The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Your opinion about room correctional software Virtual Instrument Plugins
Old 18th June 2010
  #91
SAC
Registered User
 

Unfortunately there are significant problems for doing it at only one position as well, as you still cannot mitigate the superposition of MANY non-minimum phase signals from MANY non-linear paths modified by MANY non-linear incidences by modifying ONLY the direct signal in advance of the superposition of MANY time variant energy packets.

And your system for analyzing an extremely complex diffuse semi-reverberant sound field for later arriving reflections while processing their 3 space vector components and addressing the necessary interaural cross correlation components sufficient to modify each and every one of them by manipulation of ONLY the direct signal and maintaining the integrity of the direct signal is to be performed how? By an FIR or IIR filter????? Please! The processing does not account for the non-minimum phase superposition let alone a feasible method for the complex analysis of the soundfield and ALL of the component variable to begin with! We lack this practical ability in non-realtime measurement to sufficiently process this data into readily usable form now.

This would be most difficult for a static state with an extremely constrained limited finite number of component waveforms - while still fantasizing the maintenance of the integrity of the direct signal while simultaneously modifying the same direct signal sufficient to compensate for a limited number of superposed signals.

I find the claims that "we can do this" fantastic at best. The fact is, it is NOT being done. And the best anyone has been able to do is to evaluate and slightly modify an exceptionally overly simplified set of limited characteristics limited primarily to EQing a signal - as if the flatness of the frequency response was the critical aspect of reproduction - which it is not.

There is a radical gulf between claims and reality that is simply not borne out in either theory nor practice.

And ironically, the primary arena where such attempts are being made is in professional LIVE system processing where latency is indeed a critical factor.

But perhaps more telling, are that the claims of accomplishments do not even begin to adequately address the nature of the acoustical environment prior to processing, let alone account for all of the variable via the processing. And if the presumed environment is not accurate, subsequent processing of a limited and insufficiently accounted for topological territory is not going to improve upon it! I would be ecstatic to hear someone touting such big claims to first focus on simply accurately accounting for the basic topology - and THEN they can manipulate it! And all of this assumes a linear mapping of one topological state to another - without the oh so inconvenient manifolding and folding of the spaces.

Heck, I would be even more ecstatic if, instead of making absurdly extenuating claims, that they would limit themselves to addressing fundamental issues of driver control and propagation - which have not sufficiently yet been addressed. As such, this analogous to a manufacture claiming that they are on the verge of realizing intergalactic travel when they have yet to refine an effective reliable cruise speed control in a car.

Heck, I come from an up to date background in optical image processing where I am extremely cognizant of the extreme challenges in simply filtering out noise sufficient to allow reliable image recognition - not to mention the extreme processing power and amazing neural networks running on distributed massively parallel systems (the RS/6000SP with Lots of high speed switches tying far too many nodes together!) required to implement the very rudimentary approaches to this processing. And what most assume is possible in acoustics is in fact much more complex than the basic functions we were implementing in a reliable fashion requiring systems far beyond what the 'regular' retail sector has ever encountered!

Instead, make me excited with realistic expectations that can be reasonably realized with existing technology in a foreseeable future...not pie in the sky claims based upon models and techniques and methods not yet devised nor brought to fruition based upon real physics models that can accurately be described as sufficient and accurate!

The really humorous reality is that we lack systems that can reliably simply and accurately EQ complex spatial and time variant systems!
Old 18th June 2010
  #92
Lives for gear
 
Rod Gervais's Avatar
 

Quote:
Originally Posted by Brainchild View Post
I feel like pro-physical-treatment people who say, "and then of course you can use room correction software" are just throwing the pro-electronic-room-correction people a bone.
Brain,

what a foolish thing to say........

I and a lot of others here with me (not gonna name you all - we all know who you are) are professionals in the field.

I would not put my reputation on the line for one second for anything I did not believe in.

I throw no one a bone..... nor do I believe for a second that any of my colleagues do.

If I design a room that is within 2 or 3 dB's of being flat - there is no doubt whatsoever in my mind that DSP can make adjustments to smooth that out......

The issue it the people who believe - or present (and there are a sales persons out there presenting this) that DSP can be used in lieu of room treatments. Seeing as this is a recording site - I could give a rats arse what some audiophile wants to do in his living room........ for me it's all about control rooms......... so getting a little bit of polish is not better than nothing in my book............ we want the mixed music to be capable of being listened to across a variety of systems,

Everything in a control room begins with a well designed room that's well treated........ DSP can help with small anomalies if they exists when all is said and done..... Please note that I am speaking only of slight amplitude adjustments in a room without other acoustic imperfections, which is what I strive for when designing a control room.
Old 18th June 2010
  #93
SAC
Registered User
 

Amen!

No one I know who is knowledgeable with regards to the field is against the potential capabilities!
Quite to the contrary! We know all to well the fundamental limitations that all to many are so quick to dismiss (if they were ever actually aware of them to begin with).

If it were possible and realizable now I would be the first to embrace that, or any new tool that actually addresses and accomplishes said goals!
I have NO interest in seeing such technology fail! And I am certainly not a Luddite threatened by said technology!

But I refuse to close my eyes or even squint in the interest of overly simplifying and ignoring the significant obstacles in the path of achieving what so many are so quick to declare both possible and, even more irresponsibly, has already been realized that has in fact not been achieved.

I have spent far TOO many years directly working with systems whose significance is directly responsible for the actual lives of people - where the impetus and significance of such systems assumes one hell of a lot more than simply the recreational convenience of someone in their home or a professional listening environment!

And I am also all too aware of where the current acoustical models are...

DSP has a role to play. But long before it accomplishes what some are so quick to assert that it already does, there needs to be MUCH more done to simply address a few useful applications within the context of driver control, waveguide refinement and manipulation, and basic acurate signal processing.
Old 18th June 2010
  #94
Lives for gear
 

Quote:
Originally Posted by Rod Gervais View Post
...If I design a room that is within 2 or 3 dB's of being flat - there is no doubt whatsoever in my mind that DSP can make adjustments to smooth that out......
........ DSP can help with small anomalies if they exists when all is said and done..... Please note that I am speaking only of slight amplitude adjustments in a room without other acoustic imperfections...
After this most recent post of yours, I still have no reason to think anything other than that you are continuing to humor/patronize the concept of DSP room correction.

I refer to the post you quoted from:

Quote:
As far as I can tell, the only "correction" anyone is willing to claim with certainty will actually solve a problem is speaker response correction. One other possible point of agreement is the electronic removal or significant reduction of low frequencies that are very unevenly reproduced in the room. Apart from those things, is there anything that electronic room correction even does in a "pretty well treated" room?
So, again, just to try once more to be explicit and specific: apart from the agreed-upon methods cited above, what exact DSP process do/can you use to correct what exact room problems?
Maybe I should just ask you to define the terms flat; smooth; anomalies; and amplitude adjustments.
Old 18th June 2010
  #95
SAC
Registered User
 

Where are you going with that?

He, and others, have repeatedly stated that there are other more capable means to address fundamental room acoustical issues.

