Horizontal dispersion, same as when you cross over anything too high. The lower driver's pattern starts narrowing, then as the upper driver comes in, the polar pattern spreads out. You see this in off-axis measurements as a large dip and peak. As a consequence, the reverberant energy (and thus power response) becomes very non-flat and the speaker will sound colored in anything but a very absorbent room.
Hi,
The vertical lobing of an MTM is much worse that an MT, and gets
worse the higher you c/o.
Also the vertical dispersion of an MTM is poor compared to an MT,
this "focussing" is claimed to be an advantage for HT, again it
gets worse the higher you c/o.
🙂/sreten.
The vertical lobing of an MTM is much worse that an MT, and gets
worse the higher you c/o.
Also the vertical dispersion of an MTM is poor compared to an MT,
this "focussing" is claimed to be an advantage for HT, again it
gets worse the higher you c/o.
🙂/sreten.
SY, sreten,
Thank you for the reply.
While I did a more careful search on the subject, I found myself wondering if a digital crossover would allow me the higher cross point by delaying the drivers signal. This way I could emulate a coaxial design, with no lobing problems. Aditionaly the space between drivers will be reduced and the tweeter mounted to help the delay.
Furthermore, a dsp would resolve crossover and driver phase related problems.
All this I think I could implementate with brutefir and drc. Do yo think the trick could hold up? 😀
Thank you for the reply.
While I did a more careful search on the subject, I found myself wondering if a digital crossover would allow me the higher cross point by delaying the drivers signal. This way I could emulate a coaxial design, with no lobing problems. Aditionaly the space between drivers will be reduced and the tweeter mounted to help the delay.
Furthermore, a dsp would resolve crossover and driver phase related problems.
All this I think I could implementate with brutefir and drc. Do yo think the trick could hold up? 😀
In an MTM the maximum XO is determined by the centre-to-centre spacing of the 2 Ms. The only way to get a higher XO to work well is to use smaller Ms.
dave
dave
planet 10 and bjorno,
thanks for the input. Well I've used the same program to simulate this, only that I've set the distance between drivers near 0. The reason is that I think I can emulate this effect by delaying the signal to the drivers so that I have virtualy no distance between. I wonder if someone tried something like this...
I know Linkwitz uses some kind of delay to his Phoenix baffle (an MTM also), but I don't know if this is his objective..
thanks for the input. Well I've used the same program to simulate this, only that I've set the distance between drivers near 0. The reason is that I think I can emulate this effect by delaying the signal to the drivers so that I have virtualy no distance between. I wonder if someone tried something like this...
I know Linkwitz uses some kind of delay to his Phoenix baffle (an MTM also), but I don't know if this is his objective..
Hi,
I think in the Xdir the distance refers to the actual physical distance between the tweeter and the woofer.
What I am not sure is the "Woofer Phase lag". Is it calculated from the tweeter offset to the driver distance? or is it the actual phase difference of the two drivers taking into account of the XO network, baffle step response, and the natural driver rolloff characteristics?
I found that when the "Woofer Phase lag" is around 90 then the polar response can be acceptable even if the driver distance is a bit longer or the XO point higher. When it is around 0 then the polar response is totally unacceptable. How can this happen? Why people do time-alignment then? But what is the definition of "Woofer Phase lag" in the first place?
I hope the experts can shed some lights.
Bill
I think in the Xdir the distance refers to the actual physical distance between the tweeter and the woofer.
What I am not sure is the "Woofer Phase lag". Is it calculated from the tweeter offset to the driver distance? or is it the actual phase difference of the two drivers taking into account of the XO network, baffle step response, and the natural driver rolloff characteristics?
I found that when the "Woofer Phase lag" is around 90 then the polar response can be acceptable even if the driver distance is a bit longer or the XO point higher. When it is around 0 then the polar response is totally unacceptable. How can this happen? Why people do time-alignment then? But what is the definition of "Woofer Phase lag" in the first place?
I hope the experts can shed some lights.
Bill
As far as I know you will have to cross lower than 2khz to get a true d`appolito alignment, and use very small mids to get them close enough, and you will have to use steep filters because of the low cross point
Coherense is often forgotten, which means that the mids should play highs of the same quality as the tweeter, and the tweeter should play the mids of same quality as the mids does - just like the bass needs to play mids of same quality as the mids does, and so forth ... kind of logical, isnt it
Its possible but not easy
Remember that with MTM you will need to have tweeter precisely at listening height or theres a big risk of phase problems - generally I dont like it because it results in a very tall speaker, but with the use of multiple drivers you may have to use such designs
Another option could be the use of a wide front, which would make the small drivers act like bigger ones with loss of disspertion .... and maybe more coherent
Coherense is often forgotten, which means that the mids should play highs of the same quality as the tweeter, and the tweeter should play the mids of same quality as the mids does - just like the bass needs to play mids of same quality as the mids does, and so forth ... kind of logical, isnt it
Its possible but not easy
Remember that with MTM you will need to have tweeter precisely at listening height or theres a big risk of phase problems - generally I dont like it because it results in a very tall speaker, but with the use of multiple drivers you may have to use such designs
Another option could be the use of a wide front, which would make the small drivers act like bigger ones with loss of disspertion .... and maybe more coherent
Hi Bill,
Ok for a TM but for a MTM? See picture
The drivers are point sources, perfect FR with no roll of at all = straight FR line in plot and minimum phase type, thus when plotted also follows the FR curve parallel, no baffle step response added and…
…only.
