"The Clue" - ny amerikansk högtalare

[luminous]

James,

You're right that it's a bit difficult to discuss these kinds of very involved subjects on a forum. But then I think the purpose of a forum discussion may not be to reach scientific-quality conclusions but to have a good entertaining discussion. And for me it is always rewarding to discover ways of thinking that I haven't considered before. I do admire people such as yourself and others who are willing to contribute their great knowledge in these discussions.

I understand your standpoint in defining spatial neutrality, I think it makes good sense. A spatial aspect of sound perception that I believe is very difficult to reproduce though using a stereo system is the room "ambience", the natural ambience of the listening room will add to the recorded ambience which cannot really be reproduced naturally from only two points in space in front of the listener. So that's where I think there is some room for different listener preferences in choosing the type of ambience and degree of liveliness of the listening room.

It seems to me that it would not be very profitable for most companies in the audio business to use their resources for researching the kind of intricate low-level subjects that you are talking about in this thread. How have you found time for this?

I really like your notion that only a subset of the information in the original sound field needs to be reproduced to yield a realistic experience, and that knowledge of this subset is really vital in designing the reproduction system. It is certainly obvious that a stero-recording can only store a small part of the information in the original sound field, but can still give a very pleasurable experience when reproduced over a good system.
This also explains that different use-models of loudspeakers may make different reference sound-fields at the listener position feasable.

Perhaps I came with too many questions at once to you. :) But I think there are more questions coming...

Best,
Viktor

[Jacro]

James,

You're right that it's a bit difficult to discuss these kinds of very involved subjects on a forum. But then I think the purpose of a forum discussion may not be to reach scientific-quality conclusions but to have a good entertaining discussion. And for me it is always rewarding to discover ways of thinking that I haven't considered before. I do admire people such as yourself and others who are willing to contribute their great knowledge in these discussions.
•••••••••••
I understand your standpoint in defining spatial neutrality, I think it makes good sense. A spatial aspect of sound perception that I believe is very difficult to reproduce though using a stereo system is the room "ambience", the natural ambience of the listening room will add to the recorded ambience which cannot really be reproduced naturally from only two points in space in front of the listener. So that's where I think there is some room for different listener preferences in choosing the type of ambience and degree of liveliness of the listening room.
•••••••••••
It seems to me that it would not be very profitable for most companies in the audio business to use their resources for researching the kind of intricate low-level subjects that you are talking about in this thread.
•••••••••••
How have you found time for this?
•••••••••••
I really like your notion that only a subset of the information in the original sound field needs to be reproduced to yield a realistic experience, and that knowledge of this subset is really vital in designing the reproduction system. It is certainly obvious that a stero-recording can only store a small part of the information in the original sound field, but can still give a very pleasurable experience when reproduced over a good system.
This also explains that different use-models of loudspeakers may make different reference sound-fields at the listener position feasable.
•••••••••••
Perhaps I came with too many questions at once to you. :) But I think there are more questions coming...

Best,
Viktor

Victor,

I agree, that the purpose of a forum like this is to entertain, be entertained, and have fun. I’m always trying to balance the tedious rigor of research with the fun that is the real reason we are all involved in this joyful endeavor of music and audio.

•••••••••••

You are correct with your concern about the room ambience issue.

I refer to it as the “Second Venue” problem. Whenever we have a larger “First Venue”, where the live performance takes place, and then attempt to reproduce that performance in a smaller “Second Venue”, the “Second Venue” dominates and overrides many of the ambience and spatial cues of the larger original venue.

This is a major issue, perhaps THE major remaining issue to resolve, (only equaled by the need for an effective, consistent recording capture standard).

Decades ago, when Acoustic Research did their famous live vs. recorded demonstrations they always did them outdoors, eliminating the venue differential problem.

We can solve the ‘Local Variables’ at the loudspeaker quite well, and have been able to do so for a long time, but the ‘Global Variables’ of the interactive environments and sound field around the head of the listener, are still problems that are generally left unresolved in the industry.

Most loudspeaker system types have the room as an enemy. One must find ways to have the loudspeaker either ignore the room (possible, but difficult with today’s technology) or to work with the room.

Loudspeakers that work with the room are usually boundary coupled, such as Carlsson, Larson, Ino-Audio, (the clue), Allison Acoustics, early Snell, early Audio-Pro, and very few others (a surprisingly small percentage of the total loudspeaker market). But, coupling to the first boundary is merely a first step toward the total solution required, and only a few loudspeaker purveyors design their products to resolve this issue and also very few provide any guidelines to allow the user to maximize their system to address the “Second Venue” issue.

This area is ripe with opportunity for manufacturer and end-user alike.

•••••••••••••••••••

First, I would have to respectfully disagree with you that it is not profitable for companies to use their resources to research these subjects, or to at least leverage the research that has been done by others. It is my belief that once these things are well understood, one can build much better products for much lower cost. While I enjoy attempting to advance the state of the art, it is even more exciting to be able to provide state of the art performance at a low price.

Ultimately, quality can be achieved in a cost effective manner if efficient principles are employed. One of my favorite concepts is the idea that the goal of any product should be that it is; “so advanced, that it is inexpensive”. When one understands the issues thoroughly, one can focus the expense only on the attributes that result in significant sonic improvements, and reduce cost in all other areas.

In terms of the products in the marketplace, I believe the problem is not a lack of technology. It is a problem of marketing choices/priorities and lack of understanding of what solutions are available today.

As the science fiction writer William Gibson has stated,
“The future is here, it just isn’t well distributed yet”.

•••••••••••••••••••••••

In terms of your question of how I have found time for this, I would have to say that I have been fortunate to have the freedom to pursue what I most enjoy, and get paid for it (The secret I never tell my employers, is that I would probably do it for free ☺ ).

I’ve done much research on my own, and I’ve had the good fortune to be well funded and provided with good facilities by different organizations over the years.

But, basically, I am too curious to stop working these things, whether I’m at work or during my spare time.

••••••••••••••••••••••

In terms of a “stereo” recording only storing a small part of the information in the original sound field, you are correct, but I think the standard methods of “stereo” recording is a vastly underutilized application of two-channels. Stereo is limited, but two-channel information storage does not have to be nearly as limited as current “stereo”.

I think there is an opportunity to standardize a more advanced form of two-channel spatial/timbral perceptual coding that could significantly advance the effectiveness of two-channel reproduction systems. Unfortunately, the generic stereo recording process is a rather haphazard practice that is substantially void of scientific guidelines.

If we merely utilized the state of current psychoacoustic knowledge to standardize the recording process, such that playback system designers had a form of Rosetta-stone codex to define playback system calibration, we could all be enjoying lifelike reproduction in our homes today.

Once again, to paraphrase, “The future is available to be here now, it just isn’t properly organized to be well distributed.”

•••••••••••

Don’t worry, more questions are welcome…as long as there is patience with my answers.

•••••••••••

All the best,

- James

[roren]

Jacro,

You have learnt me a lot about speakers. An other thing that I admire is the
way you handle all of our questions.
I have experienced this before, and it has been from guys like you with a practical knowledge.
To mention a few of them I have had this experience with Ingvar Öhman and a mr Robert C White. You are all guys with real knowledge.
Maybe this is what makes the different between a bussines man and a scientist.

Just wanted to say that.

