| Home |
| Search |
| Today's Posts |
|
#1
|
|||
|
|||
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
I have a Tact Room Correction System 2.0. It has a resolution of 2 Hz
(below 200 Hz I think, Tact is quite vague about their specs). Tact offers an upgrade for $1000 to a 0.4Hz resolution. The unit needs to be sent to Tact so there is no easy way to listen for the difference. I'm interested in opionions whether the upgrade is worthwile (also considering price). I'd like to get some theoretical background and possibly some practical experience. Thanks, Guido |
|
#2
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
The frequency resolution translates 1:1 to the length of the impulse
response it can produce / correct. If you're talking of 0.4Hz resolution, this translates into a response length of 2.5 seconds. It means the unit will be trying to compensate for room reverberations that may still be occurring up to 2.5s after the event. This is unrealistic. The large number of "bounces" the signal has gone through makes that the sound field is effectively diffuse, even at the lowest bass frequencies (20Hz and barring standing waves). The correction would only be valid for the exact measurement condition / microphone placement. Even just yourself moving in your chair (or wearing different socks) is enough to nullify the correction at such a long time after the impulse. Worse still, the extra "corrections" will simply add to the reverberation, making the sound worse. Realistically you can correct the anechoic response of the speaker and to a lesser extent the first reflection and some low-frequency fundamental modes of the room. Imparting much more precision to the correction unit won't get you any further. Cheers, Bruno "Guido Vogel" wrote in message news:VclIb.699419$Tr4.1745164@attbi_s03... I have a Tact Room Correction System 2.0. It has a resolution of 2 Hz (below 200 Hz I think, Tact is quite vague about their specs). Tact offers an upgrade for $1000 to a 0.4Hz resolution. The unit needs to be sent to Tact so there is no easy way to listen for the difference. I'm interested in opionions whether the upgrade is worthwile (also considering price). I'd like to get some theoretical background and possibly some practical experience. Thanks, Guido |
|
#3
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
I can't see any reason to be concerned with frequency deviation of
only a percent. If you have mis-read the spec and it turns out to be db (in place of Hz), then that is another matter. - Bill www.uptownaudio.com Roanoke VA (540) 343-1250 "Guido Vogel" wrote in message news:VclIb.699419$Tr4.1745164@attbi_s03... I have a Tact Room Correction System 2.0. It has a resolution of 2 Hz (below 200 Hz I think, Tact is quite vague about their specs). Tact offers an upgrade for $1000 to a 0.4Hz resolution. The unit needs to be sent to Tact so there is no easy way to listen for the difference. I'm interested in opionions whether the upgrade is worthwile (also considering price). I'd like to get some theoretical background and possibly some practical experience. Thanks, Guido |
|
#4
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
"Uptown Audio" wrote in message news:H%HIb.699279$HS4.5011373@attbi_s01...
"Guido Vogel" wrote in message news:VclIb.699419$Tr4.1745164@attbi_s03... I have a Tact Room Correction System 2.0. It has a resolution of 2 Hz (below 200 Hz I think, Tact is quite vague about their specs). Tact offers an upgrade for $1000 to a 0.4Hz resolution. The unit needs to be sent to Tact so there is no easy way to listen for the difference. I'm interested in opionions whether the upgrade is worthwile (also considering price). I'd like to get some theoretical background and possibly some practical experience. I can't see any reason to be concerned with frequency deviation of only a percent. Clearly, then, you do not understand one of the most fundamental properties of signals, a concept which is often described as the "time-frequency uncertainty principle." One cannot overemphasize the importance of this concept. Despite that, this and other fundamental concepts are often completely ignored by people in the high-end audio industry, much to their peril and the detriment of those they seek to serve. In its broadest form, the time-frequency uncertainty principle takes the form: Df * Dt = 1/2 This states that the product of uncertainty of frequency (Df) and the uncertainty of time (Dt) can NEVER be less than some fundamental constant. It can be greater, but never less. In this particular case, what this principle tells us is that if we want to know something about a signal for a given amount of time, say, 1 1/4 seconds (Dt), then that sets the minimum uncertainty in frequency (in other words, the frequency resolution). Rearrange the above relation so that: Df = 1/2 * 1/Dt Plugging in our value for Dt (1.25): Df = 1/2 * 1/1.25s Df = 1/2 * 0.8/s Df = 0.4 Hz So, what the "spec" of 0.,4 Hz resolution is telling us is that that is the MINIMUM frequency resolution needed for a correction system that has a time window 1.25 seconds long. Going back to the original "spec," a resolution of 2 Hz, we can calculate what that time window is, again by rearranging the fundamental time-frequency uncertainty relationship: Dt = 1/2 * 1/Df Dt = 1/2 * 1/2 Dt = 1/4 second A system with a resolution of 2 Hz has, AT LEAST, a time window that is 1/4 second wide. |
|
#5
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
"Bruno Putzeys" wrote in message news:QvFIb.698375$HS4.5001008@attbi_s01...
