[marvda1]

With the speakers facing straight ahead, is there a difference between toeing the front of the speakers in as opposed to toeing the rear of the speakers out ?

[Gordy]

Hi Marvda1,

Yes, toeing in the front will also change the driver distance / listener distance ratio slightly, while toeing the rear out will not.  That's the only difference that readily comes to mind... this early!

[warnerwh]

The bottom line is that in your room with your speakers and taste the speaker position needs to be experimented with. This is a free and very important tweak and absolutely necessary if you want the best sound.  The answer to your question is it really doesn't matter.  Like in real estate the three most important things for speakers is location, location, location.

[andyr]

Hi Marvda1,

Yes, toeing in the front will also change the driver distance / listener distance ratio slightly, while toeing the rear out will not.  That's the only difference that readily comes to mind... this early!

Gordy,

Mathematically, you are correct.  However, sitting up straight in your listening couch/chair/whatever rather than slouching back, will move your ears a far greater distance forward than the difference in driver-centre-to-your-ears from toeing in/toeing out ... so your post is rather spurious, to my way of thinking.

Regards,

Andy

[ekovalsky]

If you have line arrays or other speakers with side by side drivers, the time alignment of the mid and tweeter lines can be adjusted by varying the toe in.  For instance, changing the listening position from 30 degrees off axis (speakers firing directly ahead, equilateral triangle arrangement) to 0 degrees off axis (toed in directly at listening position) will change the relative distance between the listener and driver lines by about 1.25".  This doesn't sound like much assuming a typical listening distance of 10'.  But it actually amounts to a 90 degree phase change at a 2700hz crossover frequency, or additional 0.12msec delay group delay to the outer driver line.  This is definitely audible !

Some drivers (particularly some planars in my experience) also exhibit major shifts in frequency response when going from off to on axis.  

Check with your speaker manufacturer regarding recommended toe in, then experiment from there.  Equipment to perform objective time and frequency domain analysis really helps.  Toe in can definitely be a 'free tweak' but it it can also mess things up  

[andyr]

If you have line arrays or other speakers with side by side drivers, the time alignment of the mid and tweeter lines can be adjusted by varying the toe in.  For instance, changing the listening position from 30 degrees off axis (speakers firing directly ahead, equilateral triangle arrangement) to 0 degrees off axis (toed in directly at listening position) will change the relative distance between the listener and driver lines by about 1.25".  This doesn't sound like much assuming a typical listening assuming a typical listening distance of 10'. But it actually amounts to a 90 degree phase change at a 2700hz crossover frequency, or additional 0.12msec delay group delay to the outer driver line. This is definitely audible !

Some drivers (particularly some planars in my experience) also exhibit major shifts in frequency response when going from off to on axis.
...

I'm aware of the phase change and I agree with you that this is a significant factor ... for this reason, I have my 3-way Maggies pointing directly at my ears.  I don't listen in an equilateral triangle ... more like a triangle which has distance from my ears to the base line between the Maggies = 1.5 this baseline distance.  And when I say "pointing directly at my ears", I mean the perpendicular from a point mid-way between the central lines of the ribbon and the mid-panel, points towards my ears (this is not the centre of the overall Maggie panel).

Maggies are carefully designed so that the "acoustic centre" of their drivers are in the same vertical plane.  I cannot understand why people listen to their Maggies with far less toe-in than what I have ... as this will affect the tonal structure of the music they hear, as well as giving them the phase problems you referred to.

Regards,

Andy

[_scotty_]

Maggies also exhibit pronounced comb-filtering effects in the horizontal
plane due to the crossover slopes employed and should be adjusted for the
smoothest frequency response possible at the listening position.  In the case of their 2way systems with a 6dB/oct crossover the key to impulse and phase accuracy as well as imaging lies in achieving time-alignment at the listening postion via the correct toe-in, this is the equivalent of driver setback in a speaker with a vertical driver arrangement.
Depending on the vintage the speakers vertical rake angle should be tweeked for the best high frequency extension at the listening postion as well.
My MGIIIs and SMGs both benefited from a deviation from the factory rake angle towards a more upright postion.  
Scotty

[andyr]

... the key to impulse and phase accuracy as well as imaging lies in achieving time-alignment at the listening postion via the correct toe-in, this is the equivalent of driver setback in a speaker with a vertical driver arrangement.

Scotty ...

Hi Scotty,

I agree with your statement, however, the acoustic centres of Maggie drivers ARE all in the same plane ... thus, surely, correct time alignment means they should be toed-in so they are perpendicular to a line from your ears?

... My MGIIIs and SMGs both benefited from a deviation from the factory rake angle towards a more upright postion. ...