Rolf

[Jacro]

Jacro,

You have learnt me a lot about speakers. An other thing that I admire is the
way you handle all of our questions.
I have experienced this before, and it has been from guys like you with a practical knowledge.
To mention a few of them I have had this experience with Ingvar Öhman and a mr Robert C White. You are all guys with real knowledge.
Maybe this is what makes the different between a bussines man and a scientist.

Just wanted to say that.

Rolf

Rolf,

Thank you for your very kind words.

It is an honor to be included in the group of Mr. White and Mr. Öhman, both of which have been very generous with their substantial knowledge of audio.

Warm regards,

- James

[paa]

Jacro,
Whith rooms and room boundries being so important for sound quality, what shape of listening room would you recommend, if one were free to build it new?
Would non-parallell walls be of a significant advantage, or are there other things more important to consider?

[Jacro]

I believe that head movements are much more useful in a real acoustic event than in a 5.1 setup. In the 5.1 case it is much more likely that head movements tend to reveal the point sources of the rear speakers rather than to achieve a more exact sound space experience, especially if one does not have multiple rear speakers like in movie theatres.
Also these headphones, I do believe, only try to keep the sound field oriented towards the image, and are not useful for improving the ears ability to pick up exact directions of projected phantom sounds.

Hi paa,

Yes, head movements allow the ear-brain to have increased ability to accurately analyze the acoustical aspects of a any sound field. That is better for improving perceptual accuracy in live events, such as mapping sound fields at musical concerts and accurately detecting the direction of a growling tiger.

But, as you surmise, the hearing enhancement from head movement actually makes our job more difficult in developing perceptually accurate sound reproduction systems, in that the greater hearing acuity derived from head movements helps the ear-brain systems catch us trying to deploy our sound field trickery, such as phantom images, and allows the ear brain to detect the actual sound sources (loudspeaker locations).

As an example, one of the tricks we can deploy is to change the spectral response to create height information from a loudspeaker. When the head is still, it depends on the frequency response above 4 kHz due to pinnae effects to determine sound source height. Our perceptual accuracy for identifying the height of a sound source is not very good compared to our excellent ability in accurately detecting tiny differences in horizontal sound sources. But, with head tilting, we can engage our horizontal source analysis ear-brain system in detecting height information more accurately, by tilting our head to the side, making one ear higher than the other, which gives us improved information about the height of an image, overriding the poorer performing spectral analysis of height information.

You are also correct to suggest that these head-movement sensing headphones are limited making stereo recordings sound more realistic. But, with pure, binaural recordings they can be very impressive.

Cheers,

- James

[Jacro]

Jacro,
Whith rooms and room boundries being so important for sound quality, what shape of listening room would you recommend, if one were free to build it new?
Would non-parallell walls be of a significant advantage, or are there other things more important to consider?

paa,

After trying many exotic room shapes over the years, a basic rectangular room form seems to be consistently the best and most predictable.

Strongly angled walls do not eliminate low frequency standing waves (just moves them in frequency in a fairly unpredictable manner) but slightly angled walls or paintings that are hung in a manner that are angled down slightly, centered at ear level can minimize higher frequency slap echo. Rooms with more than six sides can have unpredictable effects (good or bad) and peaked ceilings can have deleterious effects, sometimes causing focusing energy back at the listener, instead of diffusing the acoustic energy.

Generally, the larger the room the better, due to the fact that the first to second venue distortion is reduced and the transition to the modal region is moved to a lower frequency.

Of course there are other details relative to optimizing room acoustics, but in terms of your specific question, I would recommend the basic room shape to remain substantially six-sided and rectangular.

- James

PS - One exception is that of certain non-standard, non-rectangular room designs can be effective if the loudspeaker is truly part of the room, such as large waveguide or horn systems built into, and as part of the sound room, such that the room is an extension of the loudspeaker itself.

[paa]

I thought I was the first to invent it back in 1985, but it turns out that Bobby Beaver at Altec Lansing had developed products based on the concept at least 10 years earlier with the Model 814A “Extenda-Voice” and was granted a patent (US 3,722,616).

http://www.lansingheritage.org/html/alt ... rs/814.htm

Later work was done in 1982 by POLAR-PRO in Finland (WO8401681) and Skip Cross in America (US 4,437,541), and also in 2001 by Noselli in Italy, EP1137318A2.

Most recently, the technique can be seen in the Gradient Helsinki 1.5 loudspeaker, used as the midrange baffle:
http://www.gradient.fi/helsinki15/

Best regards,

- James

How is the width of such a device related to the cut off frequency?

[Jacro]

I thought I was the first to invent it back in 1985, but it turns out that Bobby Beaver at Altec Lansing had developed products based on the concept at least 10 years earlier with the Model 814A “Extenda-Voice” and was granted a patent (US 3,722,616).

http://www.lansingheritage.org/html/alt ... rs/814.htm

Later work was done in 1982 by POLAR-PRO in Finland (WO8401681) and Skip Cross in America (US 4,437,541), and also in 2001 by Noselli in Italy, EP1137318A2.

Most recently, the technique can be seen in the Gradient Helsinki 1.5 loudspeaker, used as the midrange baffle:
http://www.gradient.fi/helsinki15/

Best regards,

- James

How is the width of such a device related to the cut off frequency?

paa,

In its basic form, approximately 1/2-wavelength, similar to a dipolar device.

But, since it has the rear waveguide as an additional variable, one can vary the front-to-back output ratio, allowing significant modification of directivity and also changing cut-off frequency to some degree.

- James

[paa]

Thanks for the explanation.
But it seems strange that the Gradient Helsinki 1.5 can have as low crossover frequency as 200 Hz between the cardioid mid and the bass, considering the mid housing only seems to be approximately 300 mm in diameter.
Can there be any other trick in there, to lower the cut off frequency?

[Jacro]

Thanks for the explanation.
But it seems strange that the Gradient Helsinki 1.5 can have as low crossover frequency as 200 Hz between the cardioid mid and the bass, considering the mid housing only seems to be approximately 300 mm in diameter.
Can there be any other trick in there, to lower the cut off frequency?

Yes, there are a number of "tricks", and Jorma Salmi, the developer of the Helsinki, is a very capable loudspeaker engineer.

Steps to extend Di-Cardioid low frequency cut-off:

1) The output through the rear waveguide is greater than the axial energy from the front of the device, so it is most often best to use damping material in rear waveguide to equalize the front and rear amplitudes. By doing so, the effective internal transit length of the waveguide will increased such that the 'virtual' dimension of the waveguide will be effectively increased by a small amount.

2) By setting the resonant frequency of the transducer at a predetermined frequency below the cut-off frequency of the waveguide, and adapting the motor strength, suspension compliance and cone mass to provide an under-damped transducer 'Q' of between 1.1 and 2.0 at the resonant frequency, one can extend the LF response of the driver/waveguide combination by about at least a half-octave, a full octave in some cases. (This approach is used with many dipole systems to extend their response below baffle cut-off.)

3) At or near the crossover frequency, one can under-damp the high-pass filter that drives the waveguide transducer creating some narrow-band gain in exchange for a dip in the impedance curve. If one observes the separate amplitude responses of the woofer and the midrange and the impedance curve of the Gradient Helsinki loudspeaker, it will be seen that there is an impedance minimum at 250 Hz and a corresponding peak (about 2 to 3 dB) in the amplitude response of the high-pass section at the same frequency, suggesting the use of an underdamped high-pass filter, boosting the waveguide response enough to achieve an acoustic crossover at 230 Hz.

