The Advantages of Floor Coupled Up-Firing Speakers

I'd ilke to see real measurements and not just pretty marketing drawings. When Stereophile measured the LS50 it looked like this:

1212KEF50fig4.jpg
 
...
When listening, we receive two independent sets of direct/reverberant energy information:

One from the listening room and one from the recording.

The set from the listening room places the "window" (as mentioned above) at its appropriate

distance in our room. The set from the recording places the instruments/voices and hall

dimensions behind that "window".

We must make a clear difference between both sets. They don't mix, if we listen

to a large hall recording in our small place. They will mix, if we listen to an intimate

near field recording of a small ensemble. That's because the direct/reverberant informations are

vastly different in the first case and become comparable in the second case. This is one aspect

of the "we are there" against "they are here" controversity.



Rudolf,

hopefully you will not take my quotes & answers as pure "nagging".

It is just that your post is inspiring ...


1) You talk about "independent sets of revererant energy" from

a) the recording (venue) and
b) the listening room

You talk about making a difference between a) and b) and there
being a clear difference in that

- a) and b) "will mix" (under certain conditions to be defined)

and

- a) and b) "don't mix" (under certain conditions to be defined)

In my mind i'd like to imagine a) as a recording from within e.g. a concert hall using some
minimalistic microphone configuration and no artificial reverb. Even though that might be
nonrealistic ... it is just for my inner imagination.

And i like to imagine b) as a (hard walled rectangular shoebox) small room acoustics listening
room whithout any siginificant treatment. Which is the listening room standard for most of even
those humans on the planet, which do not suffer from completely different problems, than how to
treat their listening room.

This said i like to make up my antithesis using an example:

Imagine a human reciter reading some simple text in said concert hall and being
recorded by said simple microphone configuration.

We now place an artificial speech recognition system (ASR) in the listening room
and feed it with

a) the speech recording from the concert hall directly.

b) a microphone (say cardioid like or even MS stereophonic) signal, obtained
by reproducing the speech recording in said listening room by some kind of
state of the art loudspeaker.

We now measure the error rate for a) and b).

I hold any bet, that the error rate of setting b) be will be higher.

Antithesis 1):
1.1) There is always a convolution between the recording venue's reverb and the
listening room's reverb on reproduction side.

1.2) Separating the two sets of reverb (sets of reflections making up the reverb),
is a MENTAL ACTIVITY (source separation).

1.3) Listening room's reverb is always degrading to any system's performance
(say Artificial Speech Recognizer, ASR) with no capability to perform this
separation activity.


Some personal conclusions:

Whether speaker/room system's influence on reproduction of a certain recording
is estimated as "separable" or "mixed" is based on the mental capabilities of
the listener.

In terms of perceived reproduction quality the speaker/room system's influence
can basically be estimated as

- degrading
- neutral or even
- enhancing

due to the recording or even "the original event".

From a room acoustics point of view, convolving the response of a concert hall
- hopefully having s u b s o n i c Schroeder Frequency FS - with the response of
a small room having FS=120 is clearly a degradation.


Why is this? The original recording is virtually free from any room modes
to be perceived, because the modal overlap factor (MOF) is >3 in the whole audible
bass range (and even much better at higher frequencies).

That means in every room mode's bandwidth we will find another 2 modes at minimum,
which will make room modes non audible as such and thus we say the room to have
"statistical" properties in terms of bass response.

It may be that due to some other measures of reverberation the response of the
concert hall and that of the listening room "do not mix", e.g. because the decay
of the concert hall is so much longer than than of the listening room.

But still those aspects of reverb "do not mix" (if we assume they really don't ...)
only due to mental activity involved: The mental activity does the separation.

> "DO NOT MIX subjectively" does not necessarily mean "no degradation"

A different example:

We have a recording from a small room, say FS=100Hz and we reproduce it in our (also)
small listening room with FS=120.

Say somehow we are lucky that the modes captured on the recording have significantly
different frequencies than the listening room's modes excited and perceived at the
listening seat.

Then we might even experience some enhancement of the recording, because MOF seems
to be increased in comparison to the original event.

> "DO MIX subjectively" does not necessarily mean "degradation"


On the other hand, in case of accidental coinciding frequencies of modal excitation
at the recording venue and in the speaker/listening-room system, the result of
reproduction will be clearly a degradation.

Of course bass reproduction is only one aspect.

But "MIX" or "DO NOT MIX" are IMO no helpful categories.

Reproduction by a speaker/listening-room system is potentially (probably)
degrading, no matter whether the recording venue has small room acoustics
or not.

The result of interaction between a certain recording and a certain
speaker/listening-room system might just be more unpredictable, in case of the
recording has captured typical small room acoustics artefacts.

(A thing to remember IMO, especially when talking "listening experience"
in forums.)

This is because most speaker room systems are likely to produce typical
small room acoustics artefacts too (see above).
 
Last edited:
Is the LS50 the driver recommended by graaf? Or is the LS50 the driver with the "promotional" polars shown by graaf? Or even both?

I think I've seen these polars before in a KEF Uni-Q document.

