why bother with a crossover?

[lt cdr data]

Reading the AVI speaker design philosophy :

http://www.avihifi.co.uk/

They maintain the most important aspect of xover design is phase.

Now to my mind, this implies that phase is time delay, and at the measurement point, both drivers emanating in phase will have the hi freq and lo freq arriving to the listeners ear at the same moment in time, thus from physics theory, no wave cancellations will occur, which will disrupt the frequency response.

Sounds good, but what if the frequency response is flat at measurement anyway, doesn't that imply a good design, and nothing bad going on?

And we cannot 'control' the phase in amplifiers (not referring to stability), its altered unpredictably by the internal capacitance in the devices or output transformers, to a bizarre degree, doing goodness knows what to our sound.

ok? The avi philosophy has much to commend though.

Well, its not been written much, but put your speakers in your room, at different temperatures, and humidities, which will alter the speed of sound, perhaps not significantly at listening distances, but

each room is uniquely shaped

with unique combinations of furnishings, curtains, tables, etc., call them soft furnishings and hard furnishings for simplicity.

basic reverb science as pioneered digitally by lexicon, TC, et al., will tell you sound is direct signal and reflected signal, too, so

at the listening point, due to the room features, there will be all manner of the frequency response and the phase response, being altered ( by reflections, refractions, cancellations, superpositions and diffractions, along with standing waves and goodness knows what else)

(which explains why a sub. and a multi way is more difficult to implement design wise or sub to speaker blend, due to position, timwise perhaps and phase, ie how many pay attention to distances from speakers when positioning a sub?, etc.)

So what you get measured in a perfect environment is no correlation to the measurement you will get in situ.

So...that means in essence that all the time and trouble given to perfect all these parameters of the xover, frequency, and phase, on and off axis is totally and utterly futile as its modified uniquely by everyone's different room.

And also, any attempt to infer the sound quality based upon those measurements again is madness as they will change totally due to the above reasons.

( and tell me how you can say what sound a speaker will make on violins from its measurements?) laughable really when you think about it, and some reviews/ers really do attempt this inference.

So, speaker measurements based on freq. and phase to perfect them are largely pointless in perfect environments, transferred to imperfect ones.

This has never to my mind been discussed, only alluded to in studio design a little.

But I think its a massively relevant point, perhaps divisive, controversial, but as I said, not been thought of really at all, and bears considerable analysis, and investigation.

So, the xover is pointless, just protect your speakers and your amplifier ( with a nice impedance curve), and get a sound good for you in your room, as that's all you can do with design, as design in a lab is futile, utterly, based on the above.

And so much for 100s of hours of xover tuning in a lab.

And to the manufacturers, you may as well simply tune by ear in an 'average' furnished room, and save all those hours slaved in the 'lab' just plug in a few parts and when you are happy...

And to the amateur struggling to learn how to design a perfect xover, you needn't worry, as this analysis is to what modern crossover design what learning the earth wasn't the centre of the universe to the church is, I believe

[J.R.Freeman]

Hello there!

With regards to the importance of a crossover: in a conventional, multiple way electrodynamic loudspeaker system, it is arguable that the crossover is indeed of great importance.

For example, consider a common 2 way system, such as the average "bookshelf" loudspeaker. Each driver in the system has a useful range that is a subset of the full audio spectrum. If driven to work outside of that range, sound quality is degraded, and damage to the drivers can occur.

Each driver will have a range through which their respective distortion numbers are best. Perhaps our fictional midbass has limited excursion, and its distortion rises heavily below 55 Hz at moderate volume levels. Well in that case, a crossover (likely active) could limit low end response to get the most out of our driver. Also, our midbass extends nicely up to the high-midrange band, but breaks up harshly after that. The crossover would make sure the tweeter handles these frequencies.

Do you see what I mean? The crossover has a few important tasks. Phase is certainly important, but no more than other concerns at play here. Another crossover function is level-matching. It is likely that your drivers have different characteristic sensitivities. The crossover adjusts the sensitivities of the drivers (in the case of a passive network, this means padding the more sensitive driver, usually the tweeter) to match. Ensuring the drivers work within their useful range also effects power handling of the system. There is also baffle diffraction step compensation to consider.

In short, the crossover serves many purposes. With regards to room interactions: you have to put the horse before the cart. You are suggesting room acoustics should be solved in the crossover. Room placement and setup are important, but you should start with as close to a neutral system as possible, I believe. If you know you are using your speakers close to a boundary, though, it's true that some tune their crossovers to take this into account. But if your room produces major acoustic anomalies, the problem is with the room, not the crossover.

JF

[Andy G]

DOH !

The AVI speakers have a crossover, an active one 4th order at 3.4k, (seems a bit high for proper power response)

And they have measured them in a "perfect" environment else how could they mention figures like +/- 2dB ?

All we have is a a basic active 2-way with the x-o built into the box with the amps, nothing more, nothing less.

How it sounds is all up to the skill of the person who has implemented the x-o and the cabinet., as with any loudspeaker.

And we all know that the room has considerable effects on the sound, that will apply to this speaker, just like any other loudspeaker.

[J.R.Freeman]

Well said, Andy.

The tone of that AVI page, from first glance, seems rather self-righteous. DSP is certainly an attractive technology, but the laws of physics are still in place. For a home-theater-in-a-box kind of setup, I'm sure this is a nice one :P

JF

[DcibeL]

Crossovers- High quality 4th order discrete analogue.

Bass driver - Very high power handling 6.5" paper cone drive unit with 1.5" voice coil and exceptionally broad bandwidth enabling phase perfect crossover at 3.4 kHz. Bass extension better than - 6dB @ 60Hz.

Analogue active crossovers are not phase perfect, they exhibit the same phase shifts over frequency that passive crossovers do. If you want a crossover with linear phase you require digital FIR filters. Look at the NHT XdS for a good example.

On top of that, 3.4kHz crossover point for a 6.5" driver is too high, beaming will be an issue. 3.4kHz wavelength is approx 4".