CTC and the CEA2034 spec

[augerpro]

The inventor of VituixCad had started a discussion around what the optimum center-to-center spacing vs crossover frequency was. The claim was better vertical reflection responses as defined by the CEA2034 spec can be had at 1.2-1.4x ctc. I had my doubts, but did a quick idealized sim in Soundeasy, and when you look at the polar plot that show the lobe shape, I was surprised that while the main lobe did get narrower, the new top and bottom lobes moved inward rather fast. This actually reduced the null slightly, and made the response smoother and more like the on-axis response. Intrigued, I built a similar idealized project in VituixCAD so I could directly see the effect on the CEA2034 vertical reflections. This dovetailed with an idea I've been toying with for a few years which is somewhat more controlled vertical directivity, because I believe the typical drywall 8" ceiling is real detriment to sound in most homes. You can see the improvement in the plot below, followed by a "typical" crossvover and ctc.

I had a prototype Klippeled to verify this performance and verify the ABEC diffraction model. Compared to the norm, I think the results are pretty outstanding. Only -20 degrees was really bad. I messed with delays in crossover CAD to mimic closer ctc and while it helped the -20 a bit, it mostly hurt the others.

So clearly there is benefit to the vertical reflections, and horizontal is really not affected. These would score very well for listener preference. But I'm just not confident if I should move forward. The main issue will be a narrower vertical sweet spot. But when is that an issue? Has anyone actually heard a real problem caused by this? I do note, that some of the better reviewed speakers like March Audio, D&D, Buchardt have a similar tightish sweet spot that I expect would be similar to mine if I moved ahead.

 

[augerpro]

I should note, the reason the responses roll off so fast >4khz is that I had to use a 4" driver as a stand-in for my waveguided tweeter. I'm only looking at the crossover area, so just ignore >4khz.

 

[digitalfrost]

So you're saying 1.2-1.4x center-to-center from mid to woofer, in relation to the crossover frequency? Does the topology play a role? It should right.

 

[augerpro]

Mid to tweeter. Yes filter type plays a role. LR2 often do real well. But LR4 can be made to work well enough - as shown above.

 

[phrog]

Are you saying that the wavelength at the crossover frequency is 1.2-1.4x the ctc spacing? What ctc spacing are you using in that vcad sim?

 

[augerpro]

Correct. My spacing is just under 8" I think with a 2.4khz crossover. About 1.4x crossover wavelength.

 

[tmuikku]

Hi, cool You could measure your listening height and evaluate how much ear height varies. My listening spot is on a couch, sometimes leaning forward, sometimes tits up etc Ear height varies some, perhaps 40cm or so. This is about 8 degree window at 3 meter listening distance or 11 degrees at 2 meters. So, I think even narrow main lobe can cater seated listening fine and its not too much of an issue as long as listening height is taken account, tweeter at ear height or something similar depending on your design.

Standing up, perhaps important? I find vertical response not too important for typical home back ground listening while doing chores standing up situation. Perhaps if critical listening was done standing up it would matter and ahould be taken account somehow.

The thing is that best stereo sound will be only within rather small area and critical listening happens there, also vertical response needs to be best there. Actually, in order to make best possible situation for critical listening my opinion is you can and even should make some sacrifices if it makes great sound where its most needed by cost of slightly worse somewhere that is not so important.

If you can make quick single speaker mono prototype with freestanding tweeter, you could use any speaker you have as bass box for it, sideways, upside down. Use lego or scrap wood as spacer and play around with the spacing. I havent tested this yet with mid and tweeter but bass and mid sounded weird with too much c-c, got kind of streched height to image.

ps.
While at VituixCAD do some ideal / real 3rd order butterworth slopes, asymmetric response whose window is really small, response is changing by degree. Also, hearing the top octave might already dictate quite narrow vertical listening window so, not much harm narrow window on xo I think, especially if it makes sound better

 

[augerpro]

I always design floorstanders to have the design axis at about 40", which is a common ear height at the listening position. So no problem there. But standmounts require the listener to properly set them up, and I worry if I design something and someone builds and doesn't like it - but never set them up at the right height. Is it their fault? Or mine for not making a more flexible design?

So I guess that is why I'm wondering if anyone has had a real world issue with wider ctc?

 

[Scgorg]

Whether or not this is worth doing seems to be reliant on a plethora of factors, the main one being listening distance. If you're sitting far away from your speaker, even a relatively small front-firing lobe will allow you ample height adjustment between slouching and sitting upright, and the greater impact of reflected sound may benefit a lot from a smoother SPDI.

If you're sitting on a desk less than a meter away from your speaker, then having a very small front lobe might be a detriment, as it constricts your movement enormously. In this case taking the hit to SPDI is probably well worth it. The same goes if you're making a speaker that is going into an area where there will be a lot of movement in both vertical and horisontal directions (e.g. at some kind of venue), though a coaxial is obviously optimal here.

I am currently in the process of designing a speaker, and have elected to go for a pretty large CTC spacing, as that seemed the best compromise for my use case.

 

[McFly]

My findings, both objective and subjective, seem to favor this theory of 1.2x XO CTC spacing for conventional multiway loudspeakers, at least for transition from mid to high frequency drivers.