Home Theater Forum and Systems banner

First Measurement!

27K views 205 replies 12 participants last post by  jtalden 
#1 ·
So proud that I finally did my first measurement tonight. :)
How does it look guys?

SETTINGS:
- Marantz SR5005 processor with settings: mode LFE (so not the LFE+MAIN setting), LPF LFE 120Hz, crossover 80Hz, all speakers set on SMALL
- Teufel Theater 8 5.1 surround set http://www.areadvd.de/hardware/teufeltheater8.shtml
- acoustic room profesionally treated with a lot of big basstraps, 6 types of different big diffusers and a Sombra-D ceiling.

I have bought a Behringer FBQ1000 (same as DSP1124) which is now in bypass mode. Looks like there s some work to do for me in the low frequencies. Should I first start to play with settings like LFE mode and crossover freq in my AVR or skip this and go right to my Behringer?

For your info: my subwoofer phase is already set properly. Move my subwoofer or main speakers is not possible anymore (only a few cm)
 

Attachments

#170 ·
A horizontal line is not possible with your system, or mine, or any “conventional” setup. All dynamic speakers and SWs have phase rotation and GD shifts that prevents obtaining a flat GD/EGD line. There are a few methods to get there, but you and I don’t have that capability. Those systems would pass a perfect square wave, but they still can have significant tradeoffs in other respects.
Please pass the square waves:

Here is very conventional system 5.25 inch woofer, 2 inch full range as tweeter. Crossover is 1kHz with DCX2496. Two dynamic drivers and electronic crossover, really no different than SW and FR here with AVR as crossover.

Raw system response:



Much as filters from REW may be exported and employed via convolution, I use filter derived by DRC type methods. This is becoming less a departure from "conventional" as more people wrap there heads around it. For each frequency in FFT of system's time domain impulse response, amplitude and phase are adjusted. Computer solves thousands of filters simultaneously.

System after EQ:



I've attached mdat for more thorough look.

Proper understanding of technique and methods makes alignments very straight forward. Multi channel sound cards may be used in place of DCX. For theater there is drawback of often needing lip sync correction via video delay. Some AVRs do this, and some software for htpc will also do this.

The key is real time convolution engine.

Here I posted project with same drivers and sound card in place of crossover. Post includes SW work as well: http://orion.quicksytes.com/viewtopic.php?f=14&t=2254&sid=f29cbbd0474374621f1cbcb7fccd94b9

It includes square wave response pics.

The sound imaging with music is spectacular. Transient behavior is spectacular.

Andrew
 

Attachments

#171 ·
No, I got it by looking at the last SPL overlay. The SPL roll off of the SW and FR traces don’t run into the noise floor until 35 dB or so. You can easily measure the noise floor by taking a measurement without any signal to the speakers (just turn the AVR off). Yours will look very good.
with noise floor you just mean the environment sounds? Is pic1 my noise floor?

LP is usually the “Listening Position” on all these audio forums unless it is clear that the context is filters where “Low Pass” would make sense. LPF for Low Pass Filter would be preferred to avoid confusion. The LP is where we tell you to put the mic. Where did you put your mic? :D [Please don't tell me where to put mine.]

The measurement signal originates from REW and passes through several devices including the speakers before it travels as an audio signal to the mic at the LP, then again as an electrical signal back into REW. With loopback engaged the IR time is the difference in time between the arrival of loopback signal and the mic signal. So if the loopback comes before the AVR the time includes the delay in the AVR. That is not a problem for us when the job is setting the relative speaker distances properly.
understood

This entire exercise is concerned with optimizing the difference in time that the signal arrives at the LP for the SW verses the FR main speaker.

You can just take it as a fact the initial large peak of the IR is representative of the time at which the “HF” portion of the sound arrives for the speaker. Be mindful that what is HF for a SW is normally called bass or LF in most all other cases, but it is still the highest freq that the SW reproduces. You will not be measuring any sound at 10k Hz using only a SW so the IR peak can’t represent arrival time of 10k Hz! So the initial large peak in the SW IR can be thought to represent maybe the frequencies on the SPL chart where just before the SPL rolls off, maybe around 100 Hz in your case.

The 5k - 20k range I mentioned for the FR is just a nice safe number to give you to understand of what to think of when I said “HF” for that speaker. There is nothing magic about the pick except that I know the arrival time of that entire range of frequencies can be thought of as equal as there will miniscule GD change in that range. The nearer to the LF range (90 Hz for that speaker) The greater the GD/EGD will be.

The “-38 to 100%FS” Means nothing to us in this context. It is the time delay of the initial peak that is of interest no matter the magnitude it has.



This should have said “cannot see that in the IR chart”. I just lost the negative somehow.



Per above.



Again I intended to type “we cannot see that in the IR chart” I copied and pasted the phrase from above so when the negative was missing above it was missing here also. I will have to discuss this with my editor, but first a glass of wine :D



Sorry, I was referring to all the hash that occurs in the IR trace after that initial IR peak. You may want to find an IR of a clean electrical signal or a speaker in an anechoic situation to see that it has just one peak.



I guess you need a better teacher. :) I think you said you understand it your following post though so maybe it got through.