Regarding his "recommendations": Has he ever shown anything that would allow us to verify his claims? I have reason to believe the speaker and especially its "magic" properties exist only in his mind.
 
hi graaf - - my apologies for looking at the end of the thread rather than the first.

if I am seeing correctly (?) your FE206 enclosure is sealed, ~0.13 cubic feet, qtc 0.5->.7 (depending upon internal damping material), and top panel about 3.5" thick and extremely inert and rigid.

yes, it was my first attempt at FCUFS


did you apply eq or did the room position of the enclosure and up-firing nature with 206's rising on-axis response set the tonal balance?

http://www.diyaudio.com/forums/atta...floor-coupled-up-firing-speakers-p6255901.jpg

the balance was ok straight so to speak (even the bass, it was quite kicking, to my big surprise) but yes I did try equalization with DEQ2496, I didn't like the sound of DEQ's DACs though

main problem with FE206 were spurious resonances in the cone, I tried asymmetrical cone treatments and I got promising results

see:http://www.diyaudio.com/forums/full-range/126234-8-inches-paper-cone-treatments.html

but pretty soon afterwards the speakers have been destroyed by ...a kitten ...oh don't ask how this happened... :eek: :p


regarding "floater" with midrange and tweeter section angled upwards, is there a particular angle of tilt which is preferred?

difficult question, I didn't investigate this, AFAIK user Elias did
 
IMO you are better off, making a speaker in such a way that it does not confuse lateralization
cues detected on a recording by the listener relying mostly on low to mid freqauencies < 2Khz.

This was also my leading thought at the time I was actively using my dipole line arrays. High directivity at low and midrange freqs.



There is no evidence to me, that a uniform directivity pattern is necessary over the entire audible bandwith.

I think so too.
 
Since fullrange coherent loudspeakers (especially conventional CD concepts like
e.g. proposed by SL above) tend to cause reflections strongly correlated to the
direct sound in usual (non extensively treated) listening rooms, those
"reproduction induced ITDG cues" are so very detrimental to imageing, because
they strongly c o m p e t e with ITDG cues on the recording.

Instruments and voices in a real musical venue never compete this way: They keep
separable by producing individual sets of ITDG cues and also patterns of later
reflections separable in arrival time and direction.

The conventional loudspeaker (fullrange coherent and also optionally CD) is a
bad musical instrument in that respect: It competes with all the instruments and
voices to be reproduced by causing strong and uniform ITDG cues combined even with
uniform later reflections strongly correlated with the direct sound over the whole
audible bandwidth.

Such loudspeakers force the user more than it is desirable to build a reflection
free zone around the speakers in order to have long ITDG in the listening room.

What is left from strong reflections arriving at the listener has to be treated
with diffusers and absorbers, which is not a practical way to go consequently
for most listeners.


So you are suggesting a method to make reflections non correlated from the direct sound, am I wrong ? :)

You got bending waves, do you still use them ?

Have you got any data to show the correlation between the on and off axis signals ? This is what we would need to see if it is practical.
 
Stereo speakers with very high directivity (high direct/reverberant ratio) will place the window directly at our head. The listening room influence is highly subdued and we may have the front soloist playing/singing at nose distance.

In my experience I have found this to be true especially at high freqs above 1kHz (about).

It means high directivity speakers at high freqs, including horns or 'waveguides', cannot be used if any realistic sound reproduction is desired. Where in reality the sound event occurs at your nose?

What we need at high freqs is low D / R ratio. It enables sounds to be reproduced in a small room in a realistic manner.
 
IMO you are better off, making a speaker in such a way that it does not confuse lateralization cues detected on a recording by the listener relying mostly on low to mid freqauencies < 2Khz.
This was also my leading thought at the time I was actively using my dipole line arrays. High directivity at low and midrange freqs.

There is no evidence to me, that a uniform directivity pattern is necessary over the entire audible bandwith.
I think so too.
A strictly uniform directivity pattern over the entire audible bandwith is technically impossible - even with dipoles ;). But don't we want to keep the high frequency portion of the reflections as high/constant as possible? :confused:
 
But don't we want to keep the high frequency portion of the reflections as high/constant as possible? :confused:

Some do and some don't. But even those who do, they tend to 'forget' that most typical rooms do not support constant reflective energy at high freqs.

If true constant energy reflection is a goal, then speaker off axis energy must increase with frequency and not be constant as is claimed.
 
It means high directivity speakers at high freqs, including horns or 'waveguides', cannot be used if any realistic sound reproduction is desired. Where in reality the sound event occurs at your nose?
There is no realistic sound reproduction in stereo, because stereo in itself is highly artificial. You can only argue about the fidelity to the sound heard at the engineering/mixing/production facility.
The waveguides I know of don't have a directivity index which is much different from dipoles. The main difference is in the angular increase of directivity - progressive in dipoles, more linear in waveguides. So you are saying that dipoles are no good for sound reproduction too :confused:.
What we need at high freqs is low D / R ratio. It enables sounds to be reproduced in a small room in a realistic manner.
How do you do that? Adding supertweeters directed to the back of the room?

Rudolf