is not important below about 500Hz" for a loudspeaker in a room? What are the data to support that?
I only agree with this is the high end is CD. If it is a normal direct radiating tweeter then rolling off the high end will not sound right because it does not have a flat power response. The CD speaker has both a flat power response and axial response so all this extra energy into the room does not sound right, it sounds bright.
Yes , agree,
Because like mitch its the best way to tame shouty coloured horns, the same curve will not work for all types of speakers ..
How do you get both in a box (with baffle step)?
Yes, that is the general understanding . . . that direct radiating tweeters, because of the way they beam will, if equalized "flat on axis", have a falling power response and not need the HF rolloff, while "CD" tweeters will, because of their flat power response, require some rolloff (which, unfortunately, acts both on power response and direct (on axis) response.
That would seem to favor the direct radiator, since it does naturally what our ear seems to want to hear . . .
Last edited:
What I disagree with is the idea that if the recording is monitored on flat speakers then it should sound right when played back on flat speakers, That plot some pages ago, and the text, seemed to imply that the rolled off response was the response of the speaker used to monitor the recording. If that is the case then i would agree that the playback system would need rolled off highs. The B&K pdf was from 1974 though. I don't know how relevant it is today.
Another point I have made is that the characteristics of the listen room will have a major impact on whether or not the highs need to be rolled off. I believe that has more to do with reflected sound and the spectral content of the reverberant field then the direct sound.
Whether a speaker is flat or rolled off, if you place it is a glass lined room it will sound a lot brighter than if placed in a room with carpets, curtains, etc. My point has always been that the room is more of a factor in this that the speaker. Remember that any wall has three parameters, reflectivity, absorptivity and transmitivity. All you have to do is listen in another room and you get a good idea of what frequencies tend to be transmitted directly through a wall. Transmitivity in the simplest form goes sort of like 1/f.
And again, I always allow the listener to make such adjustments to my speakers to suite his environment and source material. I am not trying to dictate what is correct. I design my systems to a reference, flat axial response and try to maintain uniform polar response up to a point.
FWIW, I'll post a 1/3 octave pink noise response of my system in my room later on. Currently I have the NaO II set up since I have been experimenting on how to make it sound more like the Note and Note II.
I don't know . . . thus the question.
Regarding the room, I'd speculate that regardless overall absorption first reflections may favor the highs (look at the ceiling, and ceiling/wall corners). But I'm not suggesting that accounts for all, or even most, of it . . .
The fact that we are not sensitive to sound reflections below about 500 Hz means that there is little need for controlled directivity below that frequency. The directivity simply does not matter. Its fine if its narrower, like a dipole or a cardiod, but its effectiveness at those lower frequencies is not a major sound quality issue as it is at say 2 kHz where our auditory acuteness is extremely high.
Is the rising power response an issue? - well that's a different thing. Below the Schroeder frequency directivity and power response are all moot issues because its all modal anyways. Between say 200 Hz and 500 Hz is the rising power response an issue? I don't really know the answer to that one, but its such a small consideration in the big picture that I don't worry too much about it. I know of no psychoacoustics that could address this issue and no actual test data that is applicable either, other than we do not seem to have hearing that is very acute below 500 Hz. (not when compared to above that frequency).
I have done designs that use dipoles to get a narrower directivity in the 200 - 500 Hz range but they never seemed practical enough to build. As a product they would be a dog because the costs would skyrocket. There are rational practical priorities to everything I am afraid.
The trouble with the direct radiator is that the transition of its directivity is not subtle, it is quite pronounced. With a CD speaker it can be made subtle. The direct radiator in a shallow waveguide seems to be a good compromise as just about everyone who is doing this has found (to wit the Behringer referenced in this thread). It still lacks the dynamics and power handling of a compression driver but does yield some of the directivity improvements.
Yes, we're in general agreement here (our last posts seem to have overlapped). The overall polar response of the speaker can, however, have significant impact on what the room reflects. It's why I equalized the rear tweeter differently from the front (on my ORION), and probably why I so hated the ribbon tweeter in the MGIII (which I more or less successfully "tamed" only by hanging strips of fiberglass insulation over the back of the speaker).
I'm still not quite convinced that's all there is to it, though, especially for orchestral recordings (where common microphone placement favors the highs) . . .
Are we not sensitive to the frequency response variations that go hand in hand with those reflections? Do you have actual data showing this lack of sensitivity? I understand the commercial impracticality, but a one-off for actual testing of the hypothesis is a different kettle of fish.
Sorry for my persistence about this, but I'm more of a data-driven rationalist than most, and want to understand exactly what objective data you're working from.
It's all about marketing semantics. There is no such data you are asking for.
Anyone who has experienced a high directivity speaker in 100-1kHz range in a small room will appreciate the benefits of it over low directivity speaker.
The fact that gedlee is not doing it is he cannot sell it.