Hi, Agree, I think this fact is often overlooked when discussing higher frequencies are forming 'phantom spreading' from a set of speakers.
FYI: A cut and paste picture considering this matter:
b
This made my day !
Heureka !
"... the central image disappears, and a source appears located at each loudspeaker."
and
"A further increase in frequency ... finally all components cluster around the location of the loudspeakers."
This is what I've been explaining all along happening in my perception !
And further:
Heureka^2 !
"it is interesting to note that listeners sometimes considered that the side images possessed higher frequency components than the central image - as the theory would suggest."
The image split is reality ! "as the theory would suggest"
I'm happy now
- Elias
Bjorno, Elias,
thank you for the paper/link respectively.
I think as long as there is enough low and mid frequency content,
this may be sufficient for localization and the overtones being
"mapped" to the estimated locations of phantoms.
But it is clear, that this will not work well for all kinds of
"sounds" (instruments) reproduced and there may also be differences
due to different listeners.
Stereo's cues are "fragile" and may also be "contradictory" when
estimating the horizontal angle of phantoms, which may lead to
the speakers getting noticed as "the real sound sources".
This is my personal classification of "degradation":
A) Phantoms localized independently from speakers
B) Speakers getting noticeable as "speaker clouds", but phantom imaging still in effect.
C) Sounds are "glued" to - and audibly emanating - from the speakers.
The degradation may depend on (and start with) a particular sound (intended phantom source)
under question in a given recording.
Example from my personal listening experience :
In a certain recording/speaker/room situation imaging may work quite well
for many instrument groups. A female's sung vowels e.g. may also be "phantom
sourced" near the median plane quite well ...
until the first consonants (especially fricatives f,s) are "sung" or
a solo violin is played with high overtone (spectral) content :
Game over (for quite many speakers) !
Once the illusion has been seriously damaged, the speaker's position
is calculated more exactly. From now on further more cues are collected to
estimate the speaker's position - we will find them ! - and the whole
illusion may be destroyed even for those phantoms which seemed to have
worked well up to now ...
A friend of mine pointed me (again) to this:
http://www.moultonlabs.com/more/loudspeaker_as_musical_instrument/
There are truly some interesting thoughts (observations) IMO, which may also
be quite independent from (quick) conclusions on a technical (speaker design) level.
thank you for the paper/link respectively.
I think as long as there is enough low and mid frequency content,
this may be sufficient for localization and the overtones being
"mapped" to the estimated locations of phantoms.
But it is clear, that this will not work well for all kinds of
"sounds" (instruments) reproduced and there may also be differences
due to different listeners.
Stereo's cues are "fragile" and may also be "contradictory" when
estimating the horizontal angle of phantoms, which may lead to
the speakers getting noticed as "the real sound sources".
This is my personal classification of "degradation":
A) Phantoms localized independently from speakers
B) Speakers getting noticeable as "speaker clouds", but phantom imaging still in effect.
C) Sounds are "glued" to - and audibly emanating - from the speakers.
The degradation may depend on (and start with) a particular sound (intended phantom source)
under question in a given recording.
Example from my personal listening experience :
In a certain recording/speaker/room situation imaging may work quite well
for many instrument groups. A female's sung vowels e.g. may also be "phantom
sourced" near the median plane quite well ...
until the first consonants (especially fricatives f,s) are "sung" or
a solo violin is played with high overtone (spectral) content :
Game over (for quite many speakers) !
Once the illusion has been seriously damaged, the speaker's position
is calculated more exactly. From now on further more cues are collected to
estimate the speaker's position - we will find them ! - and the whole
illusion may be destroyed even for those phantoms which seemed to have
worked well up to now ...
A friend of mine pointed me (again) to this:
http://www.moultonlabs.com/more/loudspeaker_as_musical_instrument/
There are truly some interesting thoughts (observations) IMO, which may also
be quite independent from (quick) conclusions on a technical (speaker design) level.
Last edited:
First of all these "trading effects" IMO point to the imperative to have a speaker
as free from resonance (and the frequency response smooth even at narrow band
level especially in highs) as possible, to not emphasize these effects.