And while acknowledging the limited abilities of DSP to address limited issues, that in NO way implies that he or anyone else either relies or need rely upon them.

So, exactly what is your point asserting that it is his responsibility or obligation to specify "what exact DSP process do/can you use to correct what exact room problems?"

His, and others, assertions have merely posited the potential for the limited use of such tools. You leap to assume that one necessary does is invalid and does not follow from what has been written.
Old 18th June 2010
  #96
Here for the gear
 

I just got back from a meeting with an extremely skilled DSP-developer. We investigated a lot of measurements and tried to figure out where FIR can be used, and where it should not be used at all.

The initial idea was that we should be able to correct selected problems with FIR by doing an extended analyse to categorize the room issues and feed the convolver with a manipulated response curve. We were not convinced that this could be done, but the idea was interesting enough.

However, what we found out was that we could not even correct the speakers correctly with FIR. The problem was really simple. Even the single driver speakers we tested produced a surprisingly large amount of excess phase (we have a simple method to filter it from the acoustic phase). When correcting with FIR the excess phase component is consequently being threated well on paper, but either it generates loads of pre ring, or we can not properly correct it at all. In both cases it performs less good than an IIR based filter.
Old 18th June 2010
  #97
Lives for gear
 

Quote:
Originally Posted by SAC View Post
So, exactly what is your point asserting that it is his responsibility or obligation to specify "what exact DSP process do/can you use to correct what exact room problems?"
I hope to let Rod reply for himself if he so chooses, but to answer your question, I never made such an assertion; asking a question is different from...that.

At any rate: for all the vague claims that "DSP helps with room correction once the room has been corrected," no one has actually answered the question, "How?" or, "In what way?".

Quote:
Originally Posted by SAC View Post
...while acknowledging the limited abilities of DSP to address limited issues, that in NO way implies that he or anyone else either relies or need rely upon them.
His, and others, assertions have merely posited the potential for the limited use of such tools...
It is true that Rod has himself claimed only the most modest of utility for electronic signal processing, as in:
Quote:
...none of us have taken the position that PEQ...has no value - rather that PEQ is potentially the icing on the cake - after a room is properly treated - and can be used for exactly what you refer to - minor tweaks to clean up the last of what anomalies might exist after room treatments.
This raises the question: What anomalies exist that PEQ can remedy? Is it just the playback system response, or something specifically to do with the room response?

At the same time, Rod has given his wholehearted endorsement to this statement:
Quote:
...[DSP] will not fix room acoustics but it can maximize the performance of a given system in a given acoustic...good room acoustics isn't an alternative to good DSP either. You need [emphasis added] both to maximize the sound quality.
This raises the questions: What is meant by maximizing the performance? What is meant by maximizing sound quality? What do you need DSP for in a good room? If it's just the psychoacoustic perception of more even response, that's one thing; but if it's a question of the actual behavior of the acoustical space, that's something else.
Old 18th June 2010
  #98
SAC
Registered User
 

Once a room is maximally tuned, the response can be evaluated for the frequency response, relative to the phase.

In regions of identified minimum phase behavior (and only in such regions), one can use EQ to moderate the response, and PEQ can be used to precisely make adjustments in the limited regions where they are valid. Several analysis tools have provided exactly such capability for some time and even calculated the complete PEQ setting to enter into the rig.

Thus there are limited cases where it can be used advantageously After more fundamental issues are addressed in the room.
This is already being realized in several high end measurement and analysis platforms.

Where else can DSP be used beneficially?

This has already been addressed earlier in the discussions. A classic application is in the control of the speaker Q and dispersion, as well as the alignment of the various multiple spaced driver sources in time, complete with delay line and FIR filtering and equalization of the power response. Ironically, there are valid uses for EQ and other alignment issues, and this is an area where far too many speakers need this! ~15 years ago, the TAO Saori system came amazingly close to accomplishing aspects of this.

Unfortunately, many of the applications pertinent here are invalid when attempts are made to apply them to non-minimum phase real and virtual sources in the acoustical space.

I would be happy if many would eschew the spectacular and unfounded claims and simply accomplish that for which it can be used productively!!! Doing that would make tuning rooms a much easier proposition! Just imagine, the notion of a well behaved controlled dispersion source featuring a uniformly flat power response spectrum and the advantages it would present when used in a bounded space! Its a shame more aren't focused on achieving that which needs to be done and which has the potential to actually be accomplished!

Now there's a radical concept!
Old 18th June 2010
  #99
Lives for gear
 
Bjorn Omholt's Avatar
 

Hmmm. Are you still with us juicehifi? Do you agree with what SAC and others are saying?
Old 18th June 2010
  #100
Lives for gear
 

How minimum?

Quote:
Originally Posted by Lupo View Post
Correcting minimum phase errors in the speakers is pretty obvious.

...

Quote:
Originally Posted by SAC View Post
Once a room is maximally tuned, the response can be evaluated for the frequency response, relative to the phase.
In regions of identified minimum phase behavior (and only in such regions), one can use EQ to moderate the response, and PEQ can be used to precisely make adjustments in the limited regions where they are valid...
...

juicehifi made the claim that:
Quote:
You can do nonminimum phase time domain & frequency domain correction with FIR filters. Audiolense, Trinnov, DRC and Acourate does that.
And here are some statements you have made either in direct response to that statement or to the general idea it expresses:
Quote:
But one still cannot resolve anomalies by EQ that are the result of the superposition of non-minimum phase signals.

Quote:
But you simply cannot effectively use EQ to remediate anomalies (e.g.: spatial polar lobing/ frequency response comb filtering) introduced by the the summation (superposition) of multiple direct and/or reflected (non-minimum phase) signals.

Quote:
In regions where the actual frequency response amplitude and phase are not related 1:1 by the Hilbert response (meaning that excessive phase' exists and they are not 'minimum phase') - in other words, the resultant superposed response is the result of the combination of two or more signals not aligned with respect to time, then the anomaly CANNOT be effectively addressed by any kind of EQ.

Quote:
And unfortunately, regarding non-minimum phase sources, changing the frequency of the direct signal via EQ, or causing small changes in phase depending upon the specific type of EQ used, affects both the direct and reflected signal and fails to remediate the time based differential existing between the multi-sourced superposed signals and the resultant error.

Quote:
And the notion that the direct signal can be modified such that it retains its 'original integrity' while also being modified to anticipate and correct for the later superposition due to both the original direct signal and any number of subsequent reflections simply has not been established.


Right. In regard to minimum-phase issues, you have said the following:
Quote:
AFTER all system and signal alignment, level adjustments, and room treatment is performed, it is then posible with several tools to ascertain if any of the room response exhibits a relationship of frequency amplitude and phase where the relationship of the various component superposed signals exhibit a one to one relationship (per the Hilbert Transform) and are defined to be 'minimum phase'.

Quote:
Several programs are able to evaluate the relationship of the actual frequency response amplitude and phase responses of the room for a given location with the Hilbert transformed evaluation of said relationship. If there is agreement between the two, they are minimum phase. If not, and "excess phase" exists, they are not minimum phase.