Yes, How much the relative signal needs to be attenuated to make its effect trivial is a matter of opinion. Maybe 15 dB is enough?



I’m lost here. You don’t understand that we started out trying to maximize the SPL output in 40 - 90 Hz range to help fill the dip you have there? We did not say we wanted to align the phase at 90 deg or minimize the SW GD/EDG or...
what if I just take the following measurements and get the dips and peaks as good as possible and hear from you if everything is ok:

SW+FRONT LEFT
SW+CENTER
SW+REAR RIGHT
SW+LEFT RIGHT
 

Attachments

#172 ·
with noise floor you just mean the environment sounds? Is pic1 my noise floor?
Yes

what if I just take the following measurements and get the dips and peaks as good as possible and hear from you if everything is ok:

SW+FRONT LEFT
SW+CENTER
SW+REAR RIGHT
SW+LEFT RIGHT
We could do that, but we will likely end up with LP IR timing of each of the main speaker being different from each other. It is better to have the LP IR timing the same for all main speakers.

We established good IR timing for the FR. I suggest we just replicate that same timing to the other main speakers. (We don't need the SW turned on for this purpose)

The process:
[SW off, Mic at LP, Set Loopback on.]

> Measure FR main speaker. [This is our reference IR timing.]
> Set the FL distance in the AVR same as the FR since the FL speaker is probably about equal distance from the LP. [This is not necessary, but sets the IR timing close to the correct setting].
> Measure the FL speaker.
> Zoom in as tight as possible on the 2 IR peaks in the overlay window so we can compare the timing of the FL to the reference timing of the FR. Like this:

Text Line Green Plot Font


> Now adjust (see note below) the FL distance in the AVR and remeasure until the 2 IR peaks fall as closely as possible using the highest zoom setting for the time scale. [When you get the best IR alignment possible you can delete any trial measurements it took to get there. The IR chart should now approach this:

Text Green Line Plot Font


> Do this for each main speaker - then the final IR overlay chart will come as close to this as possible:

Text Line Blue Plot Font


Because of the limits of the AVR distance settings we will not get all the other IRs to fall directly on top of the FR reference IR as shown here. The objective is just to get them as closely aligned as possible. A small error like 0.1 m, or so is not significant. Just use the AVR distance setting to minimize the error.

Note on adjustments: We can use trial and error to adjust the IR timing or it is possible to measure the IR offset distance using Control Key and right mouse button to click and drag between the IR peaks on the chart as shown above. The cursor indicates the needed adjustment.

Now the IR timings (AVR speaker distance settings) are optimized for all main speakers.

We can now measure the SW+Main for each speaker and overlay the SPL responses.

Look particularly at the SW+FR SPL in the XO range to confirm that the alignment we have chosen works for the SW+FL as well as the SW+FR. If the SPL looks poor for the SW+FL in the XO range then we may want to revisit the original choice we made of using the SW+FR to establish the target timing. In that case we may find that a compromise may be in order between using the SW+FR optimized timing as the reference and using the SW+FL optimized timing as the reference. Hopefully there will be no significant problem in this respect.
[The XO range of the other main speakers is not as critical as the FL and FR so I would make SW to Main alignment decision based only on the FL and/or FR.]


Keep in mind this method:
> First "Optimized" the target timing using the SW+FR, but that doesn't mean that the SW+FR timing could not have been been done in other ways with equal or better results sonically. It was actually just one good compromize alignment of several possible good alignments.
> Then Optimized the timing alignment of the other main speakers to match the target. This is always a good idea.

Then it's off to EQ. :sn:
 
#174 ·
Hi John,

I just did a new measurements with your new instructions. I m a bit proud of myself for accomplishing the next screenshot. All credits goes to your help of course!

The red measurement is my FRONT RIGHT and the green is my FRONT LEFT. Am I correct that there is a time difference when the sound arrives at LP of 7.2805ms and that now need to adjust the distance setting of the LEFT SPEAKER in my AVR settings with the result that the large peaks (100% FS) overlap as close as possible?

Hope I m on the right track now. And another question before I move on. After this I will need to measure my CENTER, RIGHT SURROUND and LEFT SURROUND too and adjust them the same as I will have to do for my FRONT LEFT speaker. But how do I measure my center, right surround and left surround channels? What I did for my LEFT FRONT measurement was switch the RIGHT AUX IN of my avr to the LEFT AUX IN of my avr which results in the measurement of my left front speaker but what do I do for measuring the center and surrounds??

Here s also the .mdat of the right front and left front measurement I just did:
http://www.sendspace.com/file/67yp1w
 

Attachments

#175 ·
Hope I m on the right track now. And another question before I move on. After this I will need to measure my CENTER, RIGHT SURROUND and LEFT SURROUND too and adjust them the same as I will have to do for my FRONT LEFT speaker. But how do I measure my center, right surround and left surround channels? What I did for my LEFT FRONT measurement was switch the RIGHT AUX IN of my avr to the LEFT AUX IN of my avr which results in the measurement of my left front speaker but what do I do for measuring the center and surrounds??
It sounds like he means for you to only do this on left and right (which would typically be considered the "mains"). Then you just set the center and surrounds by taking physical measurements or possibly using the settings Audyssey figured out originally.