But this is just one imperative, it is not solving the problem at its roots ...
which would call for a strategy to significantly attenuate stereo's deficiencies in that
respect.
According to this thread's subject "use a center speaker" is not an appropriate
answer, right ?
If you want to (gradually) circumvent interference between two sources there's only
one way to go IMO, that is (gradually) circumventing phase correlation in the frequency
range under question.
Since HRTF and IDT is exploited for localization, it is at least questionable, whether room
reflections can make up for these deficiencies in localization.
For some situations (sounds) that sound field "smoothing" coming from room reflecions may be
helpfull, but to repair the inherent deficiencies of the stereo "direct sound" those are simly too late ...
which may be a reason for some to prefer dense early reflectons to make "speakers disappear".
The problem i see is: Even those are too late. If you go for decorrelation, it has to be builtin to the
transducers with the lowest amount of group delay possible.
as free from resonance (and the frequency response smooth even at narrow band
level especially in highs) as possible, to not emphasize these effects.
But this is just one imperative, it is not solving the problem at its roots ...
which would call for a strategy to significantly attenuate stereo's deficiencies in that
respect.
According to this thread's subject "use a center speaker" is not an appropriate
answer, right ?
If you want to (gradually) circumvent interference between two sources there's only
one way to go IMO, that is (gradually) circumventing phase correlation in the frequency
range under question.
Since HRTF and IDT is exploited for localization, it is at least questionable, whether room
reflections can make up for these deficiencies in localization.
For some situations (sounds) that sound field "smoothing" coming from room reflecions may be
helpfull, but to repair the inherent deficiencies of the stereo "direct sound" those are simly too late ...
which may be a reason for some to prefer dense early reflectons to make "speakers disappear".
The problem i see is: Even those are too late. If you go for decorrelation, it has to be builtin to the
transducers with the lowest amount of group delay possible.
Last edited:
Secondly, there was no imaging !
And thirdly, I can locate the speakers on the floor
so it's truly a mystery as You were once unable to locate a single flooder playing mono:
Yes, because adding a center speaker to the stereo triangle has been done for decades and is conventional
However, discussing about the use of a center only speaker would be appropriate since it's still kind of unconventional, maybe even too much for this forum I think phase decorrelation is an interesting idea and I will add it to my palette too ! But what matters is the decorrelation of the in ear signals of the listener at the listening position, and there may be several methods to achieve a required level of it.
Well, but we have IACC
Does it serve the purpose ? Obviously a new analysis based on the correlation measure is needed ! This may be true for a conventional stereo triangle !
But a system that is designed to depend on reflections is other matter ! As I appear to have been using the decorrelation in my current system (SSSx5) without explicitly thinking it.
An easy test playing mono wide band pink noise, for a conventional stereo triangle listening at the listening position there is strong interference notches perceived if moving slightly sideways and/or turning the head even slightly.
But for SSSx5 playing same mono pink noise I cannot perceive any interference notches even moving 1 m sideways and/or turning my head all the way 360 degrees at the listening position ! This must be an indication of strong phase decorrelation taking action.
The clue how is this possible may be in the temporal delay between the first and second order side reflections, which is comparable to the time constant of human sound processing in time-freq domain (Think Bark, ERB, Gammatone etc. models of auditory systems).
- Elias
Yes it's a mystery, but not very intriguing one since the discovery/reinvention of SSSx5
Or maybe my perception has improved ?
Or maybe a higher standard of reference has been gained ? (Yes, the SSSx5
)- Elias
I think phase decorrelation is an interesting idea and I will add it to my palette too ! But what matters is the decorrelation of the in ear signals of the listener at the listening position, and there may be several methods to achieve a required level of it.
Well, but we have IACC
...This may be true for a conventional stereo triangle !...
...But a system that is designed to depend on reflections is other matter ! As I appear to have been using the decorrelation in my current system (SSSx5) without explicitly thinking it.....
...But for SSSx5 playing same mono pink noise I cannot perceive any interference notches even moving 1 m sideways and/or turning my head all the way 360 degrees at the listening position ! This must be an indication of strong phase decorrelation taking action...