By applying these techniques, it one can see how the Helsinki achieves the extension of its 575 Hz baffle cutoff frequency down to the 230 Hz crossover frequency.

There are other subtle aspects that are incorporated into the Helsinki to smooth the midrange response of the Di-Cardioid waveguide.

A centrally mounted transducer on a circular Di-Cardioid waveguide will have ripple in the on-axis response at each half-wavelength multiple. By mounting the midrange transducer off-center in the waveguide, and tilting the waveguide back, such that it is pointing upward somewhat, the ripple in the response is distributed more evenly, and the waveguide still operates down to substantially the same cut-off frequency.

I hope this adds some clarification about how to use these types of devices.

Cheers,
- James

[Kraniet]

I think the topic has been dealt with to some extent in the thread. But I still would like to hear a little more about your views on the shape of the frequency curve of the speakers. Its something that have been discussed quite alot on this forum. Mr Öhman calls it "stereo-system compensation", sayin that the "psychoacoustical correct frequency response isnt the "straight line" that so often is discussed.

Youve mentioned earlier that most companies just do an arbitrary "BBC-dip".

But Toole and harman (Revel) say that the curve should be straight as an arrow with an even fall in dispersion towards high frequencies. Is this just because their evaluation process is mono only?

One example is



Being a schematic picture it lacks any dB-reference. But the general idea is visible.

What are your views on this? Do you have a particular "target curve" that you aim for? Do "the clue" have any "compensations" for a better resolution in regards toward the "stereo-system fault"?

Are there any measured curves out (of the clue) on the web that you know of or could you even provide some yourself?

[Jacro]

I think the topic has been dealt with to some extent in the thread. But I still would like to hear a little more about your views on the shape of the frequency curve of the speakers. Its something that have been discussed quite alot on this forum. Mr Öhman calls it "stereo-system compensation", sayin that the "psychoacoustical correct frequency response isnt the "straight line" that so often is discussed.

Youve mentioned earlier that most companies just do an arbitrary "BBC-dip".

But Toole and harman (Revel) say that the curve should be straight as an arrow with an even fall in dispersion towards high frequencies. Is this just because their evaluation process is mono only?

One example is



Being a schematic picture it lacks any dB-reference. But the general idea is visible.

What are your views on this? Do you have a particular "target curve" that you aim for? Do "the clue" have any "compensations" for a better resolution in regards toward the "stereo-system fault"?

Are there any measured curves out (of the clue) on the web that you know of or could you even provide some yourself?

Kraniet,

This is an interesting topic. There are many different frequency ranges that can be purposefully adapted and they all interact with each other, so once again, analyzing and applying any correction factors in isolation from other interactive parameters has limited usefulness. But, for the purposes of this discussion, we’ll limit ourselves to elements relating to the band of frequencies that relate to spectral shifts due to two channel cross-talk compensation.

First, I do have a target curve, but it is a three dimensional set of power response curves that also include other interactive variables. And, ( the clue ) does incorporate a specialized response to address the "stereo-effect".

I believe that Toole and Voeks remain fixed on a simplified response characteristic, partly, due to the fact that they appear to want to sustain a non-specific, generalized use model that a large company like Harman maintains so as to not limit customer interest. I believe Toole has provided a good service to the audio industry, but it is aimed more towards how to optimize standard configuration loudspeakers applied in the most general arrangements.

If one participates in their loudspeaker comparison test room, it becomes obvious that their approach is biased towards standard free standing loudspeakers, and a boundary coupled (or many other innovative types of systems) cannot be effectively evaluated within their test setup). It is unfortunate, in that they have tremendous resources, but their focus in on how to optimize standard devices, not how to create a new, state of the art device that would involve any new level of packaging, room coupling, and radiation pattern novelty.

While specialized frequency responses are useful, they can only be advantageous if a specific use-model is deployed. Otherwise, any modification to the frequency response will be heard as merely an undesirable coloration.

That said, while the phantom image creation works well with most listeners (not all listeners hear phantom images) the “stereo-effect” has a number of well known, undesirable elements. The unintended crosstalk of the sound from the right loudspeaker arriving at the left ear and mixing with the left loudspeaker output causes both a spatial limiting effect and a spectral comb filtering that form linear distortions of the original signal.

This can either be ignored or addressed, but to address it requires a balanced, organic approach with all aspects interacting to achieve a correction without causing further perceptual errors. Some of what I do is proprietary, but I am glad to share some of the basic concepts that are involved in providing a corrective frequency response.

A simplified portion of one solution is, first, to optimize the listening angles, and polar amplitude responses to achieve the minimal inter-aural cross-correlation (IACC). With the loudspeaker/listener referenced for optimal IACC minimization, then the second step is to observe the remaining cross-correlated mixing and to apply a correction curve.

If one attempts this correction curve without first minimizing IACC by way of establishing loudspeaker/listener angles, the correction curve will have to be too severe, and tonal coloration from the correction will be heard.

If one takes a single sound source and move it around the head, the balance of minimized IACC without introducing too much spectral distortion, happens at about 21 degrees.

Once this is established, then one can introduce a correction curve of partial inverse ripple to reach a perceptually neutral balance.

By partial, I’m referring to both the amplitude of the ripple and the number of center frequencies in the series that one addresses.

I’ll give a little specificity here; I address the first three frequencies of the series with a weighting function that is descending as the amplitude is adjusted from the fundamental, to the second and third ripple harmonic (barely touching the third). Any harmonics above the third tend to be variable due to environmental and human interface issues and also, at those frequencies, one is starting to influence the height perception pinnae response, which is a separate, but interrelated, issue.

This is a one dimensional view of the issue, in that as one applies these types of corrections, one has to account for other influential variables depending on radiation pattern type and change with frequency, and multi-driver interactions and their influence at different angles, etc.

One aspect is that the induced correction ideally varies with angle, such that the off axis response ideally has the corrective curve fade out, in that one doe not want the reverberant response to include a dominant correction curve.

This isn’t a comprehensive answer (impossible without the context of a given set of variables), but hopefully one that shows a direction to explore for those who are interested in doing so.

Each optimization of this type may seem subtle to some listeners, but as one combines all of the techniques, to achieve a systems approach, they result in an organic whole that provides a musical perception that can be quite special, elevated from the standard, but “optimized for general use” loudspeakers that are derived from standard practice.

These approaches are particularly advantageous with boundary-coupled loudspeakers and/or loudspeakers that also minimize early reflections.

First, these types of systems can remove many of the extraneous acoustical arrivals.

Secondly, some claim that a boundary-coupled loudspeaker has reduced perception of front to back depth. While, that can be true for some boundary-coupled designs that don’t incorporate the full potential of the design approach, a well-designed directivity controlled, boundary coupled device will create a sound field that has greater front to back depth, by moving the frontal portion of the image forward. This is a natural effect of minimizing IACC and also reducing early reflections.

In general, I’d suggest that one pursue the conventional approaches, because the diligence and rigor it takes to balance all the parameters, once one attempts to truly optimize for all the system variables, such as the weakness of stereo, is rather tricky.

But, if one is willing to be adventurous, and diligent, the potential is for a sonic presentation that is very special.


Cheers,

- James

[Kraniet]

very interesting and thank you for finding time to answer these questions.
I assume that these corrections can only be evaluated properly in a stereo setup?