It means the unit will be trying to compensate for room reverberations that may still be occurring up to 2.5s after the event. This is unrealistic. The large number of "bounces" the signal has gone through makes that the sound field is effectively diffuse, even at the lowest bass frequencies (20Hz and barring standing waves). I don't think this is entirely correct. Bass peaks are minimum phase phenomena and can be corrected completely --- they don't diffuse in the acoustic sense (ie. become position independent). Informal studies by Michael Gerzon showed that the precision or length of filter necessary is around 1 second, so I think, if the TacT is implemented correctly, this may be a worthwhile upgrade. Realistically you can correct the anechoic response of the speaker and to a lesser extent the first reflection and some low-frequency fundamental modes of the room. Imparting much more precision to the correction unit won't get you any further. Yes, but we don't know exactly what the TacT unit does internally. It could just correct for bass room modes, and do only speaker correction at higher frequencies (as Gerzon and Craven did). If anyone's interested in Gerzon's writings, here's an archive of some of his interesting work: http://www.audiosignal.co.uk/Gerzon%20archive.html --Andre |
|
#6
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
So you are saying that you can easily distinguish between, say a 31hz
and a 33hz room mode? I can easily see where any tighter tolerance is better, but the question was is it worth $1000. Do your loudspeakers have a frequency response flat to +/- .4hz below 45hz or so? It just seems to me that the loudspeaker and the listener are the weakest link at that resolution, but then I am the ignorant one. Perhaps you could explain in some practical sense how that might be especially useful. - Bill www.uptownaudio.com Roanoke VA (540) 343-1250 "Dick Pierce" wrote in message ... "Uptown Audio" wrote in message news:H%HIb.699279$HS4.5011373@attbi_s01... "Guido Vogel" wrote in message news:VclIb.699419$Tr4.1745164@attbi_s03... I have a Tact Room Correction System 2.0. It has a resolution of 2 Hz (below 200 Hz I think, Tact is quite vague about their specs). Tact offers an upgrade for $1000 to a 0.4Hz resolution. The unit needs to be sent to Tact so there is no easy way to listen for the difference. I'm interested in opionions whether the upgrade is worthwile (also considering price). I'd like to get some theoretical background and possibly some practical experience. I can't see any reason to be concerned with frequency deviation of only a percent. Clearly, then, you do not understand one of the most fundamental properties of signals, a concept which is often described as the "time-frequency uncertainty principle." One cannot overemphasize the importance of this concept. Despite that, this and other fundamental concepts are often completely ignored by people in the high-end audio industry, much to their peril and the detriment of those they seek to serve. In its broadest form, the time-frequency uncertainty principle takes the form: Df * Dt = 1/2 This states that the product of uncertainty of frequency (Df) and the uncertainty of time (Dt) can NEVER be less than some fundamental constant. It can be greater, but never less. In this particular case, what this principle tells us is that if we want to know something about a signal for a given amount of time, say, 1 1/4 seconds (Dt), then that sets the minimum uncertainty in frequency (in other words, the frequency resolution). Rearrange the above relation so that: Df = 1/2 * 1/Dt Plugging in our value for Dt (1.25): Df = 1/2 * 1/1.25s Df = 1/2 * 0.8/s Df = 0.4 Hz So, what the "spec" of 0.,4 Hz resolution is telling us is that that is the MINIMUM frequency resolution needed for a correction system that has a time window 1.25 seconds long. Going back to the original "spec," a resolution of 2 Hz, we can calculate what that time window is, again by rearranging the fundamental time-frequency uncertainty relationship: Dt = 1/2 * 1/Df Dt = 1/2 * 1/2 Dt = 1/4 second A system with a resolution of 2 Hz has, AT LEAST, a time window that is 1/4 second wide. |
|
#7
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
|
|
#8
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
"Uptown Audio" wrote in message news:MUnJb.34929$I07.100932@attbi_s53...
