Audyssey odessey

[Sir Terrence the Terrible]

For some reason, I have more confidence in your ability to understand logic and reject impossible claims than do you. Let's review Audyssey's explanation as to how it arrives at The EQ filter applied to each speaker:

"MultEQ looks at patterns in the time domain responses and classifies them into clusters based on the similarities in those patterns, typically in 3-5 groups. A representative response is created from each cluster, and a final response is then created from grouping the representatives. That response is then used to create The EQ filter."

Sure, groups are analyzed, weighted and evaluated. Each group creates its own response which later results in a final response used to create The EQ filter assigned to each speaker. The Filter. It. Uno. One per speaker. One possible way for each speaker to influence the soundfield. Asserting that seven (or eight) speakers delivers a unique and specifically directed output to thirty two different places in the room defies logic and reason. You're a recording engineer and don't understand what they're saying?

FAQs

rw

Can you read?

I said this didn't I

The correct wording would be "each position is EQ'd to attain a flat response from each speaker".

Now what part of this don't you understand? You can understand English can't you?

I don't know why you are trying to explain something I already know, maybe you should be explaining this to yourself.

First, I never said anywhere in this thread that one speaker spits out 32 different profiles. I said that you can measure 32 different points, not get 32 different profiles out of one speaker.

[E-Stat]

Can you read?

Perfectly well. I continue to quote your words.

The correct wording would be "each position is EQ'd to attain a flat response from each speaker".

Clearly from their description, that is not what happens. It is true that each position is initially weighted during the analysis, but there is only one EQ calculated for the sum of all groups.

Audyssey does not EQ listening positions that do not need it.

Since each and every position gets the very same averaged EQ Filter, your statement is patently ridiculous. The sound produced from the speaker is not like a Cruise Missile where it can selectively target individual places in the room and not others. Every place in the room gets the same averaged correction from each speaker ! Position A gets the same equalized signal as does position B as does position C as does position D, etc.

Sheesh!

rw

[Sir Terrence the Terrible]

Perfectly well. I continue to quote your words.

The correct wording would be "each position is EQ'd to attain a flat response from each speaker".

Clearly from their description, that is not what happens. It is true that each position is initially weighted during the analysis, but there is only one EQ calculated for the sum of all groups.

And that sum would be a flat or close to flat frequency curve at each seat, just like my words stated.

Audyssey does not EQ listening positions that do not need it.

Since each and every position gets the very same averaged EQ Filter, your statement is patently ridiculous. The sound produced from the speaker is not like a Cruise Missile where it can selectively target individual places in the room and not others. Every place in the room gets the same averaged correction from each speaker ! Position A gets the same equalized signal as does position B as does position C as does position D, etc.

Sheesh!

rw

Ralph, you can see on the screen of your television set what equalization calculation that each measured position gets. So I know for a fact it does not apply equalization to places that already measure reasonably flat. The system would not be useful if it re-equalized a position that already meets the target curve. If you go through the process, you can see how it works. You keep quoting how MultiEQ pro works, but you have yet to make one point about MultiEQ XT32 pro works. Multi EQ XT32 pro has more filters, with more resolution, more horsepower which leads to the ability to provide finer calculations for those many more points. Unlike MultiEQ pro, it does not work in clusters, it works on individual positions. If position A falls into the target curve, it does not re-equalize that position.

[E-Stat]

And that sum would be a flat or close to flat frequency curve at each seat, just like my words stated.

At one seat, perhaps. At another, not exactly given the averaging of all the individual points. Averaging multiple answers gives you good answers for all, but one most certainly CANNOT say that it only affects certain areas of the room. Such is impossible using seven speakers.

Ralph, you can see on the screen of your television set what equalization calculation that each measured position gets.

More specifically, the correction factor for each position. Which never is found at the output of the device. Instead, each is combined with all the others to provide a lump sum product. One answer for all positions. Ten positions. A thousand positions. Doesn't matter. You've only got one output! That is all that matters because the output signal never sends any one individual position's EQ. One speaker is incapable of sending ten unique signals as you continue to assert.

You keep quoting how MultiEQ pro works, but you have yet to make one point about MultiEQ XT32 pro works. Multi EQ XT32 pro has more filters, with more resolution, more horsepower which leads to the ability to provide finer calculations for those many more points.

And still results in a single correction applied via a single speaker to each and every position in the room. Smarter and better weighting factors for sure. In the end, one solution. The only possible way to deliver position specific EQ is to provide position specific speakers! Comprende, senior?

Unlike MultiEQ pro, it does not work in clusters, it works on individual positions. If position A falls into the target curve, it does not re-equalize that position.

That provides a more accurate weighting factor which gets summed with all the other weighting factors. There remains, however, only one averaged output. There is only one summarized solution for the room.