...The clue how is this possible may be in the temporal delay between the first and second order side reflections, which is comparable to the time constant of human sound processing in time-freq domain (Think Bark, ERB, Gammatone etc. models of auditory systems)...
Hi, Adding more food for thoughts..Read the *Hirofumi/Yoshio and Takeshi paper + note the reference #9 that really got my attention in the mid 80:ties and ever since that time I've added similar circuits to my designed speaker systems(mono systems too) as shown in the Hirata publication for De-correlation purposes..
I.e.: One or two short frontal placed Bessel arrays with pink noise shaped FR from ~60-800Hz. in order to improve the reproduction live recorded Double-Bass like heard in Jan Johansson/Georg Riedel records that I'm grown up with (early-mid 60:ties).
*
http://www.acoust.rise.waseda.ac.jp/publications/happyou/jasje/jasje84-5-1988Nov.pdf
b
Attachments
"Beveridge" officially sucks acording to BBC ?
Found this BBC recearch report
"R&D Report 1963-01 : Influence of loudspeaker directivity and orientation on the effective audience area in two-channel stereophonic reproduction"
They investigate stereo speaker placement with 90 degree toe in i.e. face-to-face, which is similar to Beveridge
http://downloads.bbc.co.uk/rd/pubs/reports/1963-01.pdf
They complain blur imaging among others and conclude:
"The face-to-face arrangement cannot be recommended for broadcast monitoring purposes; its employment by the individual listener is a matter of taste"
and that it will
"make the stereophonic presentation ... uniformly bad"
Well
- Elias
Found this BBC recearch report
"R&D Report 1963-01 : Influence of loudspeaker directivity and orientation on the effective audience area in two-channel stereophonic reproduction"
They investigate stereo speaker placement with 90 degree toe in i.e. face-to-face, which is similar to Beveridge
http://downloads.bbc.co.uk/rd/pubs/reports/1963-01.pdf
They complain blur imaging among others and conclude:
"The face-to-face arrangement cannot be recommended for broadcast monitoring purposes; its employment by the individual listener is a matter of taste"
and that it will
"make the stereophonic presentation ... uniformly bad"
Well
- Elias
...
Well, but we have IACC
...
To estimate the quality of a center phantom (especially with dominant
HF spectral content) i would propose measuring anechoic frequency response
of the stereo speakers at some defined positions, which also are outside the
median plane. That is simple and revealing i guess.
Then a "smoothness" measure is needed for the frequency responses and the
change with microphone position as well.
Of course only a mono source having flat FR and which is CD will perform
to 100% ... but we are talking stereo here.
I would expect coherently radiating stereo speakers to perform rather poor
in that task and probably not very different from each other, being it omnis,
"regular multiway" or (even "true") dipoles ...
For reverberant conditions why not use IACC ?
Last edited:
Found this BBC recearch report
"R&D Report 1963-01 : Influence of loudspeaker directivity and orientation on the effective audience area in two-channel stereophonic reproduction"
They investigate stereo speaker placement with 90 degree toe in i.e. face-to-face, which is similar to Beveridge
...
Well
- Elias
oh well well
c'mon Elias! "Beveridge" officially sucks acording to BBC?
what is the point of this provocation?
read about Beveridge ESLs to find out the difference between them and BBCs used in those tests, not to mention a flooder
No way
I'm just quoting what the unquestionable authority has written about the face-to-face speaker configuration ! Dont you believe
It is suggested that it is not suitable for a critical listening, but rather is a matter of taste of the listener.
I'm of course aware that Beveridge includes ELS and all that, but what is the main contribution of the concept ? Face-to-face ?
I see a real problem with the Beveridge concept, namely the strong contralateral wall reflection.
In order to make it work maybe 180 degrees dispersion is needed ? BBC were not using such a speakers, as far as can be extracted from the document.
So I conclude, face-to-face is not a single dominant feature of stereo system.
- Elias
Elias,
i have a question concerning your single speaker stereo experimental
approach having x=0.5 .
