Exactly. That's why I programmed a little tool many years ago (probably before SL put out his CD) that creates a set of shaped burst sequences (.WAVs, 16bit, 44.1kHz) in quarter-tone steps from 27.5Hz to 440Hz. Each sequence is based on a burst of several cycles with raised cosine fade-ins/fade-outs of 3 periods (IIRC). Several parameters are varied systematically within each sequence: The total burst lenght, the gap in between (in multiples of cycles) plus the phase realationship between the bursts (so that the next burst starts either in phase with the room hangover or in counter-phase by making the gap half a period longer), and the interchannel phase/polarity. The sequence consist of four blocks, each block starts with long bursts reaching steady state, and long gaps in between. Then burst and gap length is successivly shortened. The next block has antiphase gap lengths (x.5 periods). These two blocks are then repeated, with one channel inverted.
From listening to this broad variety of repeatable and systematic but music-like modulations I learned more about room behaviour than from any measurments or any elaborated mental efforts before. Burnt on CD, with some additional test signals for convenience, it became several people's main tool for quickly judging and adjusting the PA sound at smaller clubs (the original application) and for helping friends to find bass-friendly positions of speakers/subs and listening chair as well as for checking the effectiveness of room treatments.
Markus knows these signals already. I may upload the (DOS) .exe and source-file (for Turbo-C++) if anybody is interested.
What you are saying about nulls doesn't make sense. Nulls are the result of several modes which cancel each other. Second, the room source-to-listener transfer function is linear. The response at the listener is the source response times the room source-to-listener transfer function. Reducing the source output at a null frequency will not reduce the depth of the null.
I don't have them because at that time I had no PC available and you didn't want to make the sound files available for download. Maybe now is a good time, please, pretty please
Thanks Dr. Geddes. So it seems that starting at 125 Hz, the narrower radiation pattern of a dipole compared to a monopole is more suited to delivering constant directivity over a wider BW than monopoles....which is a desired goal for all speakers.
Secondly, as JohnK pointed out, the lack of room pressurization by a dipole might be a reason a dipole sounds more natural to "some". Thinking more broadly, most bass musical instruments also do not pressurize the listening environment. True that a monopole can be equalised and LF extension decreased depending on the room, so that pressurisation does not occur. However it requires sophisticated measurements. OTOH, a dipole with sufficient displacement and LF extension could just fit in into any small to moderate sized room with minimal need for tailoring its response to the room.
Thirdly, if a dipole woofer is used beyond 125hz as in most 2 way applications, the bidirectional radiation at mid frequencies (around and above 500 Hz) weaves a more ambient/natural soundfield, as opposed to the almost unidirectional radiation by monopoles at these frequencies.
So in summary, even if there is no agreement (or scientific evidence) about the "perceived" difference of bass transient response between dipoles vs monopoles, based on the three ideas above, it should be logical as to why a dipole woofer may just sound more "natural" and real to some.
Last edited:
Soundaatma - that dipoles sound different is not contested. That they are inherently "better" is. People can convince themselves that "their dipoles" "sound better", but thats not proof of anything.
I don't really buy room pressurization as a factor because it is simply not a major effect. It is true that monopoles are more efficient as the frequency falls, and this could be a factor. Room pressureization only occurs below the first mode and then only if the room is highly sealed, which is extremely rare. My room is probably sealed far better than most and I don't see any response below the first mode with my five monopoles. I have seen this effect in small cars, but the volumes are vastly different than a home listening room.
There may be some truth to a dipole placed at random being preferable to a monopole placed at random, but that's simply a dumb thing to do, so what's the point? Isn't the question "How do I get the best bass in my room?" To that question, I do not find the dipole to be an advantage, more of a disadvantage due to its low efficiency. And the dipole has to be EQ'd, not doing so is not an option, so it should only be compared to an EQ'd monopole and then I think the results might be a wash.
