Most modern loudspeakers have a problem with resolution at low level. That is because of complicated frequency dividing networks and high damping in the drivers in a desire to make the frequency response as flat as possible. Many of this speakers use soft plastic cones and surrounds that have high loss. Most of this constructions have also major acoustic phase and radiation pattern problems.
The speakers i have shown here do not have that problem and actually listening to a high resolution system at low volumes opens up the ears after some time and satisfaction can be high. It is like the eyes adapting after a while to darkness.
The speakers i have shown here do not have that problem and actually listening to a high resolution system at low volumes opens up the ears after some time and satisfaction can be high. It is like the eyes adapting after a while to darkness.
I disagree. The reason some speakers need high SPL levels to come alive is a direct result of the Fletcher-Munson curve. If a speaker is voiced at 100dB, a level required by rock listeners, it will sound weak and tinny at 80dB. Long ago amplifiers had a loudness control to correct for this effect. Now tone controls of any kind are banned from "audiophile" amps and preamps under the misguided idea that anything between the amp and speaker except a straight wire corrupts the music. This, of course, disregards all of the balancing and other electronic mischief that occurs between the artist's microphone and the recorded medium.
When I voice a speaker, I apply extra BSC and may or may not accentuate the treble so that is sounds "right" at 70-80dB @2.5m. If played at 100dB, really not possible with my speakers, but I digress, my speakers will sound bass heavy and bright. Because of this , I have several loudness curves in my DSP EQ library to make the music right at whatever level I choose at the moment.
Bob
yes indeed, "flat frequency response" as engineering ideal, as a goal in itself, is the principal cause of the problem
yes indeed, and at the same time our hearing adapts especially easy to minor deviations from "flat frequency response"
on the other hand it cannot adapt to time distortions, and in case of real sound sources (which are not purely periodical) our hearing works in time domain, therefore time domain fidelity (preservation of original transient and waveform shapes) is most important from perspective of sound synthesis, sound source recognition as well as HiFi sound reproduction:
‹Z�p‚Æ—�˜_�\ TIMEDOMAIN (�S‚̃I�[ƒfƒBƒI �`Ž©‘R‚ȉ¹‚̃Xƒs�[ƒJ�[ )
and therefore the single way is the only way (plus perhaps some active crossovers that allegedly preserve time structure of sound:
http://www.resolutionmag.com/pdfs/DRAGONS/CROSSO~1.PDF )
best,
graaf
ps.
although our hearing can adapt to non-"flat frequency response", this flatness is also important but can be achieved with appropriate equalizer taking into accounts loudspeaker-room interface
designing multiway loudspeakers for flat response under anechoic conditions (or in specific model room environment) results in loudspeakers that are at the same time non-flat in most normal listening rooms and incapable of any realism because of inevitable time distortions
two words - worst case
Last edited:
I work with linear phase wideband drivers since the 70th ( Manger , Jordan, Lowther, Quad
etc. ) and presented a digital solution in 1993 with the University of Essex. The reaction from the marketplace was nearly zerro. The only phase linear speaker i sold well was the MEDEA. We made 85 pairs of that crazy expensive ( 40.000,-DM, that was a fortune in 1991) speaker but i do not think because it was phase linear but because the press got crazy. It is a mystery to me that this issue does not get more recocnition.
etc. ) and presented a digital solution in 1993 with the University of Essex. The reaction from the marketplace was nearly zerro. The only phase linear speaker i sold well was the MEDEA. We made 85 pairs of that crazy expensive ( 40.000,-DM, that was a fortune in 1991) speaker but i do not think because it was phase linear but because the press got crazy. It is a mystery to me that this issue does not get more recocnition.
Do you have any qualified evidence that your achieved level of phase linearity was audibly better than a good speaker of half that price?
Rudolf
This is an endless discussion. We made double blind tests in the 90th and the result was typical for that kind of test. Under "music conditions" a mixed group of listeners could not tell if the system was linear phase or not. We had a speaker build in cooperation with the University of Essex that could be switched to linear phase or L/R 4th order phase response with the same on and off axis frequency response. Under informal circumstances some listeners had a preference for the linear phase variety and reported better 3 dimensional imaging. When we used artificial signals like a short, sharp Dirac impulse, the puls sounded " click" on the linear phase speaker and "clock"
on the L/R4, so the impulse had a perceptable longer duration. Daniela Manger sugested to repeat the test with listeners sitting off axis because she argued that the "noise" generated by the bended phase cancels in the midle position but this has never being done.