However if you really want to try it on other channels... to get IRs for the others the easiest way is to put your receiver into "7.1 ch input" mode and use the multichannel inputs on the back (pages 21 and 50 in the SR5005 manual). This however does not steer the bass to the subwoofer. I have read on another thread here that if you connect the computer to the AVR with HDMI and use "ASIO4ALL" you can steer the sweep to any channel (this is actually something I intend to try myself), and if you do this then you can keep the AVR in surround mode and it would be processed through bass management as usual. There are probably other ways to do this with some of the community test DVDs that are available, but I'm not entirely sure how one would go about that.
 
#179 ·
FRONT LEFT@ 6.12m gives me the best result (red line is FRONT RIGHT):

It would be best to first measure SW+FL and SW+FR and overlay the SPL results. This will allow us to confirm that the SW+FR timing adjustment we chose also works well for the SW+FL as well. If that looks good then it is time to align the rest of the speakers.
10 min and I will post this
 

Attachments

#180 ·
I don t think that I have good news. I just measured (.mdat link on bottom of this reply)

Graph 1: subwoofer (4.25m) +front left (6.12m)
Graph 2: subwoofer (4.25m) +front right (6.25m)
Graph 3: no sub, front left (6.12m)
Graph 4: no sub, front right (6.25m)

Am I correct that I can make the conclusion that the SPL overlay of the SUB+FRONT LEFT is by far not as good as the SUB+FRONT RIGHT measurement? Can I also make the conclusion (please correct me if I m wrong) that with this setup (distances settings) in a stereo setup the time that the sound arrives at my LP is perfectly balanced between the LEFT and RIGHT signal (that s why I matched the IR peak as good as possible) but the SPL in the SUB+LEFT FRONT has some dips (phase errors?). And I can make those dips from the SUB+FRONT LEFT better by playing with the distance setting of the FRONT LEFT but then I will lose perfect timing because the IR peaks wil have a delay?

http://www.sendspace.com/file/e78szi
 
#181 ·
The FR and FL timing is now ideal for that LP.

I checked out the phase alignment between the SW and FR vs FL. They are both aligned as well as possible.

SPL difference between FL and FR is to be expected. I think you are overly concerned with the difference you found. The SPL differences here look to be relatively minor and very manageable. The more symmetrical the room setup the more similar the SPL response will be.

I would suggest you retain this alignment and proceed to align the CC, SL, and SL in the same way. When you have all the distances settings established, please post them.
 
#182 ·
Ok, I ll start measuring the center, surround left and surround right today with the front right IR as 'reference'. The only thing I m still a bit confused about is how to actually do this. My Marantz SR5005 has no MONO MODE. I use my Marantz SR5005 as processor and I use a separate 5-channel power amplifier (NuForce MCH-2-C5) that is connected unbalanced (rca) to the pre-outputs of the SR5005.

I found this in the Marantz SR5005 manual:

http://www.manualowl.com/m/Marantz/SR5005/Manual/222916?page=32

If you read the 'Multi Channel Playback' chapter at page 29 you see there is a option MULTI CH STEREO. Is this the one I should use? (while only playing the channel that I m measuring and disconnect the rest of the channels?)
 
#184 · (Edited)
That setup makes it relatively easy:
[For this process I will assume that we have just finished the timing adjustment with SW+FR (our reference channel) and are ready to align all the other channels to the FR IR. We still have the mic at the LP and a measurement in REW of the SW and the FR.]

> Leave all speaker wires in place for this entire process.
> Leave the REW output connection to the AVR FR input for this entire process.
> Turn system off (or just the P-amp) then disconnect all 5 RCA connections at the P-Amp.
> Connect the FR Pre-out RCA connector back into the P-Amp FL channel input. Leave all other P-amp main inputs disconnected.
> Turn on the system and set stereo mode.
> Turn the SW off
> Adjust the IR timing of the FL to match the measurement of the FR IR per the previous instructions.
> Turn the SW on and measure the SW+FL
> Repeat for CC, SL and SR.
> Save this REW file that contains the SW, 5 Mains, and 5 SW+main measurements as our record of the setup.
> Turn off the system and reset all the RCA connectors back to their original P-amp inputs.
> Record all the AVR distance settings and save them.
> Done.

EDIT: Notice that the distance adjustment will be made to the FR distance setting in the AVR for all 5 channels. This means we need to keep track of each channel distance we establish. When the process in complete the distances we found are loaded into the AVR for each of the respective channels.
 
#191 ·
That looks good.

> The SW+FR timing was aligned per “m11”; one of the 3 acceptable targets.
> The timing (distances) of all the other main speakers was found and recorded and entered in the AVR.
> The SPL and phase of the SW+FL was also confirmed to be acceptable.

With this you can now EQ the system.

Other thoughts:
> The m11 alignment was the target alignment with the most GD as you pointed out several post back. I had mentioned that m12 may provide a preferable sound. It will reduce the GD and still provide approximately the same SPL response in the XO range. The good news is that you can evaluate both pretty easily.