While the application is different, I write software that uses various weighting factors to determine supply chain demand. Using standard deviation, you eliminate outlier values. You employ time based exponential smoothing to weight demand based upon multiple factors. You build in as much intelligence as you can to provide the best single result. Unquestionably, intelligently weighted systems work better than those using simple arithmetic averaging. They CANNOT, however, under any circumstance provide separate simultaneous solutions for different scenarios.

rw

[Sir Terrence the Terrible]

At one seat, perhaps. At another, not exactly given the averaging of all the individual points. Averaging multiple answers gives you good answers for all, but one most certainly CANNOT say that it only affects certain areas of the room. Such is impossible using seven speakers.

MultiEQ averages, XT32 pro equalizes each point after the measurement of the point is taken. Since each point will measure differently, each point is equalized differently. (Left speaker) Front row left seat, measured and equalized for flat. Center seat measured, equalized for flat. Right chair measured equalized for flat. Repeat for second row. Then we move to the center speaker, repeat process. When the system is done, each speaker will deliver a flat(or near flat) signal to each of the seats. XT32 corrects each point separately, and then sends the corrected output to the speaker.

Before I purchased all of the units I did, I tested Audyssey claims with B&K lab microphone, and my RTA. Did before and after measurements at each of my twelve seats. From each seat, I got a nearly flat measurement from each speaker just as Audyssey claims. The proof is in the pudding.

More specifically, the correction factor for each position. Which never is found at the output of the device. Instead, each is combined with all the others to provide a lump sum product. One answer for all positions.

You have just describe how MultiEQ and MultiEQ XT works, not XT32 pro

Ten positions.

Ten different equalization curves separately measured and equalized to match the reference curve = a flat response for all measured positions.

A thousand positions. Doesn't matter. You've only got one output! That is all that matters because the output signal never sends any one individual position's EQ. One speaker is incapable of sending ten unique signals as you continue to assert.

It does not have to send out ten unique signals, and I never said it did. All the system does is correct those ten position to match the target curve(flat). Once those ten measurements are corrected to the target, that equalized(flat or near flat) curve is what the speaker see's.

And still results in a single correction applied via a single speaker to each and every position in the room.

Nope. XT32 pro measures and corrects each positionSEPERATLY. The result of that is a flat curve to all seats from a single speaker. MultiEQ does the single curve.

Smarter and better weighting factors for sure. In the end, one solution. The only possible way to deliver position specific EQ is to provide position specific speakers! Comprende, senior?

Still wrong. You still do not get it. XT32 measures the reference position first. If that position meets its targeted curve, it applies no equalization to that position. You move the mike, take another measurement with the same speaker. If that measurement does not meet the targeted curve, it corrects that position frequency response so it matches position A. Rinse and repeat. At the end of the process, every seat matches the same curve as the reference seat.

That provides a more accurate weighting factor which gets summed with all the other weighting factors. There remains, however, only one averaged output. There is only one summarized solution for the room.

Wrong again. It does not apply the same equalization for my surround speakers that are close to a boundary, to the front speakers that are not. It applies a different set of equalization values for each position measured.

While the application is different, I write software that uses various weighting factors to determine supply chain demand. Using standard deviation, you eliminate outlier values. You employ time based exponential smoothing to weight demand based upon multiple factors. You build in as much intelligence as you can to provide the best single result. Unquestionably, intelligently weighted systems work better than those using simple arithmetic averaging. They CANNOT, however, under any circumstance provide separate simultaneous solutions for different scenarios.

rw

Ralph, I don't give a damn what you do, you did not write the software for Audyssey did you? NO, so your what you do for a living is irrelevant to what we are talking about here.

What seems pretty clear here is you don't understand the difference between MultiEQ, and MultiEQ XT32 pro. They very clearly don't work the same way because MultiEQ does not have the horsepower to do so. Certain things are done differently because of this. MultiEQ cannot equalize surround arrays(multiple speakers with the same output), XT 32 pro can. MultiEQ cannot be used with full range speakers(you must have a 80hz crossover point from subs to mains), XT 32 Pro can. MultiEQ cannot provide individual corrections for individual points, XT 32 Pro can. MultiEQ, MultiEQ XT are found in receivers and processors, MultiEQ XT32 PRO can only be found as a standalone directly from Audyssey. MultiEQ XT32 PRO requires a laptop computer, MultiEQ and MultiEQ XT does not. MultiEQ XT32 pro can provide different EQ values to two or more subwoofers, MultiEQ and XT cannot.

Lots of difference between the two, and since you have no experience with any of them, your assumptions and conclusions should be taken with a bag of salt.
.

[E-Stat]

From each seat, I got a nearly flat measurement from each speaker just as Audyssey claims. The proof is in the pudding.

I bow to the power of a single equalization curve.

rw