Quote:
Various programs can analyze the response for the presence and 'extent' of exactly such a minimum phase relationships - if such regions exist. Examples are TEF and Room Control. If such regions do exist, then PEQ can be used to help address those particular issues.

Quote:
If such a minimum phase relationship exists, to the extent that such regions are defined and the anomalies fall within these regions, it is possible to apply limited amounts of PEQ.
Quote:
The use of DSP for limited near minimum phase low frequency correction is valid.


These statements raise the questions:
How is it that a minimum-phase net received signal including room reflections can be remediated by alteration of the original output signal such that the net received signal equals the original output signal minus the room response? How non-minimum-phase can the net received signal be and still be remediable in such a way? What's the maximum frequency in a given space and a given net signal frequency-phase relationship for which the approach works? In other words, what is the boundary between what can be addressed by original output signal alteration and what can't?


Old 18th June 2010
  #101
Lives for gear
 
Rod Gervais's Avatar
 

Brain,

perhaps it's time for you to exercise a few brain cells for yourself.....

I find your tone offensive (to say the least) but will humor you for a moment more.......

DSP (Digital Signal Processing) is just that - nothing more nothing less.

If my room has a 3dB spike at the listening position I can handle that with a good Parametric Equalizer.....

By hand I would simply run room analysis software - analyze what's happening at the listening position and lower that particular frequency until I have a flat reading....... it doesn't change my mix - it only adjusts for the "anomaly" that is caused by the room itself........

Well guess what? If I can do this by hand, it can also be done for me.

DSP is used in good automobiles for example. There is a computer analysis of engine noise and road noise performed, an inverse signal to the offending sounds is transmitted from the stereo speakers which cancels the offending sound - all while the system is playing your favorite song. It is called digital noise cancellation.

Granted the system works with a couple of very tight listening locations - usually just for the driver and front seat passenger - making it work for the entire inside of the car (or room) is possible - but requires a lot more speakers in a lot more positions, which is cost prohibitive........

If you are not familiar with any of this technology perhaps you should invest the time and effort to educate yourself - you could do what I and others do - research - read white papers - more research - more white papers - do it for the rest of your life (as we do)........

Or you could attend classes (if research is a wee bit heavy for you) then it is just a time and money investment - but someone else does the legwork for you.

In any case, as I said earlier - I am a professional in the field - I take time off from making money to come here and help people for free.

Neither I nor any of the other people here owe you an explanation - much less an education......

I might occasionally do a speaking engagement (for which I am paid) or I might teach a class (for which I am paid) or I might design a studio (for which I am paid) and the people who pay me have the right to ask me what I am doing and how I might be doing it.........

You are none of those people....... and even if you were - sometimes I even look at them and say "don't worry about it - you wouldn't understand even if I was inclined to explain - and seeing as that is the case - the explaining would be a waste of my time, and I don't waste time even for the people who are paying the bills." Remember I said the right to ask - not necessarily the right to an answer.

Rod
Old 18th June 2010
  #102
Lives for gear
 

Rod, I am genuinely perplexed by your response. After such statements as
Quote:
Originally Posted by Rod Gervais View Post
Brain,
what a foolish thing to say...
and
Quote:
Originally Posted by Rod Gervais View Post
Brain,
perhaps it's time for you to exercise a few brain cells for yourself...
the statement
Quote:
Originally Posted by Rod Gervais View Post
I find your tone offensive (to say the least) but will humor you for a moment more...
is...genuinely perplexing. I haven't intentionally set out to be offensive, and I aplogize sincerely that my tone has come across that way.
Quote:
Originally Posted by Rod Gervais View Post
Neither I nor any of the other people here owe you an explanation - much less an education...the people who pay me have the right to ask me what I am doing and how I might be doing it...You are none of those people...Remember I said the right to ask - not necessarily the right to an answer.
I'm equally perplexed that I am being interpreted as feeling entitled to an answer, or demanding an answer, or anything of that nature. If I am not regarded as even having the right to ask a question, however, I can understand how the asking of a question would be regarded as offensive.

At any rate, let me just say to you and all other audio professionals who read this that I am deeply grateful for every bit of knowledge and every reference to topics of study that you offer. I won't bore you with the details of the degree of self-teaching I have undertaken over the course of the past three years, but I assure you that any lack of comprehension is not for a lack of ongoing effort.

As for the questions themselves, I am trying to be as specific as possible. It is frustrating that my attempt to get to the heart of this matter is met with the kind of resistance, reluctance, evasion and personal attacks with which it has been met.
Old 18th June 2010
  #103
Here for the gear
 

Quote:
Originally Posted by Brainchild View Post
At any rate: for all the vague claims that "DSP helps with room correction once the room has been corrected," no one has actually answered the question, "How?" or, "In what way?".
I thought that was pretty well covered, but a good thing can't be repeated too often.

Quote:
Originally Posted by Brainchild View Post
This raises the question: What anomalies exist that PEQ can remedy? Is it just the playback system response, or something specifically to do with the room response?
Those who represents the peak of knowledge within acoustics would probably never ask that question. In an acoustically treated room, one has to choose between some goods and some bads. For example, the room LF gain can not be adjusted without affecting the decay time in the same passband. If you are willing to compromise a bit, you could adjust the tonal balance with acoustic remedies (because it is far more audible than the decay) but you would then compromise on the decay.

This too goes for the dispersion pattern of the loudspeaker. One could state that it is a part of the playback system, but then why not the room as well? The dispersion pattern is a part of the complex 3D behaviour of the room as much as anything else. If we put two otherwise identical instruments with different dispersion patterns in an acoustically perfect room, we would be able to tell the difference, so a perfect room is not the same as "the coupling between the speaker and room taken care of".

Quote:
Originally Posted by Brainchild View Post
This raises the questions: What is meant by maximizing the performance? What is meant by maximizing sound quality? What do you need DSP for in a good room? If it's just the psychoacoustic perception of more even response, that's one thing; but if it's a question of the actual behavior of the acoustical space, that's something else.
Maximizing performance is always about fine tuning the parameters that are prefered subjectively when listening. A more interesting question would therefore be:
What sounds better, a room that is treated to match tonal balance as a compromize against the decay, or a combination of perfect decay and fine tuned tonal balance?
Old 18th June 2010
  #104
SAC
Registered User
 

Quote:
Originally Posted by Brainchild View Post


These statements raise the questions:
How is it that a minimum-phase net received signal including room reflections can be remediated by alteration of the original output signal such that the net received signal equals the original output signal minus the room response? How non-minimum-phase can the net received signal be and still be remediable in such a way? What's the maximum frequency in a given space and a given net signal frequency-phase relationship for which the approach works? In other words, what is the boundary between what can be addressed by original output signal alteration and what can't?


What is this preoccupation with "frequency"? Its time to enlarge your perceptual frame of reference and to become familiar and conversant with system behavior with respect to time.

We are dealing with TIME based relationships. And it remains that you cannot EQ non-minimum phase signals using FIR, IIR or any other filters.