"Dick Pierce" wrote in message ... "Uptown Audio" wrote in message news:H%HIb.699279$HS4.5011373@attbi_s01... I can't see any reason to be concerned with frequency deviation of only a percent. Clearly, then, you do not understand one of the most fundamental properties of signals, a concept which is often described as the "time-frequency uncertainty principle." One cannot overemphasize the importance of this concept. Despite that, this and other fundamental concepts are often completely ignored by people in the high-end audio industry, much to their peril and the detriment of those they seek to serve. In its broadest form, the time-frequency uncertainty principle takes the form: Df * Dt = 1/2 This states that the product of uncertainty of frequency (Df) and the uncertainty of time (Dt) can NEVER be less than some fundamental constant. It can be greater, but never less. So you are saying that you can easily distinguish between, say a 31hz and a 33hz room mode? I can easily see where any tighter tolerance is better, but the question was is it worth $1000. ... It just seems to me that the loudspeaker and the listener are the weakest link at that resolution, but then I am the ignorant one. Perhaps you could explain in some practical sense how that might be especially useful. I think, once again, you simply fail to grasp the significance of the fundamental concepts at hand. If you have a large time window, you MUST have a narrow frequency resolution. The two are unavoidably linked. You also said: Do your loudspeakers have a frequency response flat to +/- .4hz below 45hz or so? Unfortunately, that statement simply does not make any technical sense whatsoever. This is not meant as an insult, to be sure. The response of any system of the sort we are talking about can be viewed as being comprised of three orthogonal dimensions, time, frequency and amplitude. Describing the response of a system requires us to specify at least two of those dimensions, like amplitude vs frequency, amplitude, and so on. Specifying, as you did, frequency vs frequency, gives us NO information about the system whatsoever. Let me suggest an analogy, a complete description of a system requires three orthogonal dimensions, as we mentioned, time, frequency and amplitude. With only one, you know nothing about the system. It's similar to the three orthogonal dimensions of a room, length, width and height (it's a somewhat loose analogy, to be sure, because we do not have determinant transformations from one orthogonal view to another like in signals). Now, if I were to ask you the size of the room, I would expect something like "12 feet wide by 22 feet long by 8 feet high." Or, if I were to ask you for the area of the floor, "12 feet wide by 22 feet long" would be a reasonable answer. However, describing the repsonse of a system as "+-0.4 Hz at 45 Hz" is like saying the size of the room is "12 feet +- 4 inches," It's incomplete. Now, back to our time vs frequency issue. IF you build a system whose time window is t seconds long, that system will have, by the time-frequency uncertainty relation, a certain resolution in frequency proportional to 1/t. Thats just simply a factual consequence of the proprty of the signals. Whether you WANT that resolution, whether YOU think its "necessary" or not, you got it. If you want a system that corrects a time window t seconds wide, you WILL get a system that has a resolution proportional to 1/t Hertz. Period. Look at is another way: if you want a system that does correction in the time domain that is t seconds long, the system MUST HAVE A FREQUENCY RESOLUTION proportional to 1/t Hz. If the frequency is worse than proportional to 1/t, then the time window MUST BE LESS THAN t. You simply cannot get around this fundamental relation. Now, maybe the question you are asking is, does the original poster require correction stretching out several second? That's an entirely different question. If the answer is yes, he does, then whether you want it or not, it IS necessary that such the system have frequency a resolution that it does. Putting it VERY simply, a long time window IS EXACTLY THE SAME as fine frequency resolution. |
|
#9
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
"Uptown Audio" wrote in message news:MUnJb.34929$I07.100932@attbi_s53...
