I equalize frequency response flat in the listening position, not in near field. I want to hear exactly what recording engineer did in studio. The effect of room is minimalized that way.
I am always happy with hights afterwards, it's the bass extension which is a problem. Equalizer is trying to make it flat down to 10 Hz, which is not realistic even with 18" woofers. It causes too much excursion which leads to distortion. Plus its heard far, it disturbs my neighbours. I found it just fine to roll it gently down from 30Hz. There is not much musical content below 20Hz anyway.
I am always happy with hights afterwards, it's the bass extension which is a problem. Equalizer is trying to make it flat down to 10 Hz, which is not realistic even with 18" woofers. It causes too much excursion which leads to distortion. Plus its heard far, it disturbs my neighbours. I found it just fine to roll it gently down from 30Hz. There is not much musical content below 20Hz anyway.
I highly recommend listening to Linkwitz presentation about accurate sound reproduction:
http://www.youtube.com/watch?v=VC-sxvNzC8I
http://www.youtube.com/watch?v=VC-sxvNzC8I
Rudolf said:graaf,
first and OT: Congratulations to your latest "loud speaker".
He possibly might be the most demanding one you ever build!
thanks!
o yeah! I know
Fortunately I have some experience and training as I have 4 years old daughter as well Daddy’s and Mommy’s little helper Rudolf said:
a mistake – 1 ms is one cycle of 1000 Hz wave (1000 Hz is "1000 cycles per second")
Rudolf said:
So your initial transient attack (if we talk about ~ 50 msec) is quite far away from the common step or impulse response transient, which even for a fullrange speaker is done in less than 5 msec for the most part.
(…)
In my view you are still argumenting in the realm of "transient" response when it already is "changing frequency" response.
a qui pro quo
perhaps I haven’t made myself clear enough
~ 50 msec or even more (perhaps up to 100 ms?) is a time span of a voice or musical instrument's clear, defined tone formation (time span of this "process of fighting inertia")
yes, in the physical sound source it is the realm of "changing frequency response" but not for uour hearing sense, not in the ear and brain
what I call the "initial waveform" or "the shape of the first transient attack" or more precisely "initial transient wave attack curve shape" or "the shape of transient impulse rise curve" is nly the shape of the rising curve of the wave cycle, corresponding to the "climbing up" first 90 degrees of a sine wave
or even a part of it (in all cases lasting much less than those 50 ms)
WHY?
because localization phase of hearing mechanism is ended well before the first millisecond passes!
I was thinking (as much as my loud speaker allows
) how to express most clearly how I understand all this and here is what I came to (I apologize in advance for any lack of clarity ):localization of sound sources in reverberant environment in the lateral plane (left-right) depends on interaural time difference (ITD) of the first wavefront arrival at the respective ears. Maximum ITD is around 650 microsec (0.65 ms) for a typical head size/distance between the ears (17-18 cm)
for the localization mechanism to react the two wavefronts, arriving at the respective ears one after the other, have to be somewhat related to each other, in other words – the two wavefronts have to share a common characteristic.
what can it be if it all happens <650 microsec that is when the voice or an instrument barely start to sound out and we are only at AN ONSET of "wild spectral fluctuations called the initial transient"?
IMHO this can be only the shape of the wavefront which is the shape of rising curve of the wave cycle – it is the timeline of sound pressure build up which is unique for a particular sound source as is the shape of the whole wave cycle - for example smooth and resembling sine wave for a flute tone and jagged and not resembling sine wave at all for a violin tone
our sense of hearing can be very precise in analyzing and identifying this shape because time resolution of this process is astounding - after all ~650 microsec is MAXIMUM interaural time delay
in fact psychoacoustic research showed that time resolution of binaural (and also monaural) hearing reaches the range of 10-20 microseconds!