Do phantoms which are panned to left or right sound the same as
if they were centered ?
Due to reflection paths i would at least expect them to appear more
distant than a centered phantom of equal "presence" on the recording.
Since the side (virtual mirror) sources have far larger distance to the
listener than the (real) center source.
If you feel this not being the case, why do you think it works that way ?
i have a question concerning your single speaker stereo experimental
approach having x=0.5 .
Do phantoms which are panned to left or right sound the same as
if they were centered ?
Due to reflection paths i would at least expect them to appear more
distant than a centered phantom of equal "presence" on the recording.
Since the side (virtual mirror) sources have far larger distance to the
listener than the (real) center source.
If you feel this not being the case, why do you think it works that way ?
Interesting question !
If there is no reveberation in the signal, then I think the panned phantom images appear, not on a arc of a circle, but on a straigth line. The line is located just behind the front wall.
If there is reverberation in the signal, the distance depends totally on the characteristics on the reverberation and is typically several meters, even tens of meters behind the front wall.
- Elias
In a conventional stereo triangle, do phantoms which are panned to left or right sound the same as if they were centered ?
Of course they don't !
Some thoughts on that
"Is Flat (anechoic FR) always correct"
"What is the ideal directivity pattern for stereo"
"How to achieve outstanding imaging"
matters.
1) When we judge the performance of home stereo speakers, we usually
do that by listening to commercial recordings.
Instead of discussing pros and cons of certain recording techniques
and why some recordings sound more pleasing to some listeners, i want
to focus on the mastering process in the control room.
The equipment used and the properties of the control room impose a
de facto norm on the collection of recordings, assuming that recording
engineers in common have more or less the same habits in "enjoyfull" or
"all day" listening to music.
This is a kind of positive assumption that "we all are normal hearing
persons" - being it pro or naive listeners - or at least we define
collectively, what a "normal hearing person" is.
Of course a sound engineer's working environment differs significantly
from a living room.
Mostly we assume the "norm" in that environment to be a frequency response
of the direct sound to be flat.
Assuming this, we find that there is also an implicit norm for energy
response, which is defined by the average radiation pattern of studio
monitors used.
The norm for the reverberant portion is defined by that average radiation
pattern and common habits in designing a control room ... if such thing
is existent, but i assume there is a bottom line.
A functional control room e.g. can be expected to have about less than
the half reverberation time of a common (german) living room nowadays.
Usually more care is spent on having that reverberation time independent
from frequency, which will cause a shorter decay in a control room
especially at low frequencies than in a usual hard walled living room.
But this may not be a fixed rule ...
2) Averaged studio monitor behaviour may be assumed to have a radiation
pattern considerably narrowing with frequency.
Above the modal or even transitional frequency range of the room, the
frequency range from say 120 to 300 Hz will mostly have falling energy
response as well as the range above say 6Khz.
If we distinguish a 2-way and a 3-way norm, there might also be considerable
narrowing of the radiation pattern from 300Hz upwards to say 2.5 Khz caused
by most 2-way monitors.
Keeping the assumption of a flat anechoic response being the norm, there
is a remaining frequency band from about 2Khz to 6Khz, where on axis response
as well as energy response both may be considered as "presumably" flat on
average.
3) Studio Listening: The provocational thesis (yes i know, there have been lots of studies ...):
Flat anechoic response is only the "shell" or kind of "label" of an underlying
norm, which is hidden inside that shell. The underlying recording norm may comprise
- anechoic frequency response
- energy response (and average directivity pattern) of the monitor
- preferred directions of reflections in the control room which are also dependent
from frequency
That norm as a whole is labeled "flat frequency" response, because it is the most
convenient aspect to measure using nearfield or gated measurements.
The other aspects are more diverse in comparison between even control rooms,
and the aquisition of data is by far more complex.
But the label is not the norm itself, it only comes close to a norm under
control room conditions which ought to be quite uniform.
Control room conditions are restricted by several aspects:
- The goal of having appropriate working conditions for the mastering process
- The use of currently available and affordable technology (Monitors etc.)