The measurements aren't hard, the equipment can be obtained for < 100$ and the software is free. So taking the measurements is not an issue. Knowing what to do with them, as always, can be. Someday the software that I have to adjust the subs EQ will be available because its not difficult to do, but for now I need to keep it proprietary because its my only value added.
is not really correct, I would say that both have about the same bandwidth of CD, but the dipoles is higher in directivity. People often seem to misunderstand that CD and directivity are completly different things. There can be one without the other and visa-versa. What I strive for is CD with high directivity and in the range of 125 Hz - 500 Hz a dipole has an advantage in that it can achieve a narrower directivity than a direct radiator. Below 125 Hz, I don't believe that the source type makes a difference and above 500 Hz the dipole and the direct radiator are converging on the same directivity.
I don't really buy room pressurization as a factor because it is simply not a major effect. It is true that monopoles are more efficient as the frequency falls, and this could be a factor. Room pressureization only occurs below the first mode and then only if the room is highly sealed, which is extremely rare. My room is probably sealed far better than most and I don't see any response below the first mode with my five monopoles. I have seen this effect in small cars, but the volumes are vastly different than a home listening room.
There may be some truth to a dipole placed at random being preferable to a monopole placed at random, but that's simply a dumb thing to do, so what's the point? Isn't the question "How do I get the best bass in my room?" To that question, I do not find the dipole to be an advantage, more of a disadvantage due to its low efficiency. And the dipole has to be EQ'd, not doing so is not an option, so it should only be compared to an EQ'd monopole and then I think the results might be a wash.
The measurements aren't hard, the equipment can be obtained for < 100$ and the software is free. So taking the measurements is not an issue. Knowing what to do with them, as always, can be. Someday the software that I have to adjust the subs EQ will be available because its not difficult to do, but for now I need to keep it proprietary because its my only value added.
Last edited:
I don't have them because at that time I had no PC available and you didn't want to make the sound files available for download. Maybe now is a good time, please, pretty please
I would be interested as well, but I can certainly understand not diseminating your work for free!
This surely is true (as long as we stay in the linear regime which can be quite safely assumed) but it's not our point. It's about listening to the effect in realtime with known and controlled properties of the signal. OK, your method can be used to listen to it as well -- with headphones. Further, you can walk along the room and probe with your ears, again in realtime.
Also, another purpose is to check what noises the room (plaster walls, doors, windows, furniture) does make, as well as to check the speakers themselves (distortion, rattling, port chuffing).
It was meant as a quite versatile and useful tool for evaluation of several perceptional aspects of LF reproduction in rooms.
I'll open a new thread for this, here it will get lost and it is slightly off-topic anyway.
I always use a swept tone in a room to search for buzzes and the like. But I do like the idea of having some controlled tone bursts to listen for the effects. I can see value in that.
John, I was not aware that Excel could write wav files (but I am not much of an Excel user), but anyways its very easy to generate files like this in MathCAD. But heck, if someone else has done the work for me I'll bite!
John, I was not aware that Excel could write wav files (but I am not much of an Excel user), but anyways its very easy to generate files like this in MathCAD. But heck, if someone else has done the work for me I'll bite!
Of course not, I didn't meant that.
In modal region, a null (pressure node) at the listener position is (can be) a peak (velocity node) at another position in the room. So if you reduce the level at a null at the measured position, you can reduce the level of the peak at the other position.
Yes, of course. That is why I said source-to-listener transfer function. And that is the problem with room eq. EQing a what appears to be a peak at one position may result in a deeper null at another.
But if there's already a deep null, it is no problem if you set a notch on it, thus reducing energy of that mode. It can give you more uniform response all over the room (with, of course, deeper nulls, but reduced peaks).
The problem is to define what's a "deep" null. And to conclude from the null on the right parameters for the notch.
That's the difficulty when EQ'ing based on only one measurement at the listener position. I wrote a commercial software (will run in the subwoofers processor) that does exactly that: measure at one position, calculate the parameters for the PEQs. Only for peaks, not for the nulls, I disabled that feature because it was not possible to automate that. But from my experiments, you can get very good results if you place the mic on a corner of the room; and maybe have bad luck that at your listening position you suddenly have nearly no bass. But this is rare.