To convert from m11 to m12 we can use one of the 2 methods below:
1. Subtract 5.16 m from all 5 of the main speaker distances (the SW stays at the current 4.25m).
[With the current mains settings we can’t do this as it results in a negative distance for the SL and SR. The AVR may also have a minimum allowable distance setting to deal with.]
Or,
2. Add 5.16 m to the SW (the mains all stay at their current settings).
[With the current SW setting we may not be able to do this either if we run into a maximum distance that the AVR allows.]

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

We know it is the relative distances between speakers that is important so may be able adjust all the distances to fall within the AVR limitations and allow us to adjust between m11 and m12 alignments just by changing only the SW distance.

Assuming the AVR minimum distance allowed is 0.50 m and maximum is 8.00 m, here is one set of numbers that would work:
SW: 4.25 – 3.25 = 1.00 m
FR: 6.25 – 3.25 = 3.00 m
FL: 5.99 – 3.25 = 2.74 m
CC: 5.89 – 3.25 = 2.64 m
SL: 4.91 – 3.25 = 1.66 m
SR: 5.02 – 3.25 = 1.77 m

These distances still provide the same “m11” alignment you have now, but if we want to switch to m12 alignment we just set the SW distance to 6.14 m (1.00 m + 5.17 m). In this way we can switch back as forth with only this SW distance change.

----------

Regarding EQ:
Since the SPL of XO range is reasonable close for either m11 or m12 alignment the EQ settings can be established using either alignment. The EQ settings will then work well for both alignments.

I also notice that there is a significant peak in the SPL of all main speakers around 100 – 160 Hz. The EQ of the SW using the Behringer will probably not address this issue well, so I suggest first trying to reduce this peak somewhat using the AVR EQ capabilities. If a course adjustment is made in the AVR first, then the EQ of the SW by the Behringer may be more successful.

With that, you are ready for EQ. If you want EQ help, you may want to start a new thread with all the pertinent system info and baseline measurement restated so the EQ experts here can jump in without having to dig through this one to find baseline info.

These last points are just a few suggestions to consider.
 
#192 ·
The SW+FR timing was aligned per “m11”; one of the 3 acceptable targets.
Can you summarize one last time how we did this?

The timing (distances) of all the other main speakers was found and recorded and entered in the AVR.
understand. One thing though: Auddysey measured totally different distance settings (Auddysey corresponds with tape measuring, eg. my surround speakers are @2.80m physical distance and audyssey measured 2.85m). Is audyssey THAT bad?

The SPL and phase of the SW+FL was also confirmed to be acceptable.
SPL I can see but phase is a thing I would still like to learn too how to see this. Can t one just say that the phase is correct because the SPL has the least dips in the CO range?
And another question about this: why don t we check the SPL and phase of the SW+CC, SW+SR and SW+SL? Escpecially the center speaker does a lot during a movie.

With this you can now EQ the system.

Other thoughts:
> The m11 alignment was the target alignment with the most GD as you pointed out several post back. I had mentioned that m12 may provide a preferable sound. It will reduce the GD and still provide approximately the same SPL response in the XO range. The good news is that you can evaluate both pretty easily.
what was a reduced GD again?

To convert from m11 to m12 we can use one of the 2 methods below:
1. Subtract 5.16 m from all 5 of the main speaker distances (the SW stays at the current 4.25m).
[With the current mains settings we can’t do this as it results in a negative distance for the SL and SR. The AVR may also have a minimum allowable distance setting to deal with.]
Or,
2. Add 5.16 m to the SW (the mains all stay at their current settings).
[With the current SW setting we may not be able to do this either if we run into a maximum distance that the AVR allows.]
just checked my AVR. 9.41m (4.25+5.16) is possible with my AVR.

We know it is the relative distances between speakers that is important so may be able adjust all the distances to fall within the AVR limitations and allow us to adjust between m11 and m12 alignments just by changing only the SW distance.

Assuming the AVR minimum distance allowed is 0.50 m and maximum is 8.00 m, here is one set of numbers that would work:
SW: 4.25 – 3.25 = 1.00 m
FR: 6.25 – 3.25 = 3.00 m
FL: 5.99 – 3.25 = 2.74 m
CC: 5.89 – 3.25 = 2.64 m
SL: 4.91 – 3.25 = 1.66 m
SR: 5.02 – 3.25 = 1.77 m

These distances still provide the same “m11” alignment you have now, but if we want to switch to m12 alignment we just set the SW distance to 6.14 m (1.00 m + 5.17 m). In this way we can switch back as forth with only this SW distance change.
understand this but maybe the subwoofer distance @9.41m is good to go

Regarding EQ:
Since the SPL of XO range is reasonable close for either m11 or m12 alignment the EQ settings can be established using either alignment. The EQ settings will then work well for both alignments.
understand

I also notice that there is a significant peak in the SPL of all main speakers around 100 – 160 Hz. The EQ of the SW using the Behringer will probably not address this issue well, so I suggest first trying to reduce this peak somewhat using the AVR EQ capabilities. If a course adjustment is made in the AVR first, then the EQ of the SW by the Behringer may be more successful.
understand. My AVR lets me EQ all speakers (not the subwoofer though, but that s the Behringer for right?) separate. Shall I try to EQ all speakers (FR, FL, CC, SR and SL) so they all have a good SPL response? My AVR lets me do per speaker:
-20 to +6 (0 default) for 63Hz, 125Hz, 250Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, 8000Hz and 16000Hz. And if I EQ them all (separate), are there 'rules' to keep in mind? I remember from my Velodyne SMS-1 subwoofer eq I had a few years ago that increasing (+) was an abslote no-go and that only a few minor decreases (-) at some peaks was recommended. How will that be with my AVR?
And also: will adjusting EQ in my AVR affect delays (timings) that we fine-tuned?