Based upon analysis of the signal relationships, there exists the possibility, while it is not guaranteed, that regions MAY exist where the phase relationships are minimum phase. IF such defined regions exist, precisely aligned PEQ filters can be implemented to precisely manipulate the gain of the signal within that LIMITED passband.

What regions this applies to is not determined by frequency, but by the phase relationship (to greatly overly simplify things).

-----------------------------------

And now to say something that,while often true, I already regret as it far too often results in absolutely erroneous generalizations and conceptual perversions that almost always constitute the totality of what many remember in such discussions!

OFTEN, but not necessarily so!, in a small home type system, low frequencies below say ~100 Hz are often, but not necesarily!!!, minimum phase and some modes can be moderately EQ'd. But having said this, you MUST understand that there is no NECESSARY condition that renders this true in all cases! It is due to the phase relationship - NOT the frequency!



If you want to delve into the definition and the relationships that define non-minimum phase relationships, you are going to need both the math and the S-plane diagrams. This is not something that lends itself to logical descriptions and gross generalizations,as the relationships are defined in great detail and with great specificity mathematically!

In its MOST basic, a minimum phase system has specific implications regarding how the gain and phase are correlated. In a minimum hase system, every change to the amplitude of the system has a direct and necessary corresponding change to the phase response of the system (and visa versa). to the degree that such modifications are not linked one to one, the result is an accumulation of what is referred to as 'excess phase', and this anomaly accrues with the summation of multiple time shifted signals. This aggregation cannot be remediated with inverse causal filters.

In short, and to provide an explanation sure to not satisfy you without your understanding of the math, one cannot modify a direct signals, that will simultaneously provide for the deliverance of an unmodified direct signal and simultaneously modify the direct signal that serves as the source in anticipation of the multiple unique modifications of the direct signal due to incidence with varying boundary impedance sources that can variously modify the gain, phase, and spectral energy content of the reflected signal.
And now that the direct signal is a source of the unmodified direct signal as well as the source for however many modified reflected sources, they recombine via superpositon in location specific unique time relationships such that the direct signal has been modified to compensate for the complex recombination and summation of the various complex modified signals!

You cannot simultaneously deliver an unmodified direct signal and a significantly modified signal simultaneously sufficient to address the complex issues introduced by the significant complex modifications of reflections derived from the same direct source signal. NOT EVEN IF THE non-minimum phase superposition is between simply the direct signal and just ONE reflection !!! And note, in the real world, ALL of the reflections are sourced a,from the direct signal! You don't just have one! And that incredibly complex system is unique for every point is 3space!!!!

Remember, we are not introducing a additional signal independent of the direct source signal! And you couldn't do that without destructively modifying the direct source as well! - which ironically would defeat the entire purpose of the venture anyways -as we are tying to preserve and optimize this very signal!

We are faced with a Martin Escher like dilemma, where the hand is drawing itself, and one cannot modify the other without invariably altering the source that is trying to be preserved!

But, as it is so quaint to attempt to make others prove a negative, I would love to reverse this discussion. I would like to hear just how one prescribes to solve the problem!~ I am quite tired of the assumption that all sorts of magic works, and that it is up to the Magnificent Randi to debunk it. Its time that others define with great specificity and examples how their 'realized' Holodeck actually works, instead of challenging others to state why the imagined system does not.

The fact is, we are not faced with a plethora of perfectly functioning DSP environments that solve all of our problems leaving us searching for one that does not! Quite the opposite! And while i do not dispute the potential for the application of DSP applications for direct signals, I challenge the ability for DSP to solve complex spatially distributed non-minimum phase issues. In other words, the claims of being able to 'solve' non-minimum phase "room issues" looks great in glossy marketing brochures, but I know of no competent mathematician who makes claim to solving such issues in the real world.

So...what is the real debate? That DSP can be used for limited minimum phase issues? Fine! If that is the issue, I am content to leave on that note, as DSP has many potential uses. Or is the claim that it can solve complex spatially distributed non-minimum phase issues - oft referred to as "room correction"? Other than the misnomer of referring to the equalization of minimum phase relationship regions, or of the use of adaptive digital filters to address ONLY ONE finite point, this is a spurious notion. But unfortunately, technically challenged marketing geniuses seem content to fantasize that this very limited ability warrants the misuse of the more extensive label! Unfortunately that is beyond any math or technology we have now! When that is developed and realized, please call me!

But without such a substantiated assertion, this debate lacks legs.
Old 18th June 2010
  #105
Lives for gear
 

Quote:
Originally Posted by SAC View Post
What is this preoccupation with "frequency"? Its time to enlarge your perceptual frame of reference and to become familiar and conversent with system behavior with respect to time.
We are dealing with TIME based relationships. And it remains that you cannot EQ non-minimum phase signals using FIR, IIR or any other filters.
Darn. I made a mistake in my last post addressed to you in deleting the statement, "I am on the same page as you regarding the adjustment of the net received non-minimum-phase signal via adjustment of the original output signal." It is my very lack of preoccupation with frequency - quite the opposite; I am trying to get my head around the phase issues at the boundary between minimum- and non-minimum-phase region - that is prompting my questions.
Quote:
Originally Posted by SAC View Post
What regions this applies to is not determined by frequency, but by the phase relationship (to greatly overly simplify things)...
...OFTEN, but not necessarily so!, in a small home type system, low frequencies below say ~100 Hz are often, but not necesarily!!!, minimum phase and some modes can be moderately EQ'd. But having said this, you MUST understand that there is no NECESSARY condition that renders this true in all cases! It is due to the phase relationship - NOT the frequency!
This issue is the only reason I included the mention of frequency in the questions you most recently quoted.

Quote:
Originally Posted by SAC View Post
In short, and to provide an explanation sure to not satisfy you without your understanding of the math, one cannot modify direct signals, that will simultaneously provide for the deliverance of an unmodified direct signal and simultaneously modify the direct signal that serves as the source...
And now that the direct signal is a source of the unmodified direct signal as well as the source for however many modified reflected sources, they recombine via superpositon in location specific unique time relationships such that the direct signal has been modified to compensate for the complex recombination and summation of the various complex modified signals!
You cannot simultaneously deliver an unmodified direct signal and a significantly modified signal simultaneously sufficient to address the complex issues introduced by the significant complex modifications of reflections derived from the same direct source signal. NOT EVEN IF THE non-minimum phase superposition is between simply the direct signal and just ONE reflection !!!
Remember, we are not introducing a additional signal independent of the direct source signal! And you couldn't do that without destructively modifying the direct source as well! - which ironically would defeat the entire purpose of the venture anyways -as we are tying to preserve and optimize this very signal!

We are faced with a Martin Escher like dilemma, where the hand is drawing itself, and one cannot modify the other without invariably altering the source that is trying to be preserved!
Thank you. This is exactly what I was hoping someone with the knowledge could verify explicitly.