The problem is to define what's a "deep" null. And to conclude from the null on the right parameters for the notch.
That's the difficulty when EQ'ing based on only one measurement at the listener position. I wrote a commercial software (will run in the subwoofers processor) that does exactly that: measure at one position, calculate the parameters for the PEQs. Only for peaks, not for the nulls, I disabled that feature because it was not possible to automate that. But from my experiments, you can get very good results if you place the mic on a corner of the room; and maybe have bad luck that at your listening position you suddenly have nearly no bass. But this is rare.
There is definitely a need to do both: measurements and listening tests.
When one chooses to do the measurement using signals which can be listened it becomes possible to achieve 1-to-1 correlation with the measurement results and auditory perception. This is the whole purpose.
The perception is all the matters. If I wouldn't hear any difference, say between monopoles and dipoles, I propably wouldn't bother about all this.
- Elias
Well said. I am curious what degree of 1-to-1 correlation you can reach. Is it just "dipoles are more open" or can you relate individual visual details of your diagrams to details of your listening impression? Would you also be able to relate those details to things in the physical world?
Rudolf
Ok. I have made some measurements for monopoles, dipole and cardioid. I willmake several posts.
This first figure shows the near field response of the front source of the monopole (blue), dipole (green) and cardioid (violet). The individual responses have been Eq'ed such that when the dipole and cardioid front and rear sources are summed in the far filed all three formats will have the same on axis response.
This next figure confirns that when the front and rear sources of the dipole and cardioid are summed in the far field the response matches themonpole.
The next three plots are the in room response of the three sources. Note the lables in the lower left corner of the plots. The front source of each format was at the same position. That is, all sources had the acoustic center at the same location.
Note that the dipole and cardioid are very similar and that the monopole shows some room pressurization.
This first figure shows the near field response of the front source of the monopole (blue), dipole (green) and cardioid (violet). The individual responses have been Eq'ed such that when the dipole and cardioid front and rear sources are summed in the far filed all three formats will have the same on axis response.
An externally hosted image should be here but it was not working when we last tested it.
This next figure confirns that when the front and rear sources of the dipole and cardioid are summed in the far field the response matches themonpole.
An externally hosted image should be here but it was not working when we last tested it.
The next three plots are the in room response of the three sources. Note the lables in the lower left corner of the plots. The front source of each format was at the same position. That is, all sources had the acoustic center at the same location.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Note that the dipole and cardioid are very similar and that the monopole shows some room pressurization.
Next, here are burst respons eplots presented in sonogram and CSD form. Note that the plots are in cycles so a line of constant time is shown in the sonograms for aid in interpretation. Again. see the lower left corner of the plots for the label.
Monopole:
Dipole:
Cardioid:
CSDs:
Monopole:
Dipole:
Cardioid:
Make what you like of the results. IMO, the dipole and cardioid preform pretty similarly. But recall that the cardioid is 6dB more efficient and cabn play 6dB louder. I may add some EQ to extend the bass response flat to 30 Hz.
Monopole:
Dipole:
Cardioid:
CSDs:
Monopole:
Dipole:
Cardioid:
Make what you like of the results. IMO, the dipole and cardioid preform pretty similarly. But recall that the cardioid is 6dB more efficient and cabn play 6dB louder. I may add some EQ to extend the bass response flat to 30 Hz.
Thanks John - completely plausible results. Its good to see real data.
It's clear that there is some effect below the first mode and hence the "room pressurization" (its not a dominate effect, but it is clearly there.). It's really the only significant difference in the results that I see.
I am sure people will see huge differences where I see little, but in these last few weeks before the election that's pretty common place.
It's clear that there is some effect below the first mode and hence the "room pressurization" (its not a dominate effect, but it is clearly there.). It's really the only significant difference in the results that I see.
I am sure people will see huge differences where I see little, but in these last few weeks before the election that's pretty common place.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.