With that, you are ready for EQ. If you want EQ help, you may want to start a new thread with all the pertinent system info and baseline measurement restated so the EQ experts here can jump in without having to dig through this one to find baseline info.
amost there! (I hope :))

These last points are just a few suggestions to consider
your help is so much appreciated John.
 
#193 ·
Can you summarize one last time how we did this?
Manual “SPL/IR” SW+Main timing alignment method:
[An alternative method to the “SPL/RTA” SW+main timing alignment method.]

Setup steps:
> Mic at LP.
> Choose to use either the FL or FR to establish the timing alignment with the SW. We chose the FR.
> Measure and record the physical distances from the LP of both the SW and FR. Set one of them to 5 m in the AVR and the other to the difference in physical distance. Example: SW measures about 2.8 m and the FR about 3.2 m. So we could add 1.8 m to both values and enter SW = 4.6 m and FR = 5.0 m in the AVR. We will adjust the one that is set to 5 m. In this example that is the FR.
> Previously entered EQ settings can be either left on or turned off.
> Activate Loopback feature in REW.
> Set REW preferences to use 1/3 or 1/6 smoothing.
> Use only full range sweep measurements (15 – 20k Hz) for all measurements.
> Choose the AVR distance increments to be used for the trial and error process. I recommend about 0.5 m.

Process:
[Trial and error method. (See post 140 for a calculation method.)]
> Sweep measure SW, FR, SW+FR
> View IR locations of the SW and FR initial peaks. [If the initial distances are entered properly the initial rise of the FR and SW IR peaks should fall within 1 or 2 ms of each other. This is good enough for the starting point of this process.]
> Be sure to label the distance setting on each measurement
> View the overlay SPL graph for the SW, FR and SW+FR. Identify the XO freq range as the range between where either the SW or the FR SPL falls more than maybe 20 or 25 dB from the average SPL in the area adjoining that range.
> Increment the FR distance and measure both FR and SW+FR.
> View the overlay SPL graph for each SW+FR measurement. Look to identify a setting that is nearest to the initial starting setting of 5 m that provides the highest SPL through the XO range. [I like to leave the “SW” and “FR at 5.00 m” traces visible also as I review the SW+FR increments because that better shows the SPL fill in the XO range of the traces. A good alignment will have the SPL+FR trace fall as high as possible over the SW or FR trace at all points in the XO range.]
> After we find that good alignment we can then optionally fine tune that setting by using smaller distance increments. The difference in distance between the SW and FR is the magic number we need to later set the distances in the AVR.
> The setting with the initial rise of SW and FR IRs falling closest to each other is the alignment that provides the closest phase match of the SW and FR throughout the entire XO range.
> Once we find this first solution we can predict the approximate location of alternate solutions. Another solution will fall the distance of one WL (wavelength) greater and another one WL less.
[For example if we find that a solution with the FR at 4.85 m and the XO is at 80 Hz then another solution will be found with the FR near 4.85 + 4.30 = 9.05 m and also one at 4.85 – 4.3 =
0.55 m (4.3 m being the approx. WL of 80 Hz).]
If we invert the polarity of the SW, we can expect to find solutions at about 1/2 WL more and 1/2 WL less.
> We are only interested in 3 of the alignments the one that places the initial IR rise of the SW and FR closest to the same time and the ones 1/2 WL above and 12 WL below that. We can go find those alignments and fine tune them if we want. [The distance offsets are only estimates as the actual phase timing may not actually be right at 80 Hz XO freq as was the case here for Sanders setup. To maximize the SPL fill-in in this case the alignment target was nearer 65 Hz than 80 Hz. The offsets distances are thus impacted according, WL distance for 65 Hz is about 5.16 m]
> We should also confirm that this solution works well for the FL speaker as well.
[I would think it would in most cases, but with my very asymmetrical room setup I decided to use a compromised solution falling between the best solutions found for the FR and for the FL.]
> We now use any FR and SW distance values in the AVR that retains the offset distance between them. We just need to select values that will provide enough room to allow us to set the CC, SL and SR distances to the values needed. Those that match the same IR arrival time as the FR IR.