I whould like to add that a conventional multi miked, multi tracked and mixed down recording is not a good candidate to conduct listening tests that prove the superiority of linear phase speakers because phase in those recordings is highly randomised anyway.
on the L/R4, so the impulse had a perceptable longer duration. Daniela Manger sugested to repeat the test with listeners sitting off axis because she argued that the "noise" generated by the bended phase cancels in the midle position but this has never being done.
I whould like to add that a conventional multi miked, multi tracked and mixed down recording is not a good candidate to conduct listening tests that prove the superiority of linear phase speakers because phase in those recordings is highly randomised anyway.
Well said. Many designs look to flatten FR to the nth degree and forget that the difference between a good speaker & a great speaker is 40-some dB down from the main stuff.
dave
but was it really time coherent? I mean no transient split/deformation which can be seen as one single clear step on step response measurement? and not only on-axis but in every direction , so that to make inevitable early reflections coherent with the direct sound?
isn't it that time response (as can be seen in step response measurement) off axis of a multiway loudspeaker significantly differs from its time response on axis?
best,
graaf
Only a theoretical pointsource with response from DC to light can do that graaf. And it must be a minimum phase system so that the Hilbert Transform of the Frequency Response gives the Phase Response, no Allpass Term allowed. This is not posible and not even neccesarry for realistic reproduction. Our system at Essex was coherent over the defined bandwidth. An amplifer with the same frequency response as our speaker whould have the same phase response so phase distortion in that system came only from the limitted bandwidth. We call that "Kausal" in German. It had a very well defined off axis response too with a very low crossover frequency, only defined by the L/R 4th crossover ( chosen because of well defined off axis response and no pre ringing, we avoided higher order because the preringing of that filters destroyes the off axis performance) and the physical dimensions of the drivers and mechanical layout. The sound from that system was also confined by the 16Bit 44kHz sampling we used.
The impulse response of that system was an accurate copy of the B&K measurement system used.
The impulse response of that system was an accurate copy of the B&K measurement system used.
Thanks for setting that straight. There has been a lot of research with single frequencies, isolated "clicks" and "faked" reflections to learn more about the human ear and brain. While all those experiments have helped to advance our knowledge of the hearing process for sure, only very few of them were conducted under complex "music" conditions. Sadly there are too many audiophooles around who stretch and exaggerate the relevance of each and every of those findings to promote their own preconception.
I very much like your comment about modern recordings, which are full of phase issues from the start. They have to be that way to sound "right" IMHO.
Rudolf
A good sounding speaker is always a compromise. It´s not only the phase response and the radiation pattern but also distortion, sensitivity, mechanical resonance etc. But to whom am i talking here ? You are all expets in your own way.
P.S. : A loudspeaker is a very personal thing. What i like may not float your boat. So choose your compromises carefully.
P.S. : A loudspeaker is a very personal thing. What i like may not float your boat. So choose your compromises carefully.
I am speaking of sufficient time coherency of the direct and first-reflected sound, is really a pointsource with response from DC to light required for that? I doubt
for me it is the same question all the time - why are single way speakers preferred by some people? in other words - why are we discussing here at "full range" section and not at the "multi way"?
what is wrong with multi way and better with full range?
so I would like to make my statements more precise - not in every direction but just in the direction of first reflections probably would be enough
and not over the whole range "dc to light" because I believe that in most musical sounds there is too little energy in high frequencies above ca 5 kHz for a phase shift to cause any significant - audible and objectionable - wavefront shape change
"zero phase in whole frequency range" is not needed, we just need as much time coherency as is neccesary to avoid audible deviations from the time line of reproduced natural sound source (transient and waveform shape deformations)
it's been said that the main attribute of "full range drivers" is that the harmonic structure (the tonal aspect of the sound) is consistent from low to high
I say that in the end harmonic structure is about waveform shape
and what is important is time coherency which for me is preserving the time line of reproduced natural sound source in the direct sound and in early reflections
and a good full range single driver IS practically time coherent
but on it's precise axis only?
yes, loudspeaker is a very personal thing
but I cannot see any reason for having in case of HiFi sound reproduction objective design criteria which differ from objectve criteria used in case of sound synthesis and sound source recognition - after all it is all about sound source recognition - taking sythesised sound as less or more real, taking reproduced sound as less or more real
take a look at this "Sound-Source Recognition: A Theory and Computational Model": http://sound.media.mit.edu/Papers/kdm-phdthesis.pdf
here we have literally an (indirect) answer to the question: what makes a loudspeaker a HiFi loudspeaker?
there cannot be HiFi quality - in the sense of creating realistic experience (when the listener hears virtual sound source and thinks it is real) - when the loudspeakers err in time domain - the transient is partitioned and waveform deformed
as it's been said: "Playbacked sound has very unique tone that everybody can recognize as "playbacked sound"."