Im currently in the process of constructing a crossover myself. And I find that corrections of one or two dB arent noticeable when listening to just the one speaker. Next step is to buy more parts and do the optimisation on the stereo pair.


Would you say its best to start from a prefectly flat frequency response and introduce the different "tweaks". Or could a start with a rough "BBC-template"? If so how would this shelf look like?
I realiaze you cant disclose specifics but any pointer would be welcome.

Would you say that some (or all) of these compensations relate to "personal taste" or would you call them "universally true"?

By the way its a wall hung speaker with a angled baffle to get those (approx)22 degrees that Im building. Im gonna use it with a 10cm thick absorber on the inner side of the speaker. But I still get a pretty substantial dip at 350Hz even with the absorber.
What are your thoughts on this dip (I assume the clue get a similar dip) and how did you deal with it?

Heres a imigae of the measurement.(taken, laying on the ground, outside with a good distance to other objects)

[Nattlorden]

In 1979 Robert Carver created a modified version of the JVC unit and offered a processor recalibrated for stereo recordings called Sonic Holography.

All were attempts to fix some of the fundamental spatial and tonal flaws of 2-channel stereo. Theoretically, under ideal conditions, it can be a superior approach, but is much more critical to optimize than conventional stereo. While potentially better than stereo, if it isn’t perfectly calibrated it can sound much worse than stereo. Ultimately, it is best suited for reproducing a binaural based recording.

These types of systems were always problematic with widely spaced loudspeakers, because the cancellation signals were very difficult to match, as they had to include the frequency response effects of the sound rapping around the face, to the opposite ear.

I'm actually going to give this a go... Have one of the later Carver processors on it's way by mail... just for the heck of having tried that too. Fun thing that you mentioned it here. My curiousity was actually sparked by listening to an interview with Bob... and being one thing that I haven't tried before, I felt I needed to give it a try even if I'm well aware that it most likely turn out to be a waste of money.

[Jacro]

very interesting and thank you for finding time to answer these questions.
I assume that these corrections can only be evaluated properly in a stereo setup?

Im currently in the process of constructing a crossover myself. And I find that corrections of one or two dB arent noticeable when listening to just the one speaker. Next step is to buy more parts and do the optimisation on the stereo pair.


Would you say its best to start from a prefectly flat frequency response and introduce the different "tweaks". Or could a start with a rough "BBC-template"? If so how would this shelf look like?
I realiaze you cant disclose specifics but any pointer would be welcome.

Would you say that some (or all) of these compensations relate to "personal taste" or would you call them "universally true"?

By the way its a wall hung speaker with a angled baffle to get those (approx)22 degrees that Im building. Im gonna use it with a 10cm thick absorber on the inner side of the speaker. But I still get a pretty substantial dip at 350Hz even with the absorber.
What are your thoughts on this dip (I assume the clue get a similar dip) and how did you deal with it?

Heres a imigae of the measurement.(taken, laying on the ground, outside with a good distance to other objects)

Kraniet,

You are welcome.

Many loudspeaker colorations can be evaluated on a mono basis, and even within the current topic of tonal correction for stereo cross-talk linear distortion” much can be done by listening to on a monophonic, single loudspeaker basis (preferably mounted with the same spatial relationship angle to the listener as the final use model) or dual loudspeakers with mono program, to evaluate tonal aberrations independent of spatial effects, but the final evaluation must be done in the actual, use model, dual loudspeaker mode.

Generally I recommend starting from a flat axial and well behaved off-axis response as a first step, to isolate any system resonances, driver interactive effects, diffraction issues, or other a problems before optimizing the system power response. There are many exceptions, wherein the transducers are purposely designed to have a built in ‘effect curve’.

The goal of an compensations I am discussing here are an attempt to address universal objective issues, not for the purpose of subjective personal taste enhancements.

If one were to consider adjust a loudspeaker to personal taste, I would recommend getting everything objectively correct first, so that any preference distortions that one adds are easily adjusted and identifiable, instead of just being a pleasant artifact/distortion of the loudspeaker that is left in tact. This way, one can control the system more easily, having full understanding of what is being added, with the ability to get back to an objective baseline.

If you are seeing this dip single ground plane, outdoor measurements, it may be misleading. A single boundary reflection is often difficult to eliminate completely, but the room has many boundaries, and if the loudspeaker is purposefully placed for an optimal ratio front wall, floor, and sidewall distances (and ideally, ceiling distance also), then the boundary reflections will average out in a manner that will substantially eliminate any severe dips as you observed in your outdoor measurement.

Applying absorbent to the front and sidewalls will help further, as will some form of absorbent or diffusor on the floor. All of these tactics used together will tend to smooth the ripple to an acceptable level.

Only having a single reflection boundary is rather artificial. In fact, in some room situations, if you were to eliminate all the reflections 100%, except having one strong reflection remaining, it would sound worse than if you had all the reflections distributed in an effective manner, such as establishing golden ratio (~1.6) secondary path-length ratios.

Correlated reflections are psycho-acoustically the most disturbing (those arriving equally to both ears, such as floor reflection, front wall reflection, and ceiling reflections) particularly if just one dominates.

Make sense?

- James

[IngOehman]

I think the topic has been dealt with to some extent in the thread. But I still would like to hear a little more about your views on the shape of the frequency curve of the speakers. Its something that have been discussed quite alot on this forum. Mr Öhman calls it "stereo-system compensation", sayin that the "psychoacoustical correct frequency response isnt the "straight line" that so often is discussed.

Youve mentioned earlier that most companies just do an arbitrary "BBC-dip".

But Toole and harman (Revel) say that the curve should be straight as an arrow with an even fall in dispersion towards high frequencies. Is this just because their evaluation process is mono only?

One example is



Being a schematic picture it lacks any dB-reference. But the general idea is visible.

What are your views on this? Do you have a particular "target curve" that you aim for? Do "the clue" have any "compensations" for a better resolution in regards toward the "stereo-system fault"?

Are there any measured curves out (of the clue) on the web that you know of or could you even provide some yourself?

Hi everybody!

Just for the protocol, I'd like to clarify that all the problems which I include
in what I like to call "the stereo system flaws" or "the intrinsic flaws of the
stereo system", are problems in MANY different domains.

I.e. both the timbral domain, imaging distortions, dynamic (and pseudo
dynamic (linear behaviours that still reduce dynamic range)) effects and
also time resolution loss effects.

I say this, since I sense in the above quoted, that Kraniet has reduced it
all, or close to everything to being a question of frequency response - and
even to the idea that a single univocal target curve could be "the answer to
what constitutes as a correct compensation".

Nothing can be further from the truth.

All these things are delicately entangled in quite an illusive and convoluted
manner...


I do not like simplifications, and I really do not like to have things I've said
simplified by others - and then still being hold responsible!

I'm not responsible.

(Actually - I'm one of the most irresponsible normalized earthlings that I
know. I am to be trusted about everything - but with nothing.)

- - -

The problems also occur both during "encoding" and "decoding" (recording
and replaying the recording), both are exhibiting differens set of problems,
often cross-depending in intricate patterns, which actually creates quite a
few interesting possibilities to optimize all balances.

Anyway; the possibilities to find and optimize "the stereo system compen-
sations" are no less multifaceted than the originating problem is. And thus
the compensations (optimally applied) are also addressing the behaviours
in ALL of the mentioned domains.