Ok.
When 90 degrees the different drivers are in quadrature and the FR sum flat but the polar will tilt up or down.
Increase the attenuation for either T or M but not both at the same time e.g. if phase lag is 0 and T and M are –3dB at the crossover frequency you can see what happens when woofer is changed to – 24 or vice versa.
If you increase both, only the absolute sum-level will change (FR will dip) but by decreasing the attenuation for one of the drivers it is equivalent to move into the pass band i.e. where T is receiving more power than the M.
Result: if you want to sum in-phase at crossover you must chose a steep filter to minimize the possible ‘bad’ behaviour (FR dips) within the crossover band caused by the non-coincident drivers.
http://www.geocities.com/kreskovs/TimeAligned1.html
There is no need for time alignment when using X-dir, as all types of drivers in that respect are physical identical.
An M driver phase normally lags at crossover due to the use of an inductor in series and a T driver leads in phase due to the corresponding series capacitor.
The T driver phase is set to the reference phase = 0 degrees at the crossover frequency, thus the relative phase difference that always exists in real crossovers is held by the M phase lag value.
b
Ok for a TM but for a MTM? See picture
The drivers are point sources, perfect FR with no roll of at all = straight FR line in plot and minimum phase type, thus when plotted also follows the FR curve parallel, no baffle step response added and…
…only.
Ok.
When 90 degrees the different drivers are in quadrature and the FR sum flat but the polar will tilt up or down.
Increase the attenuation for either T or M but not both at the same time e.g. if phase lag is 0 and T and M are –3dB at the crossover frequency you can see what happens when woofer is changed to – 24 or vice versa.
If you increase both, only the absolute sum-level will change (FR will dip) but by decreasing the attenuation for one of the drivers it is equivalent to move into the pass band i.e. where T is receiving more power than the M.
Result: if you want to sum in-phase at crossover you must chose a steep filter to minimize the possible ‘bad’ behaviour (FR dips) within the crossover band caused by the non-coincident drivers.
http://www.geocities.com/kreskovs/TimeAligned1.html
There is no need for time alignment when using X-dir, as all types of drivers in that respect are physical identical.
An M driver phase normally lags at crossover due to the use of an inductor in series and a T driver leads in phase due to the corresponding series capacitor.
The T driver phase is set to the reference phase = 0 degrees at the crossover frequency, thus the relative phase difference that always exists in real crossovers is held by the M phase lag value.
b
Attachments
This is interesting stuff! Probably a good job I'm about to change my MTM to 2.5 way as I'm crossing over at about 5khz as the stock crossover point made the sound unlistenably harsh.
My speaker was obviously designed on paper, and I changed it by ear and did some harm as well as the obvious good! Now I'm trying to make it all good!!! This thread is very interesting to me as it partly explains the difficulty I've found in making the crossover sound just right.
Simon
My speaker was obviously designed on paper, and I changed it by ear and did some harm as well as the obvious good! Now I'm trying to make it all good!!! This thread is very interesting to me as it partly explains the difficulty I've found in making the crossover sound just right.
Simon
MTM sounding right is one important primary issue to target but there are many other issues that make the MTM concept difficult to get right and one of them that bother me is the wider vertical localisation error compared to other configurations at shorter distances e.g. when listening close to a normal MTM.
A time ago I discussed this with a friend and saved some notes I made that maybe could wake up someone’s others interest?
See the pics that I first hesitated to submit due to the inherent information quality I really think I should improve before posting in order to avoid the risk of getting too many questions back in return….
Observe, it’s only about comparing localisation and probably the direct appearance (within time- intensity-trading < 700 uS) of apparent phantom sizes not including horizontal/vertical frequency dependent azimuth errors that smears and blurs the whole soundstage further and of course not how the MTM sounds.
That’s another story with separate questions concerning primarily listening preferences and thus beyond the scope of the pictures.
b
1(2
Attachments
Simon, you might need a "series-notch" in paralel with upper midwoofer, and maybe a "paralel-notch" in series with your low woofer 😉
Those pictures need a bit of explanation for simpletons like me to understand.
I don't really like the unnatural way the MTMs image, and nor do I like the forward striking presentation (at least with the default crossover). Also they sound very cold (again with that default crossover).
I wonder how many of these problems are because it's an MTM, and how many are just "not to my taste" features of the sound. I'm sure a 2.5 way will sound nicer, just as soon my my 3.9mH coils arrive...
Simon
I don't really like the unnatural way the MTMs image, and nor do I like the forward striking presentation (at least with the default crossover). Also they sound very cold (again with that default crossover).
I wonder how many of these problems are because it's an MTM, and how many are just "not to my taste" features of the sound. I'm sure a 2.5 way will sound nicer, just as soon my my 3.9mH coils arrive...
Simon
- Status
- Not open for further replies.