The Advantages of Floor Coupled Up-Firing Speakers

[markus76]

By the way, James Croft proposed two solutions to the "sound from the floor" problem in the thread I had linked earlier. Please see http://www.faktiskt.se/modules.php?n...110176#1110176

I don't think the HRTF-approach works reliably but the Allison-like approach of placing the woofer/mid boundary-coupled and the tweeter at ear height looks promising for a wide dispersion design.

[graaf]

Do You read the thread before posting Markus?

Everything Mr Croft has to say on the matter have been already posted here.

[graaf]

Originally Posted by Elias

There are many systems employing 3D stereo field. Ambisonic comes to mind first Others include 3D vector amplitude panning, like the one proposed by Lokki.

and what about the precedence effect?

Originally Posted by Elias

I don't understand the point. Sound direction can be localised without any reflections. In Snell Type 1 there is strong direct sound from the tweeter, which can result to image localised low.

I thought we have agreed that what is needed is optimal direct/reflected sound ratio in "Type 1" there was strong direct sound but also strong reflected sound

and the sound was NOT coming from the floor

Originally Posted by Elias

Yes there was a big difference in the treble range. I posted measurements long ago in some of these threads.

treble range that is precisely? how big? I ask because it seems to be important for the FCUFC to work best

[graaf]

Originally Posted by Elias

Is your tweeter a horn tweeter ? How is the directivity ?

it uses Dayton H12RW 12" waveguide IIRC so it´s effectively a virtual ceiling tweeter, an entirely different FCUFC implementation than "Type 1"

[Elias]

Originally Posted by graaf

and what about the precedence effect?

I thought we have agreed that what is needed is optimal direct/reflected sound ratio in "Type 1" there was strong direct sound but also strong reflected sound

As a rule of thumb, reflected and delayed sound must have a little bit higher amplitude than the direct sound in order the direct sound not to be perceived as a sound source.

In snell type 1 I cannot see how this would happen ? And, there is no vertical directivity pattern plots of Snell speaker ?

Originally Posted by graaf

treble range that is precisely? how big? I ask because it seems to be important for the FCUFC to work best

The speaker I was using for the test started to have narrowing directivity somewhere above about 3-4 kHz I think.

[markus76]

Originally Posted by Elias

As a rule of thumb, reflected and delayed sound must have a little bit higher amplitude than the direct sound in order the direct sound not to be perceived as a sound source.

In snell type 1 I cannot see how this would happen ?

If the dispersion is really wide then there are higher order reflections that are likely lumped together (perceptually). Would be interesting if someone could take the time to look at this more thoroughly.

[gedlee]

Originally Posted by Elias

As a rule of thumb, reflected and delayed sound must have a little bit higher amplitude than the direct sound in order the direct sound not to be perceived as a sound source.

Originally Posted by markus76

If the dispersion is really wide then there are higher order reflections that are likely lumped together (perceptually). Would be interesting if someone could take the time to look at this more thoroughly.

Elias - you have to be careful quoting those results because, as I have pointed out before, they refer to a shift in the "principle source location" and say nothing about image blur or coloration - the two things that I consider the most important. The precedence effect is widely over used in audio to explain things. Most of the work that was done simply does not apply to the real problems in audio - imaging, coloration and spaciousness.

[ScottG]

Originally Posted by gedlee

Elias - you have to be careful quoting those results because, as I have pointed out before, they refer to a shift in the "principle source location" and say nothing about image blur or coloration - the two things that I consider the most important.

Localization is exactly what he is describing (it's not just implied, it's expressed):

Originally Posted by Elias

As a rule of thumb, reflected and delayed sound must have a little bit higher amplitude than the direct sound in order the direct sound not to be perceived as a sound source.

"Image Blur" and "Coloration" from reflections are however interesting topics.

"Image Blur" I'm aware of - a diffused effect of an "image". Which I rarely find to be far from reality (..when compared to a real sound source with eyes closed listening). In fact sometimes the result seems to be closer to reality, but occasionally worse depending on how close the loudspeaker is to the wall and how reflective the wall is.

As for "Coloration" though, at least at higher freq.s.. I'm somewhat at a loss here for a "definition".

I have heard conditions where there were broad-band changes (like a "tilted-up" freq. response) - but not something I'd characterize as a "coloration". Moreover most people seem to adapt to the alteration from "flat".

The only time I've heard of significant coloration at higher freq.s was from direct sound (i.e. the combination and combing effects of L+R, particularly on vocals), not from any reflection. (..of course modal effects and floor/ceiling "bounce" will almost certainly provide a "narrow-band" coloration, but that's at lower freq.s.)

[speaker dave]

Originally Posted by graaf

I thought we have agreed that what is needed is optimal direct/reflected sound ratio in "Type 1" there was strong direct sound but also strong reflected sound

and the sound was NOT coming from the floor

The goal with the Type 1 is to have the reflected sound in phase with the direct. Its no different than ground plane measurements. As you take an ear level mic and move it towards the ground the air path and the ground bounce path get closer in time. The cancellation nulls move up in frequency and, eventually, all audio frequencies are in phase. The direct and reflected become inseparable.

The woofer can be above the floor surface a few inches, but the tweeter can't, hence the ramp and the notch to get the tweeter center at the reflecting plane.

On top of it all you get a bonus +6dB.

David S

[CLS]

Originally Posted by graaf

it uses Dayton H12RW 12" waveguide IIRC so it´s effectively a virtual ceiling tweeter, an entirely different FCUFC implementation than "Type 1"

Yes, correct, that's what I'm using. Thank you.

BTW, I don't know why the diffused HF can sound pretty clear. For example, the triangle at the far corner of a big orchestra is sounding just the way it should -- distant, small, crisp and clear. Both spaciousness and imaging are very good.

Flooder tweeters aside, the directivity of the central channel DML panel is very wide, too. The panel is quite large (180x90cm), with a very uniform half-space radiation.

So basically the HF in this (very reflective) room is everywhere. Why can I pinpoint that little tiangle in the whole (sound-)space? I can't explain.