- - -

For practical reasons, I try to avoid going into detailed descriptions of the
problems and solutions on internet foras.

BUT - I believe that I have been very clear (also here on faktiskt.se) that
even if you ONLY look on the timbral problems (ignore ALL the problems
that are manifesting in the other domains) caused by the stereo system
(i.e. by trying to pack the complexity of real life multi dimensional sound
into only two one-dimensional channels) and also ignoring* all radiating
directions other than the one aiming directly at the lister - I still do divide
both the problem and the sollution in 12 differens components - resulting
in quite different 0 degree frequency responce curves, depending on all
other parameters of the loudspeaker.

So there are no single target curve, and I do not see how there can even
be one! The idea of a single univocal target curve goes against everything
I know about the inner workings of our hearing.

What can be, and I believe is - is a complex equation that can be used to
create a target curve for one specific loudspeaker (assuming the environ-
ment is reasonable predictable).

I use such an equation, but again - there are at least 12 different aspects
of it where 11 are variables! (I include the 12 important enough to play a
role that is larger than the uncertainties.)

And again - the timbral part of the stereo system flaws are only a minor
part of the everything that is caused by the stereo system.

- - -

I just wanted so say that, so that no one attributes the idea to me, of the
stereo system flaws possible corrections being nothing but a target curve
in the timbral domain.

This said, I hope that no one misunderstands what I just wrote to be an
attach aimed against the BBC-dip, or the men behind it. To the contrary,
to my knowledge, the BBC's studies ware amongst the first in the word
addressing problems of the stereo system itself, and though being just a
fraction of a fraction of the truth, it does not diminish the fact that a first
step is often the most important - since it leades the way by pointing out
the direction.


Best regards, Ingvar

- - - - -

PS. Please excuse my English. Writing in English make me feel like my
head is full of brake fluid. Well, not that I've tried that...


*When I spoke about ignoring other radiation directions, I did not really
mean ignoring it (it is a vital part of the equation) only ignoring having it
adjusted separately - as a simplification, to point out that it is still very
complicated, even after such a simplification.

In real life however, engineering loudspeakers, I do no such simplifications,
but try to control everything and give each parameter the properties I like
it to have to work well in the application.

Not just "see what I got" and then try to do the best of it in regards of the
stereo system corrections.

[IngOehman]

PPS. I do not agree that using +/- 21 degrees (minimum IACC-points)
minimizes the intrinsic flaws of the stereo system. Actually, it depens on
the relation between the listening distance, the directivity index of the
loudspeakers and the acoustic radius of the room - in my opinion/from
my experience.

I suggest a somewhat wider opening angle of +/- 23 degrees for front
row listeners.

Two interesting positiv bi-effects from such a set up, is:

1. That the problem of some of the stereo system flaws, peaking at +/-
19-20 degrees, are counteracted/compensated by the higher degree of
reflexted sound further back in the listening room.

2. That the holofonic soundstage will be less widht-modulated by the
change of distance to the soundstage opening (the loudspeaker line), i.e.
creating a much more stable staging, wich corresponds much better to a
true sound source in "the room behind the loudspeakers".

3. That compatibility to the silver screen standards (+/- 22,5 degrees in
directors seat) are better.

[IngOehman]

PPPS. Sorry for the kidnapping of this thread.

I just ment to do one posting, to clear up what was said regarding a target
curve - mentioned together with my name, that I by accident happened to
see. But NOW I'm done!


/iö

[Jacro]

I think the topic has been dealt with to some extent in the thread. But I still would like to hear a little more about your views on the shape of the frequency curve of the speakers. Its something that have been discussed quite alot on this forum. Mr Öhman calls it "stereo-system compensation", sayin that the "psychoacoustical correct frequency response isnt the "straight line" that so often is discussed.

Youve mentioned earlier that most companies just do an arbitrary "BBC-dip".

But Toole and harman (Revel) say that the curve should be straight as an arrow with an even fall in dispersion towards high frequencies. Is this just because their evaluation process is mono only?

One example is



Being a schematic picture it lacks any dB-reference. But the general idea is visible.

What are your views on this? Do you have a particular "target curve" that you aim for? Do "the clue" have any "compensations" for a better resolution in regards toward the "stereo-system fault"?

Are there any measured curves out (of the clue) on the web that you know of or could you even provide some yourself?

Hi everybody!

Just for the protocol, I'd like to clarify that all the problems which I include
in what I like to call "the stereo system flaws" or "the intrinsic flaws of the
stereo system", are problems in MANY different domains.

I.e. both the timbral domain, imaging distortions, dynamic (and pseudo
dynamic (linear behaviours that still reduce dynamic range)) effects and
also time resolution loss effects.

I say this, since I sense in the above quoted, that Kraniet has reduced it
all, or close to everything to being a question of frequency response - and
even to the idea that a single univocal target curve could be "the answer to
what constitutes as a correct compensation".

Nothing can be further from the truth.

All these things are delicately entangled in quite an illusive and convoluted
manner...


I do not like simplifications, and I really do not like to have things I've said
simplified by others - and then still being hold responsible!

I'm not responsible.

(Actually - I'm one of the most irresponsible normalized earthlings that I
know. I am to be trusted about everything - but with nothing.)

- - -

The problems also occur both during "encoding" and "decoding" (recording
and replaying the recording), both are exhibiting differens set of problems,
often cross-depending in intricate patterns, which actually creates quite a
few interesting possibilities to optimize all balances.

Anyway; the possibilities to find and optimize "the stereo system compen-
sations" are no less multifaceted than the originating problem is. And thus
the compensations (optimally applied) are also addressing the behaviours
in ALL of the mentioned domains.

- - -

For practical reasons, I try to avoid going into detailed descriptions of the
problems and solutions on internet foras.

BUT - I believe that I have been very clear (also here on faktiskt.se) that
even if you ONLY look on the timbral problems (ignore ALL the problems
that are manifesting in the other domains) caused by the stereo system
(i.e. by trying to pack the complexity of real life multi dimensional sound
into only two one-dimensional channels) and also ignoring* all radiating
directions other than the one aiming directly at the lister - I still do divide
both the problem and the sollution in 12 differens components - resulting
in quite different 0 degree frequency responce curves, depending on all
other parameters of the loudspeaker.

So there are no single target curve, and I do not see how there can even
be one! The idea of a single univocal target curve goes against everything
I know about the inner workings of our hearing.

What can be, and I believe is - is a complex equation that can be used to
create a target curve for one specific loudspeaker (assuming the environ-
ment is reasonable predictable).

I use such an equation, but again - there are at least 12 different aspects
of it where 11 are variables! (I include the 12 important enough to play a
role that is larger than the uncertainties.)

And again - the timbral part of the stereo system flaws are only a minor
part of the everything that is caused by the stereo system.

- - -

I just wanted so say that, so that no one attributes the idea to me, of the
stereo system flaws possible corrections being nothing but a target curve
in the timbral domain.

This said, I hope that no one misunderstands what I just wrote to be an
attach aimed against the BBC-dip, or the men behind it. To the contrary,
to my knowledge, the BBC's studies ware amongst the first in the word
addressing problems of the stereo system itself, and though being just a
fraction of a fraction of the truth, it does not diminish the fact that a first
step is often the most important - since it leades the way by pointing out
the direction.


Best regards, Ingvar

- - - - -

PS. Please excuse my English. Writing in English make me feel like my
head is full of brake fluid. Well, not that I've tried that...