The mention of the low-frequency-phase relationship in small environments and how that relates to the boundary between minimum- and non-minimum-phase regions has piqued my interest. I am initiating my own investigation [including math!] into the phenomenon, but in the meantime if you might care to expand upon the workings of that mechanism I would be grateful.
Old 18th June 2010
  #106
Gear Guru
 
Ethan Winer's Avatar
 

Lightbulb

Folks, this could be settled conclusively, easily, and very quickly. All the pro-DSP people have to do is measure a room at the listening position before DSP. Then measure again with DSP. Then measure again a foot away with DSP, but without recalibrating the DSP. If the DSP flattens the response and reduces ringing for an area larger than one cubic foot, the pro-DSP people will go a long way toward making their case.

I have been asking pro-DSP people to do this for several years now! But they have never shown data proving that DSP can improve "all seats" as is often claimed. Versus my Audyssey Report that shows the response is made worse just a few inches away. (As in, it's better at one ear but worse at your other ear.) Now, I admit my test is just one test. So let's see more tests! Arguing can go on for years. A test takes only an hour to set up and make a few waterfall graphs.

--Ethan

________________
The Acoustic Treatment Experts
Old 18th June 2010
  #107
Here for the gear
 

DSPs are pretty commonly implemented quite badly.

If you tell a DSP to compensate for the tonal balance that is caused by the room size in the low end, it should be able to do that in more than a cubic foot.
Old 18th June 2010
  #108
Gear Guru
 
Ethan Winer's Avatar
 

Lightbulb

Quote:
Originally Posted by Midgard Audio View Post
If you tell a DSP to compensate for the tonal balance that is caused by the room size in the low end, it should be able to do that in more than a cubic foot.
That's what I hear again and again. But I have never once seen proof. If someone would only offer the proof, it might change a few minds.

--Ethan

________________
The Acoustic Treatment Experts
Old 18th June 2010
  #109
Here for the gear
 

So you are not convinced that if you lower the entire bass range by 3dB, it will be lowered by 3dB in the entire room?
Old 18th June 2010
  #110
SAC
Registered User
 

So you are asserting that it takes DSP to effectively EQ often minimum phase LFs in a room???? Or is it that this is the best that DSP is claiming to be able to do in a room?

I can do that with a standard PEQ!

And that proves DSPs value how? Or does it simply demonstrate that someone is not smart enough to address a modal peak, provided that it is minimum phase and not due factors such as a coupled space, with standard PEQ?

And we keep hearing of DSP addressing "tonal balance" and all sorts of amorphous qualities.

Define exactly what it is addressing!

Why is this so hard! We can measure every aspect of the performance! So let's define exactly what attribute(s) DSP addresses.

And start by describing what attributes it is evaluating in the system and what data is gleaned from this process and how that data is them used to effect a modification in the behavior in the system. This should not be difficult! All of these attributes are understood! We can discuss this in terms of time, frequency, energy content, potential or kinetic energy, the real or imaginary real....whatever.

Yet all I read are claims of amorphous mystical analysis in the time domain or of the impulse response! Whoopee!

Is it just me that has goose bumps at the mention of those 'oh so mysterious' terms that I would guess many marketeers are assuming are a bit intimidating to the average consumer who should therefore assume that this DSP stuff is a mystical otherworldly phenomenon beyond the grasp of mere mortals?

Why is this so hard? I am quite tired of DSP being talked about as if it is some mystical process that affects mystical amorphously defined attributes of a system's response!

I suspect the real problem is that those pushing it have no idea what exactly it is designed to modify and thus instead rely on the 'belief' that 'something' is being corrected. And the more esoteric that undefined 'something' sounds, the more effective it must be! And since it is "DSP (with the Twilight Zone theme playing in the background) that it must be really special and mystical.

Nonsense!

Thus far not one person has made a specific claim. And specific claims are not that "it lowers the bass frequencies 3 dB". I can do that to the direct signal with just several passive components! That's not DSP! That is what a capacitor is for!

I'm sorry, but this has become and remains a faith based claim rather than a scientifically based exercise, and if anything, DSP is more precise in its targeting of attributes than analog - yet nothing has been suggested here that accomplishes more than what an passive analog circuit can do!

Oh, but then we have the amorphous assertions that "it evaluates the impulse response" and it utilizes 'time domain analysis'! OK. Discuss it!

We KNOW time based phenomena! And we well understand the manipulation of signal attributes with respect to time. Yet such references are made on a more elementary level by the 'advanced claims of the DSP marketeers than in any discussion we have had when trying to introduce the concept of the ETC response where some folks were utterly confused by the concept of specular energy having an amplitude and arriving with respect to time!

Hell, I can make a better case for what DSP is capable than do its proponents! But then, perhaps that is why I am much more aware of its limitations as well and get tired of the unsubstantiated overblown hype! (As I am definitely not against the use of properly implemented DSP!)

Yet to date, I have not heard one ACTUAL claim of proficiency or the ability to modify any specific attribute other than EQ (whoopee!) despite all of the brochures claims to evaluate the "room's response" or the impulse response and to make amorphous references to the 'time domain'.

OK. I can measure and display any aspect related to this, from any of the responses in the Domain map to the 3 space composite Heyser spiral with respect to time or frequency!

Yet the DSP boys have yet to qualify a single claim. Heck, they have yet to substantially even discuss how they will modify the easily modified direct signal!

Let alone how they propose to retain the integrity of the direct signal while at the same time modifying the direct signal sufficient to compensate for its recombination (superposition) with multiple first order reflections of the same direct signal, delayed in time that have had their gain, phase and spectral energy content modified by incidence with a boundary of undetermined impedance such that these modifications are appropriately addressed in some meaningful fashion - all the while the direct signal is rendered unmodified!!(and we will ignore for the example the distinct possibility that each boundary surface can have radically different impedance characteristics and modify each reflection is radically different ways!)

Or are our claims suddenly reduced to those of being able to EQ non-minimum phase regions of a rooms response???? If this is all it is, then this technology is much ado abut nothing. As this is both basic and fundamental to what we can do every day with a complete tuning of a space in accordance to accepted acoustical room models that have been subjected to the cream of the worlds foremost acousticians for over 30 years.

Its time to put up or shut up. As amorphous claims of being able to do 'stuff' is utterly inadequate and tells me that those pushing it have far less of an understanding of both how it works and what specifically they are trying to modify then the fancy glossy brochures assembled by marketing folks!

As I said...I am quite tired of this "prove it doesn't work" crap.

When a proponent can adequately describe what it is that is being evaluated, what attributes are being modified, and how these actions are then utilized to productively improve well understood quantifiable attributes - come back and let's talk.

Recent acoustics history is filled with such spurious claims. And this does not do justice to the sincere legitimate DSP design folks who understand its capabilities and its limitations and function responsibly within such an awareness.

Until then, go commiserate with the morons behind the Monster 'frequency balanced' cable claims and those at Bose who, after having their posteriors and claims literally handed to them at the University of Illinois and where Dr Bose attempted to suppress the publication of said measurements by asserting that such should not occur as Bose had donated funds to the underwriting of SynAudCon's functions - only to have Don promptly produce a check book and 'refund' said funds rather then to succumb to such blackmail, that Bose then decided to amp up the marketing machine rather than to refine and modify their flawed direct-reflecting acoustic designs based upon an increased awareness and knowledge of what was actually happening - as they had instead been mistaking an incredibly chaotic distributed polar lobing and comb filtering dominated soundfield that was so 'uniformly bad' that one could not distinguish an optimal response position from a 'bad position - and thus enabling the promotion of such a 'mediocritously' large soundstage as an asset!!