One thing though: Auddysey measured totally different distance settings (Auddysey corresponds with tape measuring, eg. my surround speakers are @2.80m physical distance and audyssey measured 2.85m). Is audyssey THAT bad?
Most comments I see indicate that Audyssey does very well in aligning the SW to Mains. I have not seen measurements (other than my own experiments) that do a good job of comparing the Audyssey results to a well aligned manual setup. My own experience with Audyssey timing is not good, but I have mostly tested it with non-standard setups. If I set my DCX filters to simulate the standard main speaker roll off rates at about 80 Hz then Audyssey did in fact find a good XO setting and good timing. It still boosted my SWs excessively (about +9 dB at 20 Hz; their lower limit). Maybe like many automated setup systems (for any type of system) as long as the expected conditions are found the setup will be acceptable?? Speaker designs and room responses vary so much it is probably difficult to get it right in all cases.

Your 0.05 m difference for the surrounds is insignificant. The initial timing of your FR was much worse however. I think that is what started the initial concern.

I speculate that, had you just reversed the polarity on the SW (or all the main speakers) then Audyssey may well have gotten it right. I say that because your initial alignment was close to correct if the polarity of the SW was switched. The IRs were aligned well, but it was about 160 deg out of phase alignment leaving a big dip in your SPL response.

SPL I can see but phase is a thing I would still like to learn too how to see this. Can t one just say that the phase is correct because the SPL has the least dips in the CO range?
Yes.

why don t we check the SPL and phase of the SW+CC, SW+SR and SW+SL? Escpecially the center speaker does a lot during a movie.
I think that in movies the CC is primarily a dialog channel and there is little if any bass in it? You can test this with the REW RTA if you like. Maybe someone else here already knows. Even if it does contain bass would you decide to prioritize the timing for the CC at the expense of the FL and FR?
Your particular speakers are designed to properly to go together in an HT setup. The CC and surround designs are both complimentary to the FL/FR so that the phase timing would be expect to be very similar. I did indeed check this out and, as expected, the phase timing of all 5 speakers is similar so no worries.

what was a reduced GD again?
I was referring to adjusting the timing such that the GD (Group Delay) of the mid bass (maybe 40 – 65 Hz) would be minimized with the tradeoff being that the phase tracking at the extremes of the XO range is somewhat less than ideal. The phase and GD near the XO will still be the same however.

just checked my AVR. 9.41m (4.25+5.16) is possible with my AVR.

maybe the subwoofer distance @9.41m is good to go
Yes, that works fine.

My AVR lets me EQ all speakers (not the subwoofer though, but that s the Behringer for right?)

Shall I try to EQ all speakers (FR, FL, CC, SR and SL) so they all have a good SPL response? My AVR lets me do per speaker:
-20 to +6 (0 default) for 63Hz, 125Hz, 250Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, 8000Hz and 16000Hz. And if I EQ them all (separate), are there 'rules' to keep in mind? I remember from my Velodyne SMS-1 subwoofer eq I had a few years ago that increasing (+) was an abslote no-go and that only a few minor decreases (-) at some peaks was recommended. How will that be with my AVR?
I’ll let someone else help you with EQ questions. You might want look through the sticky threads and do some general searches on the topic. There is lots of EQ related info available including regarding the boost vs cut situation. You should be able to easily find info on these kinds of basic questions and read through them before you get started.

will adjusting EQ in my AVR affect delays (timings) that we fine-tuned?
Not significantly.
 
#194 · (Edited)
Manual “SPL/IR” SW+Main timing alignment method:
[An alternative method to the “SPL/RTA” SW+main timing alignment method.]

Setup steps:
> Mic at LP.
> Choose to use either the FL or FR to establish the timing alignment with the SW. We chose the FR.
> Measure and record the physical distances from the LP of both the SW and FR. Set one of them to 5 m in the AVR and the other to the difference in physical distance. Example: SW measures about 2.8 m and the FR about 3.2 m. So we could add 1.8 m to both values and enter SW = 4.6 m and FR = 5.0 m in the AVR. We will adjust the one that is set to 5 m. In this example that is the FR.
> Previously entered EQ settings can be either left on or turned off.
> Activate Loopback feature in REW.
> Set REW preferences to use 1/3 or 1/6 smoothing.
> Use only full range sweep measurements (15 – 20k Hz) for all measurements.
> Choose the AVR distance increments to be used for the trial and error process. I recommend about 0.5 m.