*When I spoke about ignoring other radiation directions, I did not really
mean ignoring it (it is a vital part of the equation) only ignoring having it
adjusted separately - as a simplification, to point out that it is still very
complicated, even after such a simplification.

In real life however, engineering loudspeakers, I do no such simplifications,
but try to control everything and give each parameter the properties I like
it to have to work well in the application.

Not just "see what I got" and then try to do the best of it in regards of the
stereo system corrections.

Hello Ingvar,

I both agree and empathize with you regarding the concern and danger of having ones forum comments reduced to oversimplifications.

The multi-dimensional aspects of effective endeavors in loudspeaker engineering are most often far more complex that what can be reduced to a few paragraphs, but, we are a social animals with a shared passion, so we embark on a dialog about the things we care about.

It seems that maybe the best one can do in this type of forum is to entertain, bring some joy and to provide incomplete ideas that hopefully at least point in a direction that empowers others to be a little more effective at discovering their own answers….and in the mean time, hope that one’s statements aren’t terribly misconstrued as they are passed on and repeated.

Anyhow, for my part, I appreciate your use of English and tolerating the sensation of a brake fluid filled head.

All the best,

- James

[single_malt]

I do not like simplifications, and I really do not like to have things I've said
simplified by others - and then still being hold responsible!

I'm not responsible.

Yes you are. By writing lots of incoherent portions of information often concluded with sentences like "at least 12 different aspects"(without further explanation) and "All these things are delicately entangled in quite an illusive and convoluted manner... "
No wonder people start tryning to make their own interpretations to get the whole picture understandable..

[UrSv]

If anybody is incoherent it is you. Again and as always. Could you please stop bringing your personal agenda och stalking IngOehman out of this thread please? We have a competent and sharing international contributor here and it would be nice of you not to soil this thread with your ever repeating attacks on IngOehman. Grow up.

[stefanolo]

I do not like simplifications, and I really do not like to have things I've said
simplified by others - and then still being hold responsible!

I'm not responsible.

Yes you are. By writing lots of incoherent portions of information often concluded with sentences like "at least 12 different aspects"(without further explanation) and "All these things are delicately entangled in quite an illusive and convoluted manner... "
No wonder people start tryning to make their own interpretations to get the whole picture understandable..

Here we (you) go again ...speaking of predictable and provocative manners....As you might have noted Jacro, and I figure a lot of others (not occupied by harassing IngOehman whenever the possibility occurs), sympathize with Ingvar in not allways explaining every aspect of a technical issue by explaining every atom involved.

I hope IÖ doesn't respond to your provocation and this thread goes down the drain.


Sorry for OT

/stefan

PS. Jacro, I very much appreciate you beeing around and giving your views on these very complicated matters even though I'm not technically educated at all. Thanks!

[Kraniet]

Kraniet,

You are welcome.

Many loudspeaker colorations can be evaluated on a mono basis, and even within the current topic of tonal correction for stereo cross-talk linear distortion” much can be done by listening to on a monophonic, single loudspeaker basis (preferably mounted with the same spatial relationship angle to the listener as the final use model) or dual loudspeakers with mono program, to evaluate tonal aberrations independent of spatial effects, but the final evaluation must be done in the actual, use model, dual loudspeaker mode.

Generally I recommend starting from a flat axial and well behaved off-axis response as a first step, to isolate any system resonances, driver interactive effects, diffraction issues, or other a problems before optimizing the system power response. There are many exceptions, wherein the transducers are purposely designed to have a built in ‘effect curve’.

The goal of an compensations I am discussing here are an attempt to address universal objective issues, not for the purpose of subjective personal taste enhancements.

If one were to consider adjust a loudspeaker to personal taste, I would recommend getting everything objectively correct first, so that any preference distortions that one adds are easily adjusted and identifiable, instead of just being a pleasant artifact/distortion of the loudspeaker that is left in tact. This way, one can control the system more easily, having full understanding of what is being added, with the ability to get back to an objective baseline.

If you are seeing this dip single ground plane, outdoor measurements, it may be misleading. A single boundary reflection is often difficult to eliminate completely, but the room has many boundaries, and if the loudspeaker is purposefully placed for an optimal ratio front wall, floor, and sidewall distances (and ideally, ceiling distance also), then the boundary reflections will average out in a manner that will substantially eliminate any severe dips as you observed in your outdoor measurement.

Applying absorbent to the front and sidewalls will help further, as will some form of absorbent or diffusor on the floor. All of these tactics used together will tend to smooth the ripple to an acceptable level.

Only having a single reflection boundary is rather artificial. In fact, in some room situations, if you were to eliminate all the reflections 100%, except having one strong reflection remaining, it would sound worse than if you had all the reflections distributed in an effective manner, such as establishing golden ratio (~1.6) secondary path-length ratios.

Correlated reflections are psycho-acoustically the most disturbing (those arriving equally to both ears, such as floor reflection, front wall reflection, and ceiling reflections) particularly if just one dominates.

Make sense?

- James

Yes it makes sense. One can really understand how the "established" HiFi-brands arent into these kind of compensations.
But I find it to be a pity aswell. Feels I bit of a shame that companies like B&W or Revel spend all that money on designing new high-tech transducers and put them in advanced boxes with "high-end" expensive crossover parts only to have a sub-optimal crossover that doesnt make the best of the stereoflaw.
On the other hand those companies seem to have the best ability to incorporate the right compensations. Would you say the concept of the stereoflaw is to difficult a matter for "normal" people(even HiFi-nuts)? As always I assume its the salespeople that dictates what can be done.

Regarding the compensation curve I posted above Im avare that it isnt as simple as that. And I apologize if thats how it seemed. I just brought it in as an example of how it could look like. Being a generalist I tend to speak in general terms even when I know it isnt as easy as that. :)

I guess these compensations are percieved differently if its a big speaker or a small. Different placements of the speaker in the room would also affect this. So a firm advice on placement should be a minimal when it comes to loudspeakers, wonder why the industry are so afraid of that..

What are your feelings when it comes to multichannel sound? Are these compensations still good to have or should they look different (not exist even)?
What are your views on multi channel sound in general. Reading Tooles book I got the expression that he seem to think multi channel are the way to go forward? Seeing how difficult it is to make a stereo pair work in an optimal way, it feels daunting to have 5 or more speakers with the "right" compensations applied. Or maybe it isnt all that critical?
The lack of an stardard when mixing surround sound it might be hard to generalize?

[Jacro]

Kraniet,

You are welcome.

Many loudspeaker colorations can be evaluated on a mono basis, and even within the current topic of tonal correction for stereo cross-talk linear distortion” much can be done by listening to on a monophonic, single loudspeaker basis (preferably mounted with the same spatial relationship angle to the listener as the final use model) or dual loudspeakers with mono program, to evaluate tonal aberrations independent of spatial effects, but the final evaluation must be done in the actual, use model, dual loudspeaker mode.

Generally I recommend starting from a flat axial and well behaved off-axis response as a first step, to isolate any system resonances, driver interactive effects, diffraction issues, or other a problems before optimizing the system power response. There are many exceptions, wherein the transducers are purposely designed to have a built in ‘effect curve’.

The goal of an compensations I am discussing here are an attempt to address universal objective issues, not for the purpose of subjective personal taste enhancements.