see: www.mrc-cbu.cam.ac.uk/~roy/RoyCV/pdfs/KPNFjasa03.pdf
so the two wavefronts arriving at the two ears have to share a common characteristic to form any sense of a "direct sound"
this characteristic can be only "the shape of rising curve of the wave cycle", "the timeline of sound pressure build up" because THIS IS ALL THAT IS THERE PHYSICALLY at this early phase of voice or musical instrument tone formation
the frequency content is not yet defined (for longer waves even one cycle is not yet completed!) or is at best very unstable and chaotically changing ("wild spectral fluctuations", "chaos ensues")
my hypothesis is that if the direct sound is identified "as such" (in the sense that "the two wavefronts are coming from the same sound source and indicate its direction") on the basis of "the shape of rising curve of the wave cycle" then only this characteristic can serve as a basis of comparison by the brain of reflected sound waves to the direct sound wave as well
Rudolf said:
it has been observed many times that linear frequency response is perceived as excessively bright
most listeners prefer response gently tilting down towards high frequencies
first of all - this is all very subjective - "dark" for one listener is "bright" for another, it is a matter nto only of preferences but of varying hearing sensitivity at high frequencies - therefore women are more sensitive to brightness as they hear better at high frequencies than men (they had to find a crying baby in the woods in the old days
)as to Your case and physics behind what You observe I believe that it is not a time domain effect
it is a matter of power response and absorptive characteristics of the room (room response)
boosting the highs as measured on axis (to linearize frequency response on axis) means boosting the highs also off axis - You end having more highs in the power response in Your room
in the nearfield You hear less of the room sound (more of the sound of the loudspeakers) but at Your listening position the sound can be dominated by the room sound
perhaps when You are boosting the highs to linearize frequency response on axis You get too much for Your room with its particular absorptive characteristics, relatively bright from perspective of Your preferences and hearing sensitivity to high frequencies
what You hear at Your listening position is not linearized on axis response but power response in which the highs can be excessively boosted by the room response that You hear more of at Your listening position than in the nearfield
best regards,
graaf
I just pulled out my ceiling firing TQWTs for the Ciare HX201 (similar to Visaton B200) out of my storage room to do some measurements in Arta. Using it in an other room than I had them before I ended up with a higher delay for the ceiling reflection. The comparison of listening impressions with measurements showed that one only gets a well-defined spacial image when there is a well-defined impulse for the ceiling reflection. In fact I managed to get it as good as a typical direct response. The delay was 5ms, same as Etienne measured, but his reflected impulse is everything else than ideal. Strangely the direct impulse (which is always the same since I have the driver on ear-height) is never good, but causes the image being perceived in front, not above. Remember the precision of the image is determined by the ceiling reflection! I haven´t followed the psychoacoustic discussion, but I don´t think the existing psychoacoustic knowledge can explain this.
Regards,
Oliver
Regards,
Oliver
what are the factors? I mean what causes an impulse of the ceiling reflection well-defined or ill-defined? In other words - what have You done to get well-defined impulse?el`Ol said:
but it is the same in every lateral direction - and it is sufficient for precise left-right spatial sound perceptionel`Ol said:
I believe that it does because the ceiling reflection is sufficiently similar to the direct sound (waveform) to be interpreted as a reflection?el`Ol said:
anyway, frankly speaking I don’t know of any scientific research of the effects of vertical reflections from above (like in case of ceiling reflection) for spatial sound perception in the vertical plane
theories based on research of binaural perception of lateral reflections and their effects on spatial sound perception - from Helmuth Haas on - generally don’t apply in case of vertical reflections because vertical sound source localization relies on completely different cues than ITD/ILD - it has to as there are no "interaural differences" when a sound source is moving in the vertical plane
vertical sound source localization relies mainly on the floor ("ground" in general) reflection (here hearing relies on distance cues and delay of the reflection) and on head and pinnae related transfer function AFAIK and IIRC
in my case although the loudspeakers are just 20 cm above the floor I perceive sound images at the ear height
as it is in case of other short omni speakers like the Carlssons
so perhaps there are also other factors
I believe that it can and that the answer is simple - ceiling reflection carries much information, in fact more than the direct soundel`Ol said:
no problem that it arrives delayed - only a sense of lateral (left-right) direction of sound is established <1 ms (more precisely <700 microsec for a typical head size) after the first wavefront arrival
the whole process of perception of a defined sound image can last longer and depend also on other factors
el`Ol said:
but what about the reflected sound fom the main speaker that would be there anyway?
best!
graaf
Hello Graaf!
Everything I changed was the listening distance.
I find it interesting that perception on ear-height is possible when the speaker stands on the floor. I never had that experience with the HX201. I tried to get an old Sonab, but a Philips-hunter paid more
Maybe I get a similar experience as you when I have my clone of the later Carlssons ready. But this project is currently paused because I just got the lasered stators for my magnetostat. To do scientifically correct psychoacoustic tests one would have to use narrow beam constant directivity devices. I played with the idea of using one firing to the ceiling, but after my measurement I wonder whether one in front and one at the ceiling (both facing the listener) would be better.
Oliver
Everything I changed was the listening distance.