Process:
[Trial and error method. (See post 140 for a calculation method.)]
> Sweep measure SW, FR, SW+FR
> View IR locations of the SW and FR initial peaks. [If the initial distances are entered properly the initial rise of the FR and SW IR peaks should fall within 1 or 2 ms of each other. This is good enough for the starting point of this process.]
> Be sure to label the distance setting on each measurement
> View the overlay SPL graph for the SW, FR and SW+FR. Identify the XO freq range as the range between where either the SW or the FR SPL falls more than maybe 20 or 25 dB from the average SPL in the area adjoining that range.
> Increment the FR distance and measure both FR and SW+FR.
> View the overlay SPL graph for each SW+FR measurement. Look to identify a setting that is nearest to the initial starting setting of 5 m that provides the highest SPL through the XO range. [I like to leave the “SW” and “FR at 5.00 m” traces visible also as I review the SW+FR increments because that better shows the SPL fill in the XO range of the traces. A good alignment will have the SPL+FR trace fall as high as possible over the SW or FR trace at all points in the XO range.]
> After we find that good alignment we can then optionally fine tune that setting by using smaller distance increments. The difference in distance between the SW and FR is the magic number we need to later set the distances in the AVR.
> The setting with the initial rise of SW and FR IRs falling closest to each other is the alignment that provides the closest phase match of the SW and FR throughout the entire XO range.
> Once we find this first solution we can predict the approximate location of alternate solutions. Another solution will fall the distance of one WL (wavelength) greater and another one WL less.
[For example if we find that a solution with the FR at 4.85 m and the XO is at 80 Hz then another solution will be found with the FR near 4.85 + 4.30 = 9.05 m and also one at 4.85 – 4.3 =
0.55 m (4.3 m being the approx. WL of 80 Hz).]
If we invert the polarity of the SW, we can expect to find solutions at about 1/2 WL more and 1/2 WL less.
> We are only interested in 3 of the alignments the one that places the initial IR rise of the SW and FR closest to the same time and the ones 1/2 WL above and 12 WL below that. We can go find those alignments and fine tune them if we want. [The distance offsets are only estimates as the actual phase timing may not actually be right at 80 Hz XO freq as was the case here for Sanders setup. To maximize the SPL fill-in in this case the alignment target was nearer 65 Hz than 80 Hz. The offsets distances are thus impacted according, WL distance for 65 Hz is about 5.16 m]
> We should also confirm that this solution works well for the FL speaker as well.
[I would think it would in most cases, but with my very asymmetrical room setup I decided to use a compromised solution falling between the best solutions found for the FR and for the FL.]
> We now use any FR and SW distance values in the AVR that retains the offset distance between them. We just need to select values that will provide enough room to allow us to set the CC, SL and SR distances to the values needed. Those that match the same IR arrival time as the FR IR.
understand. Maybe a moderator can make a sticky of this information so other members can benefit from your knowledge too.

Most comments I see indicate that Audyssey does very well in aligning the SW to Mains. I have not seen measurements (other than my own experiments) that do a good job of comparing the Audyssey results to a well aligned manual setup. My own experience with Audyssey timing is not good, but I have mostly tested it with non-standard setups. If I set my DCX filters to simulate the standard main speaker roll off rates at about 80 Hz then Audyssey did in fact find a good XO setting and good timing. It still boosted my SWs excessively (about +9 dB at 20 Hz; their lower limit). Maybe like many automated setup systems (for any type of system) as long as the expected conditions are found the setup will be acceptable?? Speaker designs and room responses vary so much it is probably difficult to get it right in all cases.

Your 0.05 m difference for the surrounds is insignificant. The initial timing of your FR was much worse however. I think that is what started the initial concern.

I speculate that, had you just reversed the polarity on the SW (or all the main speakers) then Audyssey may well have gotten it right. I say that because your initial alignment was close to correct if the polarity of the SW was switched. The IRs were aligned well, but it was about 160 deg out of phase alignment leaving a big dip in your SPL response.
I understand all of this. But I just run Audyssey again (of course wrote down all settings we made) and Audyssey says (approx):
FR,CC and FL 5.00m and surrounds@ 2.50m (which corresponds with the actual physical distance of my setup).
The distances we found (for correct time alignment) are approx. 6m for frontstage and 5 meter for surrounds. This is a difference of approx. 1 meter while audyssey measures a difference between front and backstage of 2.5meter......
I think that in movies the CC is primarily a dialog channel and there is little if any bass in it? You can test this with the REW RTA if you like. Maybe someone else here already knows. Even if it does contain bass would you decide to prioritize the timing for the CC at the expense of the FL and FR?
Your particular speakers are designed to properly to go together in an HT setup. The CC and surround designs are both complimentary to the FL/FR so that the phase timing would be expect to be very similar. I did indeed check this out and, as expected, the phase timing of all 5 speakers is similar so no worries.
understand

I was referring to adjusting the timing such that the GD (Group Delay) of the mid bass (maybe 40 – 65 Hz) would be minimized with the tradeoff being that the phase tracking at the extremes of the XO range is somewhat less than ideal. The phase and GD near the XO will still be the same however.
ok, understand

I’ll let someone else help you with EQ questions. You might want look through the sticky threads and do some general searches on the topic. There is lots of EQ related info available including regarding the boost vs cut situation. You should be able to easily find info on these kinds of basic questions and read through them before you get started.
ok. But one last question about this: is it a idea to let my AVR do the EQ and after this change all settings back to the settings we made except for the dequalization that audyssey did? Or does this affect our settings? (I remember you told something about audyssey a week ago and how it affects things)
 
#195 ·
is it a idea to let my AVR do the EQ and after this change all settings back to the settings we made except for the dequalization that audyssey did? Or does this affect our settings? (I remember you told something about audyssey a week ago and how it affects things)
Audyssey will allow you to change the XO freqs, but I think that as soon as you change any speaker distance setting it will automatically disable Audyssey. I am not sure on this point, but you can easily test it yourself. If this is correct then you cannot manually align the XO timing and and still use Audyssey for EQ. If I am mistaken then you can do that.
 
#196 ·
I just did a new Audyssey auto measurement again and I still just don t understand why Audyssey gives so big different results then we do with our manual measurement we did.