But such are the perils of introducing time based analysis and turning a light onto a previously very dim playing field shrouded in amorphous and spurious marketing clams!

As Ethan suggested, show us something demonstrable and describe what is being done. If valid, we can verify it objectively with the tools we posses. As the plaintive claims that amount to little more literal exhortations of "so you don't believe us" provided as the only assertions of fact demonstrate more than demonstrations of fact. The silence is deafening.

Until then, the loud bark has been rendered into a tiny whine.
Old 19th June 2010
  #111
Here for the gear
 

SAC, I have not read your entire post, but initially you base your post on the fact that someone is supposed to have stated that EQ-ing has to be done in the digital domain to function properly. I have not seen such a statement, and since you clearly direct your fury agains me, I can inform you that I see no reason to separate EQ-ing in digital domain from EQ-ing in analogue domain. However, this thread is about software. As far as I know, software is digital.

The interesting point is what you can do with IIR and what you can do with FIR, and since IIR is based on basic filtering matematics it will do pretty much exactly the same as what you can do with an analogue circuit.

The rest of your post seems to be about everyone working with DSP having no idea what they are dealing with. Pretty useful statement...
Old 19th June 2010
  #112
Gear Guru
 
Ethan Winer's Avatar
 

Lightbulb

Quote:
Originally Posted by Midgard Audio View Post
So you are not convinced that if you lower the entire bass range by 3dB, it will be lowered by 3dB in the entire room?
Sure, if you apply an LF shelf to reduce everything below, say, 300 Hz, then the level below 300 Hz will be lowered everywhere in the room. But that's not what DSP room "correction" does or aims to do! And you'll still have the same (typically) 30 dB disparity between the multiple peaks and nulls.

It's the peaks and nulls that are the problem, and using EQ to flatten the response at one location will not have the same effect elsewhere in the room. Did you read my Audyssey Report I linked earlier? Everything needed to fully understand the issues is in that article. See especially point #1 near the beginning of the article.

--Ethan

________________
The Acoustic Treatment Experts
Old 19th June 2010
  #113
SAC
Registered User
 

You obviously have not read my posts!

First, to address you last point, I have NEVER stated that qualified DSP 'programmers' have no idea of what they are dealing with! My point has been the COMPLETE opposite! They do! (And you obviously have no idea regarding my background!)

The fact is I COMPLETELY respect the process of legitimate DSP! But legitimate DSP processing doesn't attempt to hide behind the veil of mysticism with amorphous glossy brochures and esoteric claims. Legitimate DSP is rather proud to state EXACTLY what attributes are identified and subsequently used to effect a desired outcome! If anything, I respect legitimate DSP too much to watch as too many try to leverage the esoteric image of legitimate DSP processing to make claims they are not willing to objectively explain and verify on many folks all too wiling to believe whatever is printed in a glossy brochure or posted on a fancy website dominated by 'testimonials', but little else.

Unfortunately, regarding "room correction", there is a substantial chasm amounting to a complete disconnect that exists between the marketing claims and the fundamental acoustics!

"Room correction" DSP remains a mystical 'black box' where the physics is totally obscured. But I guess that is consistent with the utterly nonsensical view held by too many that acoustics is unknowable by science - ironically postulated by those who know next to nothing about the science of acoustics!

It is no mistake that not one technical process has been postulated or mentioned for evaluation here! All we get are amorphous claims that it can 'do things'. Wonderful things! And unfortunately, the only magical process I discern is taking the money of those who, by virtue of a foregone rationalization of a large expenditure of money, are convinced that a 'difference' has been made!

And who cares about whether EQ is done in the digital or analogue domain??? I certainly don't! In fact, to jump into this arena more completely, the only advantage digital holds over analogue is that once a signal is in the digital domain, signals are more easily processed with out the complex artifacts that so often plague analogue due to the significant complexity of heterodyning and filtering analogue signals.

Take video processing for example, I would MUCH rather encrypt and decrypt video signals at baseband rather than messing with the incredibly complex process of using sine-wave synch suppression of modulated time variant analogue signal which is almost impossible to do without artifacts!

And to cite DSP's 'amazing'(sic!) ability to EQ MINIMUM phase LF signals? Give me a break! EQing minimum phase signals is a rather trivial accomplishment - and in no way requires investment in DSP (nor digital since you seem concerned about that!) enabled systems.

What is not addressed by the proponents of DSP is the process and procedure for magically providing "room correction" - and this is THEIR term! They refer to the "time domain" and to analyzing the impulse response as if those terms alone are supposed to be sufficient to impress.

But never is it explained what attributes they evaluate those domains or responses for, and how such information is subsequently used! Its as if to do so would trivialize the process and diminish the esoteric accomplishments! An that is nonsense!

What it in fact does, is relegate said claims to one of belief - as there is no legitimate limitation to clearly expressing what is being evaluated and what is subsequently done in terms of the specific processing, and what is accomplished. And every legitimate claim can easily be not only evaluated and verified on the basis of first principles; but the results can be verified as well. And all we get are unsubstantiated subjective emotional claims based on the testimonies of folks - just like the infomercials with folks who claim to have lost 82 pounds in an afternoon or who now make $120K a week from posting recipes in classified ads! And all of which are deemed as insubstantial by the courts, as testimonies from those unqualified to evaluate a system have no legal standing. A coincidence? Yeah....right.

And nothing of the sort is provided either in terms of the logical explanation of the process nor in terms of easily provided verification in terms of readily available time domain testing and analysis.

So before you utterly misrepresent what I have said and posit a red herring in place of my objection, I would suggest that you actually try reading that which you have decided to disagree. But what seems to be consistent is that the disagreement is just as (un)substantial as the many marketing claims for "room correction" DSP, which, not surprisingly, feature the same lack of specificity.



You want to discuss DSP claims? Then have the specific process: from what exactly is evaluated, what attributes are quantified, and how said attributes are then processed such that a strictly defined outcome is achieved. Such specificity is easily verifiable using the same tools so many claim to use to process the system for DSP! In fact, I suspect some of us have more substantial means to do exactly that! In fact, let us know what 'special' tool is required! If such claims are correct, they can be easily and objectively verified. And just think, you won't even need to rely on 'Betty from down the street' to provide a her testimony that the system has reduced the pain of her neuralgia.
Old 19th June 2010
  #114
SAC
Registered User
 

To state once more for clarity.

This 'discussion' is a waste of time if all we do is continually discuss the varieties of 'feelings' about DSP.

That not only misses the point, but its worthless to debate preferences regarding what flavor of ice cream one likes. Not only that, but I am tired of the legitimate purveyors of DSP technologies having their discipline and reputations denigrated by others making fancy claims without substantiation!

So, here are the proposed rules for future discussion:

Put forth the specific goal to be accomplished by the system, the specific method of system evaluation, the specific attributes for which the system is evaluated, the rationale for processing and the specific algorithm - meaning the logical block diagram - for said processing , and what it is that one maintains has been objectively achieved.