So you are saying that you can easily distinguish between, say a 31hz and a 33hz room mode? And, actually, to hone in on this statement directly, indeed, I can EASILY differentiate betyween 31 and 33 Hz, as that is grater than a semitone. When tuning one of my harpsichords, for example, and can EASILY distinguish to a significantly finer resolution in frequency than this. If, while I'm tuning the instrument, I give myself 10 seconds to listen for the beats, I can distinguish between 31 and 31.2 Hz without any difficulty whatsoever. Indeed, this is just another example of the time- frequency uncertainty principle: if I want to tune to an accurate frequency, I've got to listen for a long time. The more accurate the pitch I want, the longer it takes me to hear the pitch error. And, the converse is, the longert I want to be listening, the finer the frequency resolution I have as a direct consequence. |
|
#11
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
(Dick Pierce) wrote in message ...
(Guido Vogel) wrote in message ... (Dick Pierce) wrote in message ... I can't see any reason to be concerned with frequency deviation of only a percent. In its broadest form, the time-frequency uncertainty principle takes the form: Df * Dt = 1/2 This states that the product of uncertainty of frequency (Df) and the uncertainty of time (Dt) can NEVER be less than some fundamental constant. It can be greater, but never less. ... A system with a resolution of 2 Hz has, AT LEAST, a time window that is 1/4 second wide. So the time window would be at least 1,25 seconds? It looks like the 2Hz spec gives a shorter time window, which sounds like a goal to achieve. As I'm not familiar with this stuff, my interpretation is probably wrong.. The interpretation is quite simple: * Less uncertainty in time, more uncertainty in frequency * More uncertainty in time, less uncertainty in frequency Given the above, I also have a hard time figuring out whether the upgrade might be worthwile (is there a significant audible benefit to gain or will it be minor?). Unfortunately, this is an impossible question to answer, if for no other reason than the very problem itself is effectively an intractible one. All due respects to the people invested in and contributing to the products, it's a very difficult problem and no one in the high end audio business has come even close to solving the problem. The discussion above describes the absolute limits of certainty inherent in the paradigm, it describes why a system has the relationship between two complementary parameters that it does. However, it describes nothing about the quality of the algorithms or how well they are implemented. Extending the time window may contibute to a more complex set of filter coefficients that are harder to generate but may result in a more stable filter with a different level and set of artifacts. So your question comes down more to a case of selecting which partial/imperfect solution has the set of flaws that your more willing to live with. Thanks for your help, Dick. Going from 2Hz to 0.4Hz resolution increases the uncertainty in time. Whether that is desirable or not (how well is the implementation done) can only be evaluated by listening I guess, but that requires 2 Tacts, one with 2Hz and one with 0.4Hz resolution... |
|
#12
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
Hi Guido,
Thanks for your help, Dick. Going from 2Hz to 0.4Hz resolution increases the uncertainty in time. Whether that is desirable or not (how well is the implementation done) can only be evaluated by listening I guess, but that requires 2 Tacts, one with 2Hz and one with 0.4Hz resolution... Although it doesn't help you, can perform the test with a Tact RCS 2.2X. In "2.0" mode (2 channels of correction - mains without subwoofers), the DSPs perform the correction with 0.4Hz resolution. In "2.2" mode (4 channels of correction - mains & subwoofers) the DSPs perform correction at 2Hz resolution. The 2.0S upgrade has been well received by those who have had it performed. It also comes with additional software features like ParEQ. You should search the Tact Audio Users Group on Yahoo for comments on the upgrade. Note: make up your mind soon - the 2.0S upgrade offer expires at the end of January! Regards, Tip |
|
#13
|
|||
|
|||
|
Upgrade from 2Hz to 0.4Hz resolution for Room Correction worthwile?
|
Linear Mode