Audyssey measures:
FL: 3.84m
FR: 3.93m
CC: 3.72m
SL: 2.64m
SR: 2.79m
SW: 4.17m

We measured:
FL: 5.99m
FR: 6.25m
CC: 5.89m
SL: 4.91m
SR: 5.02m
SW: 4.25m

I know the distances are relative but they don t 'match' at all. So like I asked before: Is Auddysey THAT wrong?
 
#199 ·
Note that if we subtract 2.25 m to the manual distance setting for all the speakers there is no impact on the timing at all. So if we do that we get:

Text Line Font Number Design


The distances to the mains now fall within +/-0.10 m for all 3 methods. Since the Audyssey mic was not placed exactly where the REW mic was placed and the ends of tape measure are also slightly different that is very excellent agreement. There is no evidence or any reason to assume that Audyssey cannot align the main speaker IRs just as well as we did manually. Similarly they can be aligned well with a tape measure with enough care.

The difference then is just in the SW timing that was targeted. Audyssey established different target timing for the XO handoff between SW and the 5 Mains than we did manually and both differ from the tape measurement.

In our measurements above we have good evidence that any one the 3 timing solutions we found manually is a preferable starting point if we intend to manually EQ.
 
#200 ·
Completely understand John! Thanks for clarification.
Auddysey did not set a CO at all because Auddysey set my speakers on LARGE which disables the CO. I already learned before I started this thread that LARGE is a no-go and that speakers always have to be set on SMALL (only when you own 20k B&W 800D's or something similar that can do full 20-20000Hz you can set speakers on LARGE).

But I think (checked this; will let you see tomorrow, not in my HTroom atm) Auddysey did a pretty good job at equalizing my FR, FL, CC, SR and SL. However, the values that Auddysey set for EQ the channels are not shown in my AVR so I cannot write them down and use them in our manual setup. I ve been trying to manually EQ my FR to begin with this afternoon. Spend almost 4 hours on it but it gives me headaches which is the right thing to do. Am I correct that my goal is to get all 5 speakers have a approx. similar SPL responses? (completely similar is not possible due to the limitations of the EQing in the SR5005. One important thing is that I cannot set a width and cannot change the freq. I want to adjust, it are just a few fixed frequencies that can be EQd).
 
#204 ·
Sure, you can use the CC as the reference channel for the SW timing if you like.

No need to rerun everything if you don't want to. You can just fine tune the current setting. It should be very close to ideal already for the CC.

If a small adjustment is made it can be done to the SW distance. That way all the other channels are impacted the same way and the mains timing is not effected. You can just measure each main with the SW to assure the SPL response through the XO range is still good.

If you adjust CC distance instead that is okay also, but it is then necessary adjust all the other mains the same amount to retain the proper mains timing.
 
#205 ·
I did some more thinking last night John. For a 5.1 setup wouldn t it be best to work like this:

1.
Measure and set the distances from the FR, FL, CC, SR and SL so all the 5 main speakers have the same IR delay.
(From this moment DO NOT touch the FR, FL, CC, SR and SL distances anymore!)
2.
Measure all main speakers separate while running the subwoofer and look for the best subwoofer distance per main speaker (highest SPL output in XO range).
3. Compare the best IR distances with the best 'subwoofer phase' distances per speaker.
4. Find a subwoofer distance with all these distances you found to get the best compromise for all 5 main speakers+subwoofer where both the IR and PHASE are good.

You will always have to make compromises but this would result in the 'best/avarage' subwoofer distance setting.
 
#206 ·
The functional difference in that procedure is that all 5 speakers are used to find a compromise timing alignment. I had suggested that the FL and FR mains be used for the compromise alignment. You also found a source that indicated it is best to use the CC for the timing alignment.

My suggestion was based only on my experience as a hobbyist. I have not seen any studies or other opinions on the matter, so value it accordingly.

---------
Some of my logic:
Even when using identical main speakers (like mine), the room modes near the XO can shift the phase of the main speakers greatly in the XO range. In my case the SL/SR are similar and nearly 180 deg different than the FR/CC speakers. The FL is significantly different than both of the other sets (about 80 deg). If I were to compromise between all 5 the result might work okay for the FL, but not for the others. For a while I was aligning per the FR/CC. This left a significant dip in the FL SPL. That still worked reasonably well in my case, but recently I have been setting a compromise setting between the FR/CC and FL. I now feel that works a little better in my setup as the EQ boost needed in the FL is reduced a little.

When deciding to generalize a recommendation I judged that most people would be better served to compromise between the FL and FR mains as I judge them to be the more important main speakers. I also hope that, with a reasonably symmetrical room setup, their phase results would not be too dissimilar. I was guessing this would be better general advice than the other options. I don’t see any sense in providing any weight to the SL/SR if it means significant compromise to the front mains. Everyone’s situation will be different however and no simple guideline can be considered “best” in all cases.]
-------

I have not seen an entire set of measurements for your particular setup so I don’t know what you are looking at for the other 4 main speakers. It’s very possible an average of all 5 mains will work very well in your case. Anyway, now that you have a good feel for the objectives and tradeoffs, you can easily judge that for yourself.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top