If one were to consider adjust a loudspeaker to personal taste, I would recommend getting everything objectively correct first, so that any preference distortions that one adds are easily adjusted and identifiable, instead of just being a pleasant artifact/distortion of the loudspeaker that is left in tact. This way, one can control the system more easily, having full understanding of what is being added, with the ability to get back to an objective baseline.

If you are seeing this dip single ground plane, outdoor measurements, it may be misleading. A single boundary reflection is often difficult to eliminate completely, but the room has many boundaries, and if the loudspeaker is purposefully placed for an optimal ratio front wall, floor, and sidewall distances (and ideally, ceiling distance also), then the boundary reflections will average out in a manner that will substantially eliminate any severe dips as you observed in your outdoor measurement.

Applying absorbent to the front and sidewalls will help further, as will some form of absorbent or diffusor on the floor. All of these tactics used together will tend to smooth the ripple to an acceptable level.

Only having a single reflection boundary is rather artificial. In fact, in some room situations, if you were to eliminate all the reflections 100%, except having one strong reflection remaining, it would sound worse than if you had all the reflections distributed in an effective manner, such as establishing golden ratio (~1.6) secondary path-length ratios.

Correlated reflections are psycho-acoustically the most disturbing (those arriving equally to both ears, such as floor reflection, front wall reflection, and ceiling reflections) particularly if just one dominates.

Make sense?

- James

Yes it makes sense. One can really understand how the "established" HiFi-brands arent into these kind of compensations.
But I find it to be a pity aswell. Feels I bit of a shame that companies like B&W or Revel spend all that money on designing new high-tech transducers and put them in advanced boxes with "high-end" expensive crossover parts only to have a sub-optimal crossover that doesnt make the best of the stereoflaw.
On the other hand those companies seem to have the best ability to incorporate the right compensations. Would you say the concept of the stereoflaw is to difficult a matter for "normal" people(even HiFi-nuts)? As always I assume its the salespeople that dictates what can be done.

Regarding the compensation curve I posted above Im avare that it isnt as simple as that. And I apologize if thats how it seemed. I just brought it in as an example of how it could look like. Being a generalist I tend to speak in general terms even when I know it isnt as easy as that. :)

I guess these compensations are percieved differently if its a big speaker or a small. Different placements of the speaker in the room would also affect this. So a firm advice on placement should be a minimal when it comes to loudspeakers, wonder why the industry are so afraid of that..

What are your feelings when it comes to multichannel sound? Are these compensations still good to have or should they look different (not exist even)?
What are your views on multi channel sound in general. Reading Tooles book I got the expression that he seem to think multi channel are the way to go forward? Seeing how difficult it is to make a stereo pair work in an optimal way, it feels daunting to have 5 or more speakers with the "right" compensations applied. Or maybe it isnt all that critical?
The lack of an stardard when mixing surround sound it might be hard to generalize?

Kraniet,

I understood that the curve that you showed was merely to provide a generalized example to give context for discussion. I did not see it as you reducing it all to just frequency response.

I am sorry if you felt criticized for bringing forth a good question, which you did in a clear and useful manner. You should always feel free to come forth with questions that are meaningful to yourself.

••••••••••••

As you suggest, sales and marketing often tend to drive the product decisions more than the engineers. The general aspects of the stereo flaw are not difficult to convey or understand, but to comprehend all the complex interactive details and what to do about them is a much more complex issue.

Two of the main impediments are; 1) there is a wide range of recording processes, making it difficult to coordinate with all of them, and 2) the desire to not make loudspeaker use models too specific, so that they can fit into more homes without impacting the sensitivities to interior design preference.

In terms of #1, relative to an ideal angular loudspeaker alignment, I would have to say that, unfortunately, there is not just one ideal arrangement. Numbers like 21 degrees or 23 degrees have limited meaning without carefully defining all the variables.

Due to the variation in encoding process, the recording may be encoded by the pure definition of binaural, or by any one of the almost unlimited permutations of what is called the “stereo” recording process. Again, it is only one part of the picture, but one variable that is not static, but rather dynamic, depending on the recording/playback processing, is that of optimal listening angle, which is reduced to a very low number for binaural recordings played through a crosstalk cancellation system and intermediate number for “stereo” recordings with an imbedded level of crosstalk cancellation, to a bit wider for standard stereo, still wider for certain mixes that include center channel. The particular angle numbers can be defined more precisely as one more clearly defines the myriad of associated parameters.

••••••••••••

I always find it remarkable that while a first rule for marketing is to have a product that is distinctive, with advantages over the competition, but at the same time most marketing departments will balk at anything that appears too different or unusual. They want to be differentiated and better, while at the same time remaining safely similar to the competition.
Strange.

•••••••••••••••

Your last sentence speaks to the truth of one of the most significant problems with all forms of recording/playback process.

I’ll take a risk and state that my current feeling is that multi-channel sound (meaning more than two-channels) as it stands today, and the way it is most often applied, has little value for anything other than home-theater applications, unless one is just looking for merely adding artificial sensationalism to a music program.

With home theater recordings there are at least dubbing stage standards that allow one to substantially replicate what the recording engineer is hearing at their seat in the mixing stage and for that mix also to have some correspondence between one’s experience at the movie theater and in one’s own home theater.

But for music, it is currently a rather confused state of affairs.

Mr. Toole may be correct in stating that it is the way forward, but at the very least, we must move very far forward from where we are today, before it reaches the level of becoming “the way”.

There is good work being done in this regard, but it is still in its formative stages, technically, and even more so, politically.

That said, I remain optimistic about what we can achieve.

All the best,

- James

[IngOehman]

Regarding the compensation curve I posted above Im avare that it isnt as simple as that. And I apologize if thats how it seemed.

No problems from my horizon, as long as I, mentioned by name as I
was, is allowed to clarify that I do not see it that way, nor have ever to
my kowleadge, presented it as merly a timbral property.

Actually, it is a good thing that you bought it up, since one persons view
of a matter, is often shared by others - and this way I was given a chance
to clarify that I have never spoken about phenomenons in one dimension
only.

Ana again, I'm happy as long as I get the chance to clarify that I do not
want to be hold responsible for something I have neither thought nor
said. So thank you for that chance!

I just brought it in as an example of how it could look like. Being a generalist I tend to speak in general terms even when I know it isnt as easy as that. :)

That is fine by me, as long as I'm not hold responsible.

And again I like to remind everyone that the compensation does not only
have a lot of different reasons and is therefore different in different speak-
ers - it is not a mere timbral question either.

I guess these compensations are percieved differently if its a big speaker or a small. Different placements of the speaker in the room would also affect this. So a firm advice on placement should be a minimal when it comes to loudspeakers, wonder why the industry are so afraid of that..

I believe it is a combination of three things:

1. These are relativly small errors compared with most of the actual prob-
lems that has to be dealt with just to arrive at an uncompensated speaker
of descent quality.
So if I had to suggest an order of priority, the compensation of the stereo
system flaws would come in, long after matters like controlling distortion,
dispersion, dynamik linearity, and having the speaker function in a room.
And already these are more than a handful to deal with. So in my opinion,
very few loudspeakers are suffering most from "not having stereo system
compensations".