- The possibilities of acoustical room treatment
I want to concentrate on the first aspect:
One of the main tasks of the recording engineer is creating a stereo panorama
with "believable" phantom images occuring centered, half way left etc.,
make their spectral fingerprints believable, provide "the right" amount of
presence, reverberation etc. to make the acoustic scene cohesive "as a whole".
Sitting outside the median plane of the stereo triangle, even when that deviation
is small, will be detrimental to that kind of work.
Even (especially ipsilateral) side wall reflections caused by a monitor with
non uniform radiation pattern, will cause combing (spectral deviations)
and possibly spatial deviations in localization of transients.
Especially with given technology, there are good reasons to provide that kind
of special working conditions in using a room wich is more absorbent
(and hopefully also more diffuse ...) than a common living room.
Mastering is maybe in some way comparable to looking at a printing using
a magnifier:
Like people in the print business, who know how objects "should look like" when
magnified (to look properly when looked at with laymen's bare eyes), a recording
engineer has a "working mode of listening" which will create a result that holds
under conditions that can be assumed to be more reverberant.
Nothing else makes any sense, bacause living rooms are no control rooms.
4) Home Listening
The home listener usually wants a "live like" presentation. Unless we cannot explain
properly or even to some detail what this really is, we could diverse ourselves in
describing the home listeners "common" environment, which may even be less uniform
than the production environments used to create commercial recordings.
The home listener's environment is/has usually
- more reverberant
- a geometry of the stereo triange which is less controlled in terms of symmetry
(be that caused by the room, the interior, multiple listeners listening ...)
5) Home speaker
As there seems to be a tendency for most listeners to prefer "wide" radiating
speakers, we should ask why this is the case and how that aligns with the de facto
recording norm (which is quite fuzzy) as described above.
In most real acoustic venues, listeners sit near or beyond the "critical distance"
of direct sound and reflected sound contributing equally to resulting sound pressure
levels.
Usual recordings provide remarkably dominant direct sound but with sufficient reverb
to support "cohesion" and provide "realistic" cues to form a "recording venue" at least
if the recording venue is not entirely virtual, which is even genre specific.
Why wide radiating ? There might be several reasons
- add sufficient reverb to get the direct/reflected ratio close to common preferences in "live" situations
- being more independent in choosing the listening seat
there may be even more motivations.
But what is "wide radiating" ?
- Putting the studio monitor in a more reverberant environment could be "wide radiating" enough ?
- What kind of radiation pattern over frequencies is preferable ?
- Is there a pattern, which is "more correct" than others ?
No matter how we answer those questions, as soon as we choose a radiation pattern deviating
from the "implicit recording norm", we will have additional restrictions to face.
From an esthetical point of view (e.g. "make it live like"), there is no reason at first
in maintaining a radiation pattern narrowing with frequency, because real acoustic venues
do not have those properties in the same way as it was present in the control room:
This is the case especially from upper bass to lower treble and it is questionable for the
two uppermost octaves too, where stereo btw. shows its weaknesses in an unvarnished manner.
But when deciding towards some kind of "constant directivity", we are leaving the
implicit recording norm. That holds for a virtual (fullrange) 90x90 degrees CD system as well
as for an omni (yes, even a true omni is a CD speaker).
Looking e.g. at the range >6Khz the uppermost octaves will be overrepresented in the living
room and the sound will be way too bright.
The same desire for downtilting the response will probably occur from 100Hz to about 2Khz.
Oops, we have CD now ... we cannot maintain a flat anechoic frequency response on axis
and by the same time fullfill the "implicit recording norm" (IRN i love that term).
So we - by and by - end up with noteably downtilted responses even in the "on axis
anechoic portion".
6) Tasks of a home speaker deviating from the recording norm, e.g. by having constant directivity
When deviating from the recording norm due to
- anechoic frequency response
- energy response (and average directivity pattern) of the monitor
- preferred directions of reflections in the control room which are also dependent
from frequency
that combined balance has to be re-estimated taking expected (home) room acoustics
into account and equalise the anechoic frequency response in sound pressure
(which in an ideal CD system closely resembles the energy response) accordingly.
