I prefer a high-order crossover especially in a car environment, because it minimizes the effects of spatially-dependent interference from two speakers that are not located in the same place. I say especially in a car environment, because in my home speaker, at least my tweeter and midrange faceplates are touching. In my car, the tweeter is at the top of the door and the woofer is a couple feet down at the bottom of the door. Another benefit of the sharper crossover is relaxing the requirements for the two drivers to respond identically over a wide range of overlapping frequencies. If your tweeter won't make noise below 9kHz and your woofer breaks up and sounds like hell above 3kHz, a low-order crossover isn't going to do it for you. The answer is more drivers, which I don't like philosophically, and is practically difficult in a car. Phase shifts in a 4th-order crossover are minimized by using the Linkwitz-Riley tuning (two 2nd-order Butterworth filters in cascade). I honestly don't know what filter tuning Alpine uses in their digital crossovers. They could even be using an FIR filter which could result in no phase distortion (just a constant group delay at all frequencies -- linear phase response). But since they give options for 1st-4th order filters, I expect that they are using the IIR equivalents of standard analog filters.

The time alignment is a digital delay, and therefore adds in only group delay -- that's a phase shift that's proportional to frequency -- which keeps each wave packet coherent (together). You're just using it to correct for the group delay created by the distance of the opposite speaker to your ear, which has the same effect on phase.

Moving your head slightly doesn't kill the effect. You're correcting for distances of 2-6 feet, so 2 inches is peanuts. I'm not sure, but 2 inches might be smaller than the resolution of the adjustment.

I think one of the big benefits to the time correction is in getting the relative phases/delays correct between the tweeter and woofer, to get the listener on the right axis of the directional interference pattern. This is probably even more important for low-order crossovers than for high-order crossovers, now that I think of it.

I've experimented with time corrections for driver's side, passenger's side, center of the car, and no delays. I'll tell you that double-misaligned (passenger's side setting) sounds different, but no worse, than no delay, which sounds different, but no worse, than "half-misaligned" (center of the car setting). The only real improvement/degradation occurs when switching between tuned for the position you're sitting in, and anything else.

I don't know where you're located, but if you're in the SF Bay Area you are welcome to come do some critical listening.

 

[QUOTE=modifry,Oct 28 2005, 06:12 PM] If you're saying you can use a smaller amp on each driver, yes that is correct.

 

Phil, sorry if my post was misleading in any way.

i believe i stated something about going active when trying to maximize the setup. also, tuning a crossover is not that hard. reading any type of professional tutorial is more than enough for the average person to go out and try it themselves.

i think you are getting confused about tuning an EQ and tuning a crossover.

 

i didn't really think it was misleading... i just didn't agree.

while a crossover can be a very simple device to actually 'setup' (program, configure, etc.) i do believe that doing it properly isn't as easy as people might think. an EQ can be quite difficult as well, and getting both of them 'right' is not easy at all.

to get a crossover really set up properly, without creating a frequency hole or mound, while working with driver limitations, and trying to keep the soundstage correct -- well, it just isn't easy for the average amateur - that is all my point is... sure, they can go out and try it out, but will they really end up w/ something better? will they even make par? i'd bet most won't...

*shrug*

otherwise, i'm staying out of this thread.

 

I agree, higher orders lower your susceptibility to cross-over/driver proximity induced distortion. And I agree with you on driver palcement - I'm not thirilled about having high-mounted door tweeters in the newer models, as that's the last place I'd mount them if it were up to me. (mine are in the lower door, surface-mounted next to the woofer.

I've been doing some reading on time-alignment and can share some of it because it's on the Internet here . It looks like humans can't differentiate a time difference of less than 20ms (correlates to a distance of over 20 feet) so it's not really the time difference that you're correcting for. What we DO hear are changes in the sound field so that we can't correctly discern the spatial relationships in the music, plus there can be some frequency response effects of time misalignment. The spatial differences can evidently be extreme. The article talks about a test CD with time differences of only 1ms (about a foot). The difference is the listener either places the sound directly between the speakers (no delay) or fully at one side (with delay).

Unfortunately, most recordings don't have "natural" or "correct" spatial relationships anyway because they're recorded in a recording studio instead of on a sound stage. Positional cues don't exist, so any you hear are either manufactured or are a by-product of the mixing or reproduction sytem. In most cases, time-alignment to get the correct sound stage is only reproducing a fake sound stage. But if it sounds better (or just different and you like it) who cares?

Frequency response is more readily discererned by most listeners and my guess is that it's affects are more noticeable when they're corrected with time-alignment.

Overall the author's suggestion is to go ahead and do time correction if you can do it without mucking it up. But then he also says you can't have "true stereo" unless you maintain the 40-60

 

Have to let Orthonormal comment here because I don't know what I'm talking about, but I thought one benefit of time correcting individual drivers was that it would help align the phase of the tweeter relative to the woofer even though the path lengths are very different, especially on the side you are sitting on. This effect would not be subject to the human ear limitation on time delay. I think this is basically what you are saying, right?

I agree that the stock MY02+ driver locations are not ideal and that's what's causing most of the problem here. There are however reasons to use them (convenience, stock appearance for security etc.) and as someone pointed out earlier good imaging can be achieved for these locations. I think what John has done here is an example of making it work. But then again, it might work really well in his case because the stock tweeters might have very consistent off-axis response. We'll see after this weekend hopefully whether it has a similar effect on my system (same HU but CDT CL-61T components).

Also wanted to point out another advantage of the Alpine internal active xover- level control for each driver is +/- 12db, much greater range than the typical jumpers on the passive xovers of component systems.

Peter

 

[QUOTE=PoweredByCamry,Nov 1 2005, 07:04 PM]Have to let Orthonormal comment here because I don't know what I'm talking about, but I thought one benefit of time correcting individual drivers was that it would help align the phase of the tweeter relative to the woofer even though the path lengths are very different, especially on the side you are sitting on.

 

I think the key point is that the Alpine CDA-9833 is not one of those cheap systems. The capabilities of the head unit closely match the capabilities of a Burr Brown (TI) audio processor chip - 6 channels with built in digital filters and delays and independent volume controls.

I've been daydreaming about learning how to use I2C and how to design linear-phase FIR digital filters, then using that same BB processor chip as the core of an all-digital crossover and amp for my home stereo. I hadn't really thought of the time alignment when I first came up with the idea, but I wonder if it would help any with my home speakers. They're designed to be more or less time-aligned, but you could tune for various listening heights with the time alignment.

 

Well thanks very much to Orthonormal, who cut up his hands pretty good wiring my doors last weekend, I now have a similar active setup to his. We both run the Alpine 9833 HU in 3 way mode using the internal active xover and time correction. Difference is he is running stock speakers, powering the tweeters and mids with the F and R internal HU amps. I am running CDT CL-61 components, powering the tweeters off the HU and the mids off my Pioneer 50x2 amp. I also have an Infinity Basslink.

After listening to it for a few days, I am very happy with it. The time correction does the same wonders for my system as it did for Ortho's. An added bonus is that I can go to -12dB on the tweeter level, whereas the CDT passive xover had choices of -3, 0, and +3db and appeared to have no compensation at 0db. Before I think the general tweeter level was too high, and using EQ to attenuate certain frequencies just made the response lumpy. My system now sounds smoother and the imaging is much better.

Peter