2. The main part of the speaker industry has not even reflected over mat-
ters of stereo system flaws at all, and the very small part that has - are
either unclear over their own views on it (due to an open mind in combo
with their lack of studies on it), or they do not agree with others.
So there is really no general answer available, that can be applied for any-
one who "is not afraid" to apply it.
Finding a sollution fitting ones philosophy in all other aspects of reproducing
sound timbrally and in space and time, can take decades of research.

3. Some loudspeaker manufacturers are, due to their specific philosophical
ideas on how to communicate the recorded information to the listeners,
not really in NEED of applying any compenatios at all - since they do not
manufacture loudspeakers that works in a manner that will make the flaws
appear anyway, so there is really no need for compensations!
My view on that is, good for them!

What are your feelings when it comes to multichannel sound? Are these compensations still good to have or should they look different (not exist even)?

What are your views on multi channel sound in general. Reading Tooles book I got the expression that he seem to think multi channel are the way to go forward? Seeing how difficult it is to make a stereo pair work in an optimal way, it feels daunting to have 5 or more speakers with the "right" compensations applied. Or maybe it isnt all that critical?

The lack of an stardard when mixing surround sound it might be hard to generalize?

I understand that the above questions are not directed to me, so I say
no more.


Vh, iö

[IngOehman]

Hi everybody!

Just for the protocol, I'd like to clarify that all the problems which I include
in what I like to call "the stereo system flaws" or "the intrinsic flaws of the
stereo system", are problems in MANY different domains.

I.e. both the timbral domain, imaging distortions, dynamic (and pseudo
dynamic (linear behaviours that still reduce dynamic range)) effects and
also time resolution loss effects.

I say this, since I sense in the above quoted, that Kraniet has reduced it
all, or close to everything to being a question of frequency response - and
even to the idea that a single univocal target curve could be "the answer to
what constitutes as a correct compensation".

Nothing can be further from the truth.

All these things are delicately entangled in quite an illusive and convoluted
manner...


I do not like simplifications, and I really do not like to have things I've said
simplified by others - and then still being hold responsible!

I'm not responsible.

(Actually - I'm one of the most irresponsible normalized earthlings that I
know. I am to be trusted about everything - but with nothing.)

- - -

The problems also occur both during "encoding" and "decoding" (recording
and replaying the recording), both are exhibiting differens set of problems,
often cross-depending in intricate patterns, which actually creates quite a
few interesting possibilities to optimize all balances.

Anyway; the possibilities to find and optimize "the stereo system compen-
sations" are no less multifaceted than the originating problem is. And thus
the compensations (optimally applied) are also addressing the behaviours
in ALL of the mentioned domains.

- - -

For practical reasons, I try to avoid going into detailed descriptions of the
problems and solutions on internet foras.

BUT - I believe that I have been very clear (also here on faktiskt.se) that
even if you ONLY look on the timbral problems (ignore ALL the problems
that are manifesting in the other domains) caused by the stereo system
(i.e. by trying to pack the complexity of real life multi dimensional sound
into only two one-dimensional channels) and also ignoring* all radiating
directions other than the one aiming directly at the listener - I still do divide
both the problem and the solution in 12 different components - resulting
in quite different 0 degree frequency response curves, depending on all
other parameters of the loudspeaker.

So there are no single target curve, and I do not see how there can even
be one! The idea of a single univocal target curve goes against everything
I know about the inner workings of our hearing.

What can be, and I believe is - is a complex equation that can be used to
create a target curve for one specific loudspeaker (assuming the environ-
ment is reasonable predictable).

I use such an equation, but again - there are at least 12 different aspects
of it where 11 are variables! (I include the 12 important enough to play a
role that is larger than the uncertainties.)

And again - the timbrel part of the stereo system flaws are only a minor
part of the everything that is caused by the stereo system.

- - -

I just wanted so say that, so that no one attributes the idea to me, of the
stereo system flaws possible corrections being nothing but a target curve
in the timbral domain.

This said, I hope that no one misunderstands what I just wrote to be an
attach aimed against the BBC-dip, or the men behind it. To the contrary,
to my knowledge, the BBC's studies ware amongst the first in the word
addressing problems of the stereo system itself, and though being just a
fraction of a fraction of the truth, it does not diminish the fact that a first
step is often the most important - since it leads the way by pointing out
the direction.


Best regards, Ingvar

- - - - -

PS. Please excuse my English. Writing in English make me feel like my
head is full of brake fluid. Well, not that I've tried that...


*When I spoke about ignoring other radiation directions, I did not really
mean ignoring it (it is a vital part of the equation) only ignoring having it
adjusted separately - as a simplification, to point out that it is still very
complicated, even after such a simplification.

In real life however, engineering loudspeakers, I do no such simplifications,
but try to control everything and give each parameter the properties I like
it to have to work well in the application.

Not just "see what I got" and then try to do the best of it in regards of the
stereo system corrections.

Hello Ingvar,

I both agree and empathize with you regarding the concern and danger of having ones forum comments reduced to oversimplifications.

The multi-dimensional aspects of effective endeavors in loudspeaker engineering are most often far more complex that what can be reduced to a few paragraphs, but, we are a social animals with a shared passion, so we embark on a dialog about the things we care about.

So true.

It seems that maybe the best one can do in this type of forum is to entertain, bring some joy and to provide incomplete ideas that hopefully at least point in a direction that empowers others to be a little more effective at discovering their own answers….and in the mean time, hope that one’s statements aren’t terribly misconstrued as they are passed on and repeated.

Your words are wise, but I do not like to rely on hope.

So if possible, I always try to communicate not only the answer to an
isolated small question (in a large complex), but also the "complexity of
the whole" - if so only by reminding everybody of it, with a few words.

It is not always enough.

But then again, there is often neither time nor room for more.

Anyhow, for my part, I appreciate your use of English and tolerating the sensation of a brake fluid filled head.

Well what can I say - you can try it by switching to Swedish.

All the best,

- James

Same to you, James.


Vh, iö

[paa]

James, when you try to get a rectangular box like The Clue as coupled as possible to the wall, how much absorption would you prefer behind it? What size and thickness would you consider minimum and optimum?

As a minimum I recommend 75mm thick acoustic foam that is arranged on the wall, inward from the right loudspeaker (going towards the left loudspeaker), starting at the inside corner of the cabinet that is closest to the wall, and also starting at the bottom edge of the cabinet.

The panel should be about 60 x 60cm, extending about 25cm above the top of the cabinet.

Ideally, additional amounts and placement would be adapted to the needs of a particular environment.


For a more complete/optimal arrangement, as a general rule, I would suggest a 75mm thick, 150cm x 150cm square piece centered horizontally behind the cabinet, and starting at the floor.

Also, place 100mm thick acoustic foam outward from the speaker, (to the right of the right loudspeaker) over to the corner, and from the corner extend 1/3 the length of the sidewall, OR, a piece 1-meter wide centered on the point where the first specular reflection bounces off the sidewall to the listener.

Again, each room has it’s own issues to deal with, but these are the general guidelines for minimum and optimal absorption.

Let me know if I need to provide further explanation.

Cheers,

- James

I may be thinking slowly, but now I wonder if the more full version of room damping you suggested here will make the loudspeaker more forward radiating, almost like cardioid speakers
Maybe the bass loss from cardioid boxes can thus be considered unnessesary?
Would you like to comment on this?

[vr6jos]

Nu finns ett par till salu. Någon rik herre kan väl köpa bara för skojs skull och recensera!?!

http://www.hifimagasinet.com/annonsDetalj.asp?annonsID=53611