Unfortunately combing by wall reflections might have an effect on tonal balance,
which is frequency dependent and also physiological effects have to be taken into account.
That rises the question how to make a speaker that is CD and shows little combing
by nondiffuse esp. early reflections ...
Nevertheless, that equalizing or "rebalancing" step has to be done for every speaker
deviating from the implicit recording norm by radiation pattern in order to obey the norm
in terms of tonal balance - which has high priority.
On the other hand deviating from the implicit recording norm seems to be needed to
get "outstanding imaging" under home conditions.
Kind Regards
"Is Flat (anechoic FR) always correct"
"What is the ideal directivity pattern for stereo"
"How to achieve outstanding imaging"
matters.
1) When we judge the performance of home stereo speakers, we usually
do that by listening to commercial recordings.
Instead of discussing pros and cons of certain recording techniques
and why some recordings sound more pleasing to some listeners, i want
to focus on the mastering process in the control room.
The equipment used and the properties of the control room impose a
de facto norm on the collection of recordings, assuming that recording
engineers in common have more or less the same habits in "enjoyfull" or
"all day" listening to music.
This is a kind of positive assumption that "we all are normal hearing
persons" - being it pro or naive listeners - or at least we define
collectively, what a "normal hearing person" is.
Of course a sound engineer's working environment differs significantly
from a living room.
Mostly we assume the "norm" in that environment to be a frequency response
of the direct sound to be flat.
Assuming this, we find that there is also an implicit norm for energy
response, which is defined by the average radiation pattern of studio
monitors used.
The norm for the reverberant portion is defined by that average radiation
pattern and common habits in designing a control room ... if such thing
is existent, but i assume there is a bottom line.
A functional control room e.g. can be expected to have about less than
the half reverberation time of a common (german) living room nowadays.
Usually more care is spent on having that reverberation time independent
from frequency, which will cause a shorter decay in a control room
especially at low frequencies than in a usual hard walled living room.
But this may not be a fixed rule ...
2) Averaged studio monitor behaviour may be assumed to have a radiation
pattern considerably narrowing with frequency.
Above the modal or even transitional frequency range of the room, the
frequency range from say 120 to 300 Hz will mostly have falling energy
response as well as the range above say 6Khz.
If we distinguish a 2-way and a 3-way norm, there might also be considerable
narrowing of the radiation pattern from 300Hz upwards to say 2.5 Khz caused
by most 2-way monitors.
Keeping the assumption of a flat anechoic response being the norm, there
is a remaining frequency band from about 2Khz to 6Khz, where on axis response
as well as energy response both may be considered as "presumably" flat on
average.
3) Studio Listening: The provocational thesis (yes i know, there have been lots of studies ...):
Flat anechoic response is only the "shell" or kind of "label" of an underlying
norm, which is hidden inside that shell. The underlying recording norm may comprise
- anechoic frequency response
- energy response (and average directivity pattern) of the monitor
- preferred directions of reflections in the control room which are also dependent
from frequency
That norm as a whole is labeled "flat frequency" response, because it is the most
convenient aspect to measure using nearfield or gated measurements.
The other aspects are more diverse in comparison between even control rooms,
and the aquisition of data is by far more complex.
But the label is not the norm itself, it only comes close to a norm under
control room conditions which ought to be quite uniform.
Control room conditions are restricted by several aspects:
- The goal of having appropriate working conditions for the mastering process
- The use of currently available and affordable technology (Monitors etc.)
- The possibilities of acoustical room treatment
I want to concentrate on the first aspect:
One of the main tasks of the recording engineer is creating a stereo panorama
with "believable" phantom images occuring centered, half way left etc.,
make their spectral fingerprints believable, provide "the right" amount of
presence, reverberation etc. to make the acoustic scene cohesive "as a whole".
Sitting outside the median plane of the stereo triangle, even when that deviation
is small, will be detrimental to that kind of work.
Even (especially ipsilateral) side wall reflections caused by a monitor with
non uniform radiation pattern, will cause combing (spectral deviations)
and possibly spatial deviations in localization of transients.