yes, this is what multi way speakers give us - better or worse but fundamentally the same "playbacked sound"
IMHO the problem is that "HiFi" loudspeaker design is fundamentally flawed, it is unscientific
it is based on false generalizations inherited from nineteenth century science that hearing only "analyzes sound waves in terms of sinusoids - a Fourier spectrum"
I think that the real problem is that audiophile expects and wants just better (in one way or the other) "playbacked sound" and she/he gat used to it so much that se/he likes it more than live sound...
therefore "the reaction from the marketplace" to phase linear speakers is "nearly zerro" as was also Your experience
best,
graaf
Last edited:
Thats what i say, you are the expert. One ceveat : not all wideband drivers are minimum phase.
Many drivers i measured that have a wither cone have a terible hole in the phase and frequency response at the frequency where the wither sets in. Many of them may also have good sensitivity but an extremely ragged response at other parts of the curve too.
A bad tonal ballance ( i am particular unhappy with hollowness, nasality and spittyness) is as anoying as a bad time domain behavious, maybe even more so.
Many drivers i measured that have a wither cone have a terible hole in the phase and frequency response at the frequency where the wither sets in. Many of them may also have good sensitivity but an extremely ragged response at other parts of the curve too.
A bad tonal ballance ( i am particular unhappy with hollowness, nasality and spittyness) is as anoying as a bad time domain behavious, maybe even more so.
You are the expert, I am just a frustrated music lover...
yes indeed but cannot we fix it with a parametric equalizer?
regarding resonant problems and in particular modes of the cone - what do You think of solutions like "differential wave impedance" cones or oblique cones - dispersing the resonance instead of taming it?
Have You ever tried anything like that? Would You try it now as You are now working with full/wide rangers?
below are: Lowther driver modified by Stewart Hegeman, conventional vs oblique aluminium cone response from JVC's brochure and Baker Reproducers 12 inches woofer (photo from Gramophone, March 1966, p. 110) - at Baker they coined the term "Differential Wave Impedance": "The patented cone construction has a central metallic ellipse and is said to give crisp response to transients with absence of coloration", "it has a special double ellipse of metal at the cone centre to eliminate cone break up."
best,
graaf
Attachments
Yes, that has all being done and more by me and others. We had luck to find a cone that is well behaved until 12kHz. Scanned with a vibrometer we could of cause find bending modes in that driver but coincident frequency is high enough to not cause distortion. Elastic flexing is simply too small in amplitude in that design to cause hystheresis. A driver that is linear to say 20kHz and has a very good off axis response has still to be invented. A friend of mine works on an electrostat that can do that and has up to 110dB clean soundpreasure but again this ain´t no cheep.
You mean linear in the sense of linear frequency response or in the sense of low non-linear distortions?
cannot the first be equalized? as to the second - I don't know what to think about it - isn't the second harmonic of 10 kHz - 20 kHz and the third - 30 kHz?
how critical is driver's performance over 10 kHz really? there are just the higher and the highest harmonics of musical instruments and voices
very good off axis response - that is?
question is - what is really needed and why?
Do You know my idea of the so called flooder?
It uses a coincident driver or just a fullranger (fom 5 to 8 inches).
It's off axis performance is nowhere near very good, the listener listens to it completely off axis and yet this imperfect off axis performance seems to be inaudible as such
best,
graaf
Last edited:
I mean frequency response.
Yes, you could try to build something like that and then equalize.
The Scan Speak wideband i am using is a good candidate for that.
Off axis distortion is often lower to my surprise.
I could build something acording to your plan and listen.
Do you know Sonab ? They had ideas like that too. I think you posted a photo.
What the ear is hearing then is the sound power. In the old days companies like Grundig and SABA had rooms to measure the sound power. It was a room without damping and asymmetric walls. The spaeker was put on one end and the mic on the other.
Yes, you could try to build something like that and then equalize.
The Scan Speak wideband i am using is a good candidate for that.
Off axis distortion is often lower to my surprise.
I could build something acording to your plan and listen.
Do you know Sonab ? They had ideas like that too. I think you posted a photo.
What the ear is hearing then is the sound power. In the old days companies like Grundig and SABA had rooms to measure the sound power. It was a room without damping and asymmetric walls. The spaeker was put on one end and the mic on the other.
yes I think that I did, and I've been thinking about something that could just be regarded as a more refined, audiophile quality version of Sonab V-1
sounds interesting, perhaps we could continue discussion in the original flooder thread:
http://www.diyaudio.com/forums/multi-way/121385-loudspeakers-room-system.html
but first I should read through it, gather all observations from all experimenting users and prepare a kind of interim resumé
best,
graaf