[Rudolf]

With regard to dipoles the smaller or at least different excitation of room modes is mentioned as a big advantage. Since I could not find clear and comprehensible measurements for this, I have tried it myself. The arrangement is as presented:

Measured objects are two H frames up to 200 Hz with small open baffles on top taking over from 200 Hz. For all measurements the OBs were positioned 45° to the space axes. A symmetrical arrangement of the dipoles was not possible for practical reasons. As recommended by Linkwitz and others the loudspeakers are just as far from the front wall as the listening position (Hoerplatz) is off the rear wall. All measurements took place at the listening position. Both loudspeakers were active for all measurements. Distance from the listening position to the loudspeakers is 2.4-2.5 m.

First the comparison of dipole orientation along the longitudinal axis (red) and transverse axis (black). (Caution: These colors do not correspond with the colors in the room sketch!):

One can see, how with the alignment of the dipoles along the longitudinal axis the longitudinal modes at 33 and 66 Hz are energized. The modes at 100 and 133 Hz are less pronounced.
With placement in transverse direction the first longitudinal modes are substantially less energized. But we get transversal modes at 43 and 86 Hz. Interestingly these have been excited in the longitudinal placement to the same degree. A further transverse mode with 127 Hz forms a plateau with the longitudinal mode of 133 Hz.

Generally it can be stated that the longitudinal arrangement supports all transverse room modes to the same degree, while the dipoles in transverse arrangement - working against each other – actively suppress the lowest longitudinal mode.

I find the compromise between longitudinal and transverse arrangement particularly interesting:

Here the colors of the frequency responses agree with the room sketch. In my opinion the black curve (45°-positioning) makes for a good averaging between longitudinal and transverse alignment. Have I to mention that it sounds best too?!

[DanTheMan]

   I'll have to read this again.  Sounds like interesting work to say the least.  Nice job.

[scorpion]

Yes, interesting measurements, Rudolf. The 45-45 measurement clearly stands out at least in low bass. The usual 80 Hz excitation (listening rooms are perhaps more like each other than is perhaps thought) might be less influenced. From 80 Hz and upwards the 45-45 response is better or no worse than the other alternatives. Using my experience with toe-in placement I think you are right. But do you think that 45 degree toe-in also is optimal for the whole frequency spectum ? But I suppose the angles were chosen to give clear and easily understandable measurements. 

Dipole toe-in is of course also interesting with regard to the overall reflections into the listening spot. That is also one ot the dipole strenghts that you with the help of the dipole null and the room can shape the response quite a bit. I cannot sit in the 'ideal position' in my room and the wall behind me is also not solid but with a double-door opening. The speakers are placed 1 meter from front wall and nearest corner. But I can listen to my speakers standing in the 'ideal position'.My findings are that you probably normally will benefit by toe-in the speakers a bit more than you would think. My angle is about 30-35 degrees which I think gives best clarity and that is a bit off-axis both in the listening spot and certainly in the 'ideal listening position'.

/Erling

[hurdy_gurdyman]

My speakers are set-up very similar to Erling's speakers. I have found this to have the best OB image in my room, and frequency balance is good. They are toed in enough to be crossing quite a bit in front of my listening seat. Bass seems fine here.

Dave aa

[ooheadsoo]

The 45 degree position seems to have the dipole null located at the first reflection point, which seems like a good idea.

[Rudolf]

Erling,
I see a lot of reasons for speaker toe-in. And they are quite different for woofer and mid/treble IMHO.
Dipole woofer:
I have only shown the benefit for manipulating the lowest room modes. In fact my H-dipoles are now toed in even 60° from the long room axis to give best results. That way the dipole axis points to the room corners and not perpendicular to the front wall. This results in a longer virtual speaker-to-wall-distance, which is always a benefit for dipole woofers. John K. has shown in one of his Tech. Studies http://www.musicanddesign.com/PowerMatching.html how to best match a dipole woofer to a dipole midrange powerwise:

Looks familiar, doesn´´t it?

Dipole midrange:
My midrange drivers are equalised to give best response at ~30° off axis. It seems to be a good compromise, if one wants to achieve a good direct response and a balanced room response with drivers that rise in directivity with higher frequency. In combination with my listening position this results in the midrange baffles toed in almost 20° with regard to the woofers. Needless to say that the null looks pretty much to the first reflection point as ooheadsoo already remarked.

It´s comforting to see other people with a bunch of dipole experience coming to the same geometric set up (by and large). Obviously our ears hear what theory predicts

Rudolf

[scorpion]

Rudolf,

We have an interesting discussion going on on this subject in Swedish Hi-Fi Forum. There is a spreadsheet (in Swedish, not very hard to understand I think) constructed to dimension quarterwave bass-traps. It is working very good, evidently. There is also a testsignal suit (frequencies are counted by a voice in Swedish also understandable I think) My room was excitated by my dipoles around 80 Hz very clearly (also evident in measurements) so I am constructing an 85 Hz trap to begin with.  I thought that 50 Hz was also a troublesome frequency, but not really according to the test.

Links: Spreadsheet: http://bildarkiv.hififorum.nu/esl/Baspipa/Tubecalcorgelpipa_version1.xls
Testtones: www.hififorum.nu/artiklar/downloads/HiFiForumtesttones/20to120in100sek500mV.wav

/Erling