Especially with given technology, there are good reasons to provide that kind
of special working conditions in using a room wich is more absorbent
(and hopefully also more diffuse ...) than a common living room.
Mastering is maybe in some way comparable to looking at a printing using
a magnifier:
Like people in the print business, who know how objects "should look like" when
magnified (to look properly when looked at with laymen's bare eyes), a recording
engineer has a "working mode of listening" which will create a result that holds
under conditions that can be assumed to be more reverberant.
Nothing else makes any sense, bacause living rooms are no control rooms.
4) Home Listening
The home listener usually wants a "live like" presentation. Unless we cannot explain
properly or even to some detail what this really is, we could diverse ourselves in
describing the home listeners "common" environment, which may even be less uniform
than the production environments used to create commercial recordings.
The home listener's environment is/has usually
- more reverberant
- a geometry of the stereo triange which is less controlled in terms of symmetry
(be that caused by the room, the interior, multiple listeners listening ...)
5) Home speaker
As there seems to be a tendency for most listeners to prefer "wide" radiating
speakers, we should ask why this is the case and how that aligns with the de facto
recording norm (which is quite fuzzy) as described above.
In most real acoustic venues, listeners sit near or beyond the "critical distance"
of direct sound and reflected sound contributing equally to resulting sound pressure
levels.
Usual recordings provide remarkably dominant direct sound but with sufficient reverb
to support "cohesion" and provide "realistic" cues to form a "recording venue" at least
if the recording venue is not entirely virtual, which is even genre specific.
Why wide radiating ? There might be several reasons
- add sufficient reverb to get the direct/reflected ratio close to common preferences in "live" situations
- being more independent in choosing the listening seat
there may be even more motivations.
But what is "wide radiating" ?
- Putting the studio monitor in a more reverberant environment could be "wide radiating" enough ?
- What kind of radiation pattern over frequencies is preferable ?
- Is there a pattern, which is "more correct" than others ?
No matter how we answer those questions, as soon as we choose a radiation pattern deviating
from the "implicit recording norm", we will have additional restrictions to face.
From an esthetical point of view (e.g. "make it live like"), there is no reason at first
in maintaining a radiation pattern narrowing with frequency, because real acoustic venues
do not have those properties in the same way as it was present in the control room:
This is the case especially from upper bass to lower treble and it is questionable for the
two uppermost octaves too, where stereo btw. shows its weaknesses in an unvarnished manner.
But when deciding towards some kind of "constant directivity", we are leaving the
implicit recording norm. That holds for a virtual (fullrange) 90x90 degrees CD system as well
as for an omni (yes, even a true omni is a CD speaker).
Looking e.g. at the range >6Khz the uppermost octaves will be overrepresented in the living
room and the sound will be way too bright.
The same desire for downtilting the response will probably occur from 100Hz to about 2Khz.
Oops, we have CD now ... we cannot maintain a flat anechoic frequency response on axis
and by the same time fullfill the "implicit recording norm" (IRN i love that term).
So we - by and by - end up with noteably downtilted responses even in the "on axis
anechoic portion".
6) Tasks of a home speaker deviating from the recording norm, e.g. by having constant directivity
When deviating from the recording norm due to
- anechoic frequency response
- energy response (and average directivity pattern) of the monitor
- preferred directions of reflections in the control room which are also dependent
from frequency
that combined balance has to be re-estimated taking expected (home) room acoustics
into account and equalise the anechoic frequency response in sound pressure
(which in an ideal CD system closely resembles the energy response) accordingly.
Unfortunately combing by wall reflections might have an effect on tonal balance,
which is frequency dependent and also physiological effects have to be taken into account.
That rises the question how to make a speaker that is CD and shows little combing
by nondiffuse esp. early reflections ...
Nevertheless, that equalizing or "rebalancing" step has to be done for every speaker
deviating from the implicit recording norm by radiation pattern in order to obey the norm
in terms of tonal balance - which has high priority.
On the other hand deviating from the implicit recording norm seems to be needed to
get "outstanding imaging" under home conditions.
Kind Regards
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.