The feet which came with my IIIAs stood them vertical.  Are you saying that funny wooden baseplate which IIIs have tilts them forwards?  (I've only seen one pair of IIIs and i can't remember!)

Regards,

Andy

[ekovalsky]

Scotty,

I'm not very familiar with the Maggies.  Do you know the specifics of their crossovers ?  

Just because the acoustic centers of the drivers are in the same plane doesn't necessarily mean they will be time aligned set up as you described.  For instance, a 4th order LR crossover between the mid and tweeter would introduce a 360 degree phase shift on the tweeter.  In that case, for proper time alignment you would want the speakers set up with less toe in and the tweeters on the inside, thus relatively increasing the distance between the listening position and midrange compared to the tweeter and delaying the arrival of sound from the midrange.

[andyr]

... Just because the acoustic centers of the drivers are in the same plane doesn't necessarily mean they will be time aligned set up as you described.  For instance, a 4th order LR crossover between the mid and tweeter would introduce a 360 degree phase shift on the tweeter.  In that case, for proper time alignment you would want the speakers set up with less toe in and the tweeters on the inside, thus relatively increasing the distance between the listening position and midrange compared to the tweeter and delaying the arrival of sound from the midrange...

Two Qs:

1.  What about the case of a 2nd order LR crossover between mid and tweeter?  This introduces a 180 deg phase shift so the midrange driver is inverted to end up in an "in phase" situation.  Surely this would require equal distance for midrange and tweeter to the ear?

2.  In a 3-way Maggie, which is the more critical crossover to get phase-aligned ... the base/mid or the mid/tweeter?

Regards,

Andy

[ekovalsky]

Two Qs:

1.  What about the case of a 2nd order LR crossover between mid and tweeter?  This introduces a 180 deg phase shift so the midrange driver is inverted to end up in an "in phase" situation.  Surely this would require equal distance for midrange and tweeter to the ear?

2.  In a 3-way Maggie, which is the more critical crossover to get phase-aligned ... the base/mid or the mid/tweeter?

Regards,

Andy

As a 2nd order filter will produce 180 degree phase shift, the tweeter is typically inverted like you say.  This results in "in-phase" output from the mid and tweeter, and the summed group delay should be flat.  But remember phase and time are related -- a 180 degree phase shift on the tweeter means its output is delayed by one half a wavelength compared to the midrange.  This amounts to about 4" at a 1700hz crossover frequency!   So it is probably fair to say the Maggies are phase-aligned but not time-aligned (which is a rarity in loudspeakers).

On a three way system I would think the bass/mid crossover is the more important to align, since phase problems are generally more audible at lower frequencies.

It would be an interesting experiment to bypass the Maggie crossover and tri-amp with a zero phase, steep DSP filters...

[andyr]

As a 2nd order filter will produce 180 degree phase shift, the tweeter is typically inverted like you say. This results in "in-phase" output from the mid and tweeter, and the summed group delay should be flat. But remember phase and time are related -- a 180 degree phase shift on the tweeter means its output is delayed by one half a wavelength compared to the midrange. This amounts to about 4" at a 1700hz crossover frequency! So it is probably fair to say the Maggies are phase-aligned but not time-aligned (which is a rarity in loudspeakers).

On a three way system I would think the bass/mid crossover is the more important to align, since phase problems are generally more audible at lower frequencies. ...

Hi, ekovalsky,

Am I understanding it correctly ... in the above example, you are saying the mid and tweeter are phase aligned but not time aligned?
If one of the two (the tweeter?) were located 4" behind the acoustic centre of the mid-driver (ie. the mylar sheet), then the two drivers would also be time aligned??
Does this physical location not throw out the phase alignment?
And at 2,600Hz (the actual crossover point) are we talking 2.6" not 4"???

And the bass/mid is more important to get right ... mmm.  

Thanks,

Andy

[DVV]

You might also want to try, for the hell of it, to position your speakers about 2 feet upwards, i.e. so the the tweeter is approximately 2 feet above your straight line ear level.

Depending on the speakers and the room, this can produce very interesting effects. With live music in particular, it can sometimes give you the feeling of actually being there. Small clubs don't often have ideal acoustics.

If you have classic stereo two channels, do try elevating them those 2 feet above the ear level, and also try using the speakers at zero angle to you, i.e. perfectly parallel with the walls.

If you happen to have a summing center channel (one which mixes the L and R signals into one, as proposed by Paul W. Klipsch way back in 1949, in cases when a true center channel is not available), the middle speaker should be towards you at ear level, while the L and R speakers should be above your ear level by about 2 feet, parallel with the walls.

