Hi Markus,
I'm not totally following your scenario. And when you say "precedence breaks down pretty early" which way are you going? Making the center louder or the reflection?
I was assuming a louder signal, the wall bounce, and an earlier signal, the partially blocked direct sound. If the direct sound is very low (graphs show 12 to 15dB below the later) then the sound should localize at the outer reflection. As you raise the level of the direct sound (reduce blockage), at some point it would take over as the source of the sound. this is just the flip side of saying "the earlier sound wil dominate a later reflection even if the reflection is as much as 12 dB louder" (stating precedence effect in the usual order).
David S.
Wides were delayed by 10ms in both cases. Center = L+R, L=L, R=R.
Center 0dB, wides -5dB -> slight stereo effect and located towards height of wides
Center 0dB, wides 0dB -> room wide stereo effect, soundstage has depth
It's interesting that these numbers are nowhere near those found in the literature.
Took some measurements of Geddes Nathan (left speaker only) with different toe-in. The interesting part is how the reflection pattern changes in the first 15ms.
40° toe-in:
30° toe-in:
20° toe-in:
10° toe-in:
0° toe-in:
10° toe-out:
40° toe-in:
An externally hosted image should be here but it was not working when we last tested it.
30° toe-in:
An externally hosted image should be here but it was not working when we last tested it.
20° toe-in:
An externally hosted image should be here but it was not working when we last tested it.
10° toe-in:
An externally hosted image should be here but it was not working when we last tested it.
0° toe-in:
An externally hosted image should be here but it was not working when we last tested it.
10° toe-out:
An externally hosted image should be here but it was not working when we last tested it.
If you Google "Roger Russel" you'll find a great website put together by Roger, head of acoustic research and speaker design at McIntosh from 1968 - 92. He's recently developed a wideband columb speaker that he claims sounds incredible, and works with room acoustics more effectively than any other speaker design. It has 24 - 3.25 inch drivers per side, (vertical line array), tiny footprint, uses electronic EQ to be flat from 20HZ - 15kHZ. Bass is allegedly plentiful with no boominess, imaging is said to be second to none, no audible interference pattern due to multiple drivers, and they can be placed near walls so no divorce, very wide sweet spot, etc.. Check it out. The individual drivers appear to be available at Parts Connection and Madisound (who shows full specs) at $22 each (Peerless/Vifa TG9FD-10-04). If what he says about them is true, it could be the final answer to this threads question. I'm pretty tempted to build a pair.
I see three distinct peaks, what is the direction of each of these reflections?
This one?
My Experience with Column Systems
Almost extends to the ceiling, might be an infinite line-array then. Before building, check the work of Keele:
Don Keele's CBT (Constant Beamwidth Transducer) Page
Ceiling, ipsilateral wall (you can see it in the 10° toe-out graph), contralateral wall.
But neither is actually "constant directivity". CBT is only constant beam height, the horizontal is not constant but falls just like any piston source. IMO they should be placed on their sides, or better yet just make a constant beam width in both directions! Now there is an idea!
Why we would even want to have curvature in the wavefront ?? Any corvature is indication of source vicinity. All the sound is like coming from a single point. Wavefront curvature is one indication of localisation of the speakers! This is not what we want for stereo ! Blumlein stereo assumes plane waves. His coincident figure eight microphone is not sampling the wavefront curvature at the recording venue, why we would like to generate erroneous curvature at the reproduction end ?!? This is not an accurate method to reproduce what has been recorded.
- Elias
- Elias
Would you please elaborate this? Do you have research indicating that we can hear the curvature of the wavefront? Would be interesting, but I'm a little skeptical right now
It wasn't just that one typo I was going by, I was also looking at the other comment I quoted (about the contra-lateral reflection occurring first) as well as previous comments you've made in many different posts.
The confusion is that you often haven't made clear that when you talk about avoiding reflections in the first 10ms you mean minimizing reflections in this time region as much as possible in amplitude through directivity, not pushing them out of the 10ms window entirely via sufficient physical time delay of the reflections.
They are two completely different scenarios, and in practice directivity will only sufficiently reduce the early reflections at higher frequencies.
As I mentioned, Linkwitz recommends 6ms to first wall reflection for best results (actual minimum time delay, not just reflections within this time window greatly attenuated) and many people through the course of this and other threads have taken his 6ms recommendation and your 10ms one and rolled them together as if they're referring to the same criteria, when in fact they're not.
Call me a pedant if you will, but I think the distinction is very important.
Anyway lets not quibble about it, as I'm actually on your side of the argument in most regards - I'm of the belief that in many (most?) rooms it's impossible to get 10ms to the first side wall reflection, and highly impractical to achieve 6ms, so directivity (perhaps along with some natural diffusion to avoid an early specular reflection) is the best option to get those early reflections (especially ipsilateral) down in amplitude to the point where they're in amongst the overall reverberation level in the room. (and don't stand out obviously in the ETC curve)
Last edited:
You are confusing curvature of the line with curvature of a wavefront. The line recedes towards the top to give delay to those elements. In addition there is level tapering (reduced drive) to the elements at the end of the column. This is an evolution of steered or tapered line arrays as have been used by others in acoustics and antenna design for many years. The end result is well controlled and high vertical directivity. The meausrements show good near and farfield response that is well maintained over height. The radiation pattern is approximately cylindrical.
The end result is nothing like an omni on the floor (do we have to endlessly hear about that?)
As to Blumelein demanding flat wavefronts, where does that come from? If I want a violin or guitar in my living room, wouldn't they be creating a circular wavefront? Don't most instruments create circular wavefronts, at least at low frequencies? (That approximate to flat as the distance increases.)
David S.
Of course he's going to claim that, he's selling them.
Ask someone who's heard them, there's a good chance their opinion will vary.
Is there any specific point you like to make?
Do you have research indicating that we can hear the curvature of the wavefront? Would be interesting, but I'm a little skeptical right now
Who say we hear the curvature ? Curvature indicates the velocity vector of the wavefront, the travelling direction. At least the last 100+ years have been spent to understand how human perceive sound directions in spatial domain
Example: Sit down in front of a speaker emitting curvature wavefront and the sound is coming from directly front of you, then stand up and speaker is perceived 1m lower. When I'm in a concert and stand up from my chair I don't perceive the band falling down And P.S. we do hear curvature in the nearfield as a distance cue.
Check the phase isobars ! They are not vertical straight lines but indicate the curvature of the wavefront.
The directivity can be well controlled because the wavefront has curvature
What if you like to reproduce distant source, no curvature allowed ? Anyway, since no curvature information is possible in stereo recording, it is not accurate method to generate false curvature at the reproduction end.
- Elias
Yes, they are completly different, but I would have thought my position obvious since I am always talking about "small rooms" in homes where for "code" reasons are never big enough to get anywhere near allowing for a 10 ms window. It's small rooms that are the challenge, it's a "no brainer" if you can move the walls out of the way!
Directivity can, of course only work as low as you can actually achieve any directivity - kind of obvious. Fortunatly it is reasonable to get decent directivity down into the 500-700 Hz region, while below that out localization ability is failing fairly fast. We dominately localize a complex source based on queues above 500 Hz. So you can get a fairly long ways with readily achievable directivity - most of the critical frequency range.
I'm with you Dave. All finite sources create spherical wavefronts at a great enough distance (and its not that great actually). The argument about perceiving the location of a curved wavefront is bogus. I like to think of my speakers as a point source with directivity because thats what they are. They completly disappear sonically in the room.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.