Based upon the ability to evaluate the result relative to the goal (just like a logical transfer function!) and using objective state of the art tools available to evaluate the resulting system performance - any claims can be easily and objectively evaluated.

No more feelings, testimonies or assumptions. The degree of effectiveness or ineffectiveness can be objectively verified. And THEN you can subjectively state your feelings as to whether the objective results justify the effort and expenditure.

Until then, I really don't care about how or what one 'feels' about an undefined amorphous 'objective' process which some try to legitimize by the dropping of the term "DSP"..
Old 19th June 2010
  #115
Gear Maniac
 

Quote:
Originally Posted by SAC View Post
You obviously have not read my posts!
Who has that much time on their hands? You haven't exactly mastered the art of brevity. Try training on Twitter heh
Old 20th June 2010
  #116
SAC
Registered User
 

Fair enough.
No one is obliged to read them!

But then again, if one chooses not to, don't then proceed to (erroneously!) tell me what I wrote.
Old 20th June 2010
  #117
Here for the gear
 

Quote:
Originally Posted by Ethan Winer View Post
Sure, if you apply an LF shelf to reduce everything below, say, 300 Hz, then the level below 300 Hz will be lowered everywhere in the room. But that's not what DSP room "correction" does or aims to do! And you'll still have the same (typically) 30 dB disparity between the multiple peaks and nulls.

It's the peaks and nulls that are the problem, and using EQ to flatten the response at one location will not have the same effect elsewhere in the room. Did you read my Audyssey Report I linked earlier? Everything needed to fully understand the issues is in that article. See especially point #1 near the beginning of the article.

--Ethan
I am, as I have mentioned before, working on a room correction system, and I do not want the system to do anything else to the room than to EQ the signal partially due to the remaining peaks and dips at low frequencies.

To work with resonances, one have to work directly with the resonant components, and that is the room geometry.

I also think it is important not to use Audyssey as a proof of the limitations in all other DRC-systems.
Old 20th June 2010
  #118
Gear Guru
 
Ethan Winer's Avatar
 

Lightbulb

Quote:
Originally Posted by Midgard Audio View Post
I am, as I have mentioned before, working on a room correction system, and I do not want the system to do anything else to the room than to EQ the signal partially due to the remaining peaks and dips at low frequencies.
Have you ever set an EQ for the listening position, and then measured what happens 4 inches away, 8 inches away, and so forth? It's always good to have a "proof of concept" before going to a lot of design effort.

--Ethan

________________
The Acoustic Treatment Experts
Old 20th June 2010
  #119
Here for the gear
 

Yes, indeed, I have spent many thoudsands of hours doing manual EQ-ing. It is not easy to get large sweetspots. I normally work with the frequencies that behave more or less identical in the entire measured field.

Most resonance problems could be avoided by better controlling the dispersion. I believe in letting the DSP handle the dispersion control and the room EQ in an integrated way. That is the kind of room treatment (in addition to even out the room acoustically) I have had the best results with so far. Also in acoustically imperfect rooms that technology has provided the best results. It is a bit like comparing horn speakers, dipoles and traditional speakers in the same room to find out what they do to the room acoustics.

This is in a way very interesting. Several large tests have been made to find the relationship between "subjective sound quality" and "objective parameters". What always comes out on top are the frequency response and the dispersion pattern. The rest always comes out way behind the two. One could say this is related to the speakers. One could also say both of them are closely connected to the speaker/room interaction. One is the minimum phase (the frequency response) which is also considered the most audible in those tests, the other is the excess phase (which is just as closely connected to the dispersion control).

This type of tests does, without intending to, go pretty far in suggesting that measured room response is not easy to correlate to subjective quality. This makes sense if identifying things as "being real". If the speakers are replaced by an instrument, there are some pretty basic things that could be different. The frequency response should be matched pretty closely, and that is not very difficult. Late room effects does not affect the experienced tonal balance very much. But the dispersion can put a clearly audible signature to the sound.

I do not hereby suggest that a speaker can mimic the dispersion pattern of all different instruments and vocals, but by reducing the early reflections to those within the speakers +/- 60 degrees and by evening out the dispersion vs frequency, what remains is not only pretty free from added signature, but it is also taking advantage of the usable early reflections that adds to identifying placement, defining room and so on, but since this is within the narrow dispersion field it still mimics most instruments and vocals pretty closely.

So, what I am saying is that comparing response curves is not always the best way to judge sound quality.
Old 20th June 2010
  #120
SAC
Registered User
 

Quote:
Originally Posted by Midgard Audio View Post
Yes, indeed, I have spent many thoudsands of hours doing manual EQ-ing. It is not easy to get large sweetspots. I normally work with the frequencies that behave more or less identical in the entire measured field.
What does this mean? I mean, where or how does this relate to the real world? Frequencies below the Schroeder critical frequency behave modally - and there is no uniform region where "frequencies ... behave more or less identical(ly) in the entire measured field."


And above the Schroeder critical frequency, where the wavelengths are shorter than the boundary dimensions, energy is reflected specularly and comb filtering is the result of combinations of direct and reflected energy superposing! Again, where "frequencies. (DO NOT)..behave more or less identical(ly) in the entire measured field."


While this may sound valid, the statement simply has no basis in fact!


Quote:
Originally Posted by Midgard Audio View Post
Most resonance problems could be avoided by better controlling the dispersion. I believe in letting the DSP handle the dispersion control and the room EQ in an integrated way. That is the kind of room treatment (in addition to even out the room acoustically) I have had the best results with so far. Also in acoustically imperfect rooms that technology has provided the best results. It is a bit like comparing horn speakers, dipoles and traditional speakers in the same room to find out what they do to the room acoustics.


Speaker Q and controlled dispersion have little impact on modal frequencies.

You cannot control room modes by controlling the dispersion of direct signals. LFs still impact upon boundaries and it is due to the relationship of the wavelength and the dimension of the boundary surface that dictates their modal behavior.

So, this means that you are necessarily talking about the behavior of sound above the Schroeder critical frequency where energy is reflected specularly as a result of incidence with room boundaries.

And while speaker Q and controlled dispersion can indeed help minimize the degree of early boundary incidence (e.g. - the principle behind both the LEDE and also of the RFZ as a geometric room configuration used to further facilitate the design of the LEDE by minimizing early reflections and augmenting the ISD), it cannot preclude later boundary incidence.

And with the later boundary incidence, and subsequent superposition of direct with reflected, and with reflected and reflected signals, we have resulting comb filtering and regions of distributed frequency response variation.

If only the sole issue regarding the space were with the direct signal only!

The most prudent way to address such specular energy distribution is a combination of absorption and diffusion.

EQing early reflections arriving within the time interval defining the ISD is meaningless, as the psychoacoustic elements controlling intelligibility and imaging have nothing to do with and frequency variation. Instead it is the time relationship of the direct and reflected signals as described by the Henry Precedence Effect and its corollary, the Haas effect, that determine intelligibility and imaging.