This can produce amazing results even in small rooms. However, I must also add that this works really well with now all but disappeared monitor type speakers (the classic 2-by-1 BBC format), such as for example JBL 4312 (Century 100), AR3a, AR11, etc.

Regarding driver phase coherence, this seems to have been forgotten on modern speakers. Thankfully, I own speakers (B&M Acoustics, now defunct loacl Belgrade, Serbia company) which are angled 9 degrees front to back, so at 6-7 feet, the drivers are in fact phase alligned.

Cheers,
DVV

[ekovalsky]

Hi, ekovalsky,

Am I understanding it correctly ... in the above example, you are saying the mid and tweeter are phase aligned but not time aligned?
If one of the two (the tweeter?) were located 4" behind the acoustic centre of the mid-driver (ie. the mylar sheet), then the two drivers would also be time aligned??
Does this physical location not throw out the phase alignment?
And at 2,600Hz (the actual crossover point) are we talking 2.6" not 4"???

And the bass/mid is more important to get right ... mmm.  

Thanks,

Andy

That's basically it.  But I think it is the tweeter that is delayed relative to the midrange, so the you'd want the midrange positioned further away than the tweeter to time align the system.  So, the speaker toe in would result in a convergence behind the listening position, with the tweeters on the inside.  You are correct -- at 2,600hz the distance would be 2.6" rather than 4" (1128 ft/sec divided by freq * 12 in/ft * 0.5 for half a wavelength).  And you also noticed that this physical repositionig would throw out the phase alignment, but that could be restored by running the midrange and tweeter in phase instead of out of phase with each other.  Right now they are wired out of phase to compensate for the lack of time alignment.

There are only a few manufacturers who build time aligned speakers, while phase aligned speakers are pretty common.  So probably if you can only have one, phase aligned is better.  I enjoyed Apogee Divas (not terribly dissimilar to your Maggies) for years without ever thinking about this, so I wouldn't dwell on it.  The Apogees did use an outboard electronic crossover and I am uncertain what phase abberations it induced.  Apogee did strongly recommend the speakers be used without any toe in, i.e. listening at 15-30 degrees off axis.  What does Magenpan suggest ?  Of course the ideal listening axis involves more than just phase and time considerations, as drivers (particularly planars and ribbons) exhibit frequency response swings too.

Digital crossovers are interesting in that they can have linear phase and high slopes.  They are also much slower than analog crossovers, particularly low pass filters at low frequencies.  So it is easy to achieve  both phase & time alignment but this requires multi-amping with programmable delays between driver groups.  I have the TacT RCS 2.2X which is an amazing piece of equipment but it takes additional, manual steps to accomplish this -- the software ignores the variable lag introduced by high and low pass filters and sets delays based only on the physical position of the mains and subs.  Additional delay (17.2 msec in my setup) has to be added to the mains because of the slower low pass filter.  Doing this made a huge difference and resulted in what I believe is a rare, perfect integration of mains and subs.

[_scotty_]

ekovalsky,the early series of Maggies that were 2way designs all had a
simple 6dB/oct crossover. They deviated from this when the MGIII was brought out. The MGIII had a 2nd order parallel network but the values of the
components used on the tweeter to midrange xover did not conform to any specific type filter design,[ie,buttertworth,bessel,etc]. I changed the filter
to a 12dB/oct linkwitz/riley  and this eliminated the harshness the ribbon tweeter had. From phase-coherent standpoint the the MGIII is unfixable with conventional crossover technology, I too would like to hear the results of a high order DSP crossover applied to Maggies.  In regards to vertical rake, the older Maggies had the speaker mounted towards the front of the baseplate with speaker tilted back so it wouldn't fall forward and crush something. The new designs all appear to be vertically oriented.
As far as how the toe in should be set on a three-way maggie I would set the up for the flatest response possible at the listening position. This is the only thing you can get right because of the phaseshifts induced by the crossover.
These phaseshifts cannot  be compensated for or eliminated by toe-in
any more than they can be fixed in a conventional system when this type of crossover network is used.
Because of this I wouldn't worry about phase issues,I would recommend
toeing the speakers in to suit yourself and enjoy the music,anything else may lead audiophilia -nervosa and insanity.
Scotty

[ekovalsky]

ekovalsky,the early series of Maggies that were 2way designs all had a
simple 6dB/oct crossover. They deviated from this when the MGIII was brought out. The MGIII had a 2nd order parallel network but the values of the
components used on the tweeter to midrange xover did not conform to any specific type filter design,[ie,buttertworth,bessel,etc]. I changed the filter
to a 12dB/oct linkwitz/riley  and this eliminated the harshness the ribbon tweeter had. From phase-coherent standpoint the the MGIII is un ...

Agree!