I find it interesting that perception on ear-height is possible when the speaker stands on the floor. I never had that experience with the HX201. I tried to get an old Sonab, but a Philips-hunter paid more
Maybe I get a similar experience as you when I have my clone of the later Carlssons ready. But this project is currently paused because I just got the lasered stators for my magnetostat. To do scientifically correct psychoacoustic tests one would have to use narrow beam constant directivity devices. I played with the idea of using one firing to the ceiling, but after my measurement I wonder whether one in front and one at the ceiling (both facing the listener) would be better.Oliver
I just did my first test with the mid/high unit only. Damping is not perfect yet, just 5 mm dense polyester felt. I can now understand why Stig Carlsson´s paper deals so much about purist / original acoustic recordings vs. "modern" multi-microphone recordings or even those with synthetic reverb. With purist recordings like older ones from BIS or Harmonia Mundi France I do not only get the reach out and touch effect, but also a sense of the recording room / concert hall and the placement of the instruments in it. What I experienced with a less purist recording I liked so far (Alia Vox) was a mess. The sound comes from two speakers and a 2 m diameter clump of mud between them right behind the wall. The psycho-acoustic reverb algorithms that were optimized for hiding conventional speakers do the opposite here. For those who want to experience the effect but don´t want to mourn about the recording glory of the past all day the "ceiling flooder" is probably the better solution. Many of us have the Visaton B200 and just need to build a TL with them.
Oliver
Oliver
el`Ol said:
This is an outrageous comment that does not belong in these posts.
I know David and I certainly don't agree with everything that he says, nor do I think his method of testing using his own personal opinions as data are very scientific, but he is an intelligent educated and committed professional with a great deal of background. He would never resort to such childish comments as this. Grow up!
@graaf:
The "ceiling flooder" was the speaker with the ceiling-firing HX201 with the impulse response I posted above.
@gedlee:
Of course I don´t hate David Griesinger as a person, but I have always hated the signature the Lexicon reverbs produce and with the carlsson-style speakers the spacialness I never really liked, but could live with more or less is gone, so the effect is reduced to its worst.
Fortunately I don´t listen to a lot of music that is produced that way, and I still have a second system (OB) that has a strange kind of spacialness anyway.
Regards,
Oliver
The "ceiling flooder" was the speaker with the ceiling-firing HX201 with the impulse response I posted above.
@gedlee:
Of course I don´t hate David Griesinger as a person, but I have always hated the signature the Lexicon reverbs produce and with the carlsson-style speakers the spacialness I never really liked, but could live with more or less is gone, so the effect is reduced to its worst.
Fortunately I don´t listen to a lot of music that is produced that way, and I still have a second system (OB) that has a strange kind of spacialness anyway.
Regards,
Oliver
OK, sorry David, in case you can read this.
I just played some samples from the Quantec website, and I find the QRS algorithm is really a different cup of tea. But probably people have become so used to Lexicons that the Quantec (despite being back again with full functionality for a year or so)will stay a nieche product for TV and Cinama productions.
I just played some samples from the Quantec website, and I find the QRS algorithm is really a different cup of tea. But probably people have become so used to Lexicons that the Quantec (despite being back again with full functionality for a year or so)will stay a nieche product for TV and Cinama productions.
Hello graaf!
I mean the OA-50 series:
http://www.carlssonplanet.com/oa52_2.php
However I use them with the Ciare HX160 on ear-level, but directly in front of the wall, as proposed by Stig Carlsson. The room has a wooden ceiling, less reflecting than proposed by Carlsson, so maybe this is the reason why I can´t use them in a lower position.
The frequency response is rather flat for a paper cone fullranger from 150 Hz to 17 kHz, unlike one would expect from the measurement on axis (which is really as bad as shown at the Ciare website).
Woofers are ordered.
regards,
Oliver
I mean the OA-50 series:
http://www.carlssonplanet.com/oa52_2.php
However I use them with the Ciare HX160 on ear-level, but directly in front of the wall, as proposed by Stig Carlsson. The room has a wooden ceiling, less reflecting than proposed by Carlsson, so maybe this is the reason why I can´t use them in a lower position.
The frequency response is rather flat for a paper cone fullranger from 150 Hz to 17 kHz, unlike one would expect from the measurement on axis (which is really as bad as shown at the Ciare website).
Woofers are ordered.
regards,
Oliver
gedlee said:
Gedlee,
If I understood correctly DG's approach in bass in small rooms, DG suggest to drive L-R LF (below say 120Hz) with 90deg phase shift for the best "sound" to reach reasonably high interaural time delay to make LF sound externalized. But, as you wrote, phase shift cannot decorelate LF signals. IMO because there is simply not enough time/volume for multiple reflections to do it. Quite a difference.
DG recomend (or rely upon?) adding an uncorelated reverb in the left and right channel into the record for requencies below say 1000Hz to support "envelopment and externalization". What to do if it is not the case? Adding a reverbating device into the signal chain does not seem to me as a viable solution. Does the damped resonant walls adds the reverb? Or their duty is only to damp a pressure modes?