As the behavior of this region is defined by time relationships of the arriving energy, it makes no sense to discuss resolving time domain issues by manipulating frequency - even if this had any meaning. You cannot mitigate intelligibility and imaging by EQ! And you cannot mitigate frequency anomalies by attempting to EQ the superposition of non-minimum phase direct and reflecting signals.

And after the ISD time interval, a well mixed diffuse field - assuming one has been successfully constructed by the use of diffusion, results in a very dense comb filtering that varies at every point in the room! Not only is there no reason to attempt to remediate such densely populated high Q comb filtering, but this behavior is desirable psycho acoustically! It is precisely this behavior we are building via the use of diffusion!

The fact is that a soundfield constituted of sparse focused specular reflections results in widely spaced low Q comb filtering due to the substantial time differentials between the arriving signals. And due to the time differentials, you cannot EQ the comb filtering anyway!

So I would love to have someone tell me exactly how , even if you could physically EQ such a soundfield, exactly what benefit one receives?



Quote:
Originally Posted by Midgard Audio View Post
This is in a way very interesting. Several large tests have been made to find the relationship between "subjective sound quality" and "objective parameters". What always comes out on top are the frequency response and the dispersion pattern. The rest always comes out way behind the two. One could say this is related to the speakers. One could also say both of them are closely connected to the speaker/room interaction. One is the minimum phase (the frequency response) which is also considered the most audible in those tests, the other is the excess phase (which is just as closely connected to the dispersion control).


Yes, the psychoacoustic qualities referred to here are intelligibility and imaging. And a sense of locational focus determined by the strong termination of the ISD as defined in the Henry Precedence Effect and the absence of early arriving signals within the ISD time interval as described in the Haas effect.


The intelligibility is defined in the ISD by the very fact that there are no additional early arriving signals to sum and destroy the intelligibility and imaging! EQ has NOTHING to do with that. We are dealing in this interval ONLY with an the anechoic arrival of the direct signal! And EQ is already readily available to address the direct signal only - but that has
nothing to do with 'room correction', intelligibility or imaging! EQ in this respect is more commonly known as a "tone control"!

And "excess phase" is what differentiates minimum from non-minimum phase signals! And the superposition of such signals (e.g. the direct and reflected signals, CANNOT be resolved by EQ!!!!!


Quote:
Originally Posted by Midgard Audio View Post
This type of tests does, without intending to, go pretty far in suggesting that measured room response is not easy to correlate to subjective quality. This makes sense if identifying things as "being real". If the speakers are replaced by an instrument, there are some pretty basic things that could be different. The frequency response should be matched pretty closely, and that is not very difficult. Late room effects does not affect the experienced tonal balance very much. But the dispersion can put a clearly audible signature to the sound.


Excess gain of specular signals within the later arriving soundfield are EXACTLY what affects tonality and imaging issues! At this point we are not dealing with a controlled Q direct signal at all!!!!! At worst we are dealing with the superposition of either sparse non-minimum phase specular reflections of varying gain, or, at best, we are ideally dealing with a well-behaved exponentially decaying diffuse soundfield defined by a complex diffusion of all sorts of reflected signals, preferably arriving primarily at the listening position from lateral directions. Talking about speaker Q and dispersion at this point is, well, pointless.
Q and dispersion are aspects of the direct signal prior to its incidence with boundaries. And we are dealing with energy 'long' after the boundaries have been encountered!

And replacing speakers with individual instruments??? What?

Considering most recordings are multi-tracked with instruments that may be in different rooms, or for that matter, as multi-tracking is now commonly done 'by wire', they could have been separated by many states and weeks! References to ensemble recreation or whatever is being referred to here makes no sense.

And least of all - what does the frequency response have to do with ANY of this - even if it had any meaning to EQ the interaction of various instruments! Oh wait, I know! You local high school band needs an EQ for the director - that will address sound quality issues...


Quote:
Originally Posted by Midgard Audio View Post
I do not hereby suggest that a speaker can mimic the dispersion pattern of all different instruments and vocals, but by reducing the early reflections to those within the speakers +/- 60 degrees and by evening out the dispersion vs frequency, what remains is not only pretty free from added signature, but it is also taking advantage of the usable early reflections that adds to identifying placement, defining room and so on, but since this is within the narrow dispersion field it still mimics most instruments and vocals pretty closely.


A speaker's role is NOT to replicate the Q and dispersion of individual instruments! In fact, stereo is a device that using the spacing and gain relationships to exploit the Henry Precedence Effect to create a phantom image in between the two discreet sources! And now we are reading of references where the perceived goal is to replace such an illusion with discreetly sourced separate instruments places at some amorphous spot in space- where, who knows, as most recordings are NOT made in ensemble situations where the phase relationships of the individual pieces are preserved! The mix is determined by gain and pan controls on the part of the mix engineer!
And the speakers' role is to accurately reproduce the mix created by the engineer.

Quote:
Originally Posted by Midgard Audio View Post
So, what I am saying is that comparing response curves is not always the best way to judge sound quality.


The problem with this scenario is that it ignores what we know both physically about modal and specular energy behavior. And it utterly ignores the most basic (let alone more subtle) psycho-acoustics which are very well known.


Even if we had the DSP issues down, to be effective, it assumes that we have the basic signal relationships within a bounded space down. In other words, no amount of fancy signal manipulation - least of all EQ, is going to address issues that are fundamentally time based and which are a result of the superposition of non-minimum phase signals that by definition cannot be re-mediated by EQ.


Thus, a valid answer is predicated upon a well-formed question that assumes a valid model of the physical and psycho-acoustic behavior that characterize the space.


And this is simply not present here.


This conversation simply ceases to be one having anything to do with DSP and is instead one dominated by the need to understand the basic time and frequency domain relationships that exist in a bounded space - both in their most destructive relationships and in terms of an optimal relationship where the artifacts detrimental to intelligibility and imaging and tonality are minimized. And EQ, whether DSP based, or simply implemented via passive LRC components is simply NOT the tool that addresses these factors.


My suggestion would be to stop, re-examine, and to become more intimately aware of the behavior of energy in a bounded space that with the benefit of time based analysis and psycho-acoustics has become well understood. And then, based upon a 'sound'(sorry!) understanding of such behavior and relationships, to then proceed to evaluate options that can address such fundamental relationships. But trying to address fundamental issues due to relationships with respect to time by using frequency domain tools such as EQ has been proven ineffective long ago.

----------------

But...if after all is said and done, one still desires to proceed with EQ as the solution to what ails you, this is already possible. One just needs the proper tool.

Programs such as are available on TEF and RoomCapture and SysTune can, evaluate the room for the frequency response and for the corresponding phase at a particular point in the room.. And they can precisely determine if there are regions of minimum phase behavior, and if any exist, the programs will then generate an inverse curve with precise PEQ settings that one can simply program into a quality PEQ unit.

But to put this step into its proper context - this is the VERY last step that is performed on a completed system AFTER ALL other fundamental causal issues are addressed.
Thread Tools
Search this Thread
Search this Thread:

Advanced Search

Forum Jump
Forum Jump