Crossover alignment with Subwoofers

[IWC Doppel]

Hi all I am happy with my sub alignment (three subs) and the alignment tool works the graphs looked sensible and all happy one thing I am not sure aboutis why the alignment tool is suggesting the following when I look at the centre channel played LFE only (the aligned group) and centre only. I am crossing over at 60 Hz and have checked the alignment at this frequency. the black line looks right when A+B the sub and centre and match the black line with the alignment tool, the alignment tool suggests a crazy number at 60 Hz and the predicted graph looks poor ?

What am I doing wrong ?

 

[jtalden]

I don't follow the question but...
Charts 1 and 3 look to be favorable timing.
Chart 2 is not favorable as the phase is not aligned well through the XO range. If REW suggested this timing what buttons were used to get this result? Please post the mdat so this result it can be reproduced and investigated. Possibly some indication of the reason for the poor result will be evident.

In general some room setups are very difficult to analyze, REW as well as most other automated setup routines can sometimes provide poor timing suggestions due to the room influences through the XO and cursor position chosen. REW does provide the calculated SPL and phase charts to review the impact of the suggested timing and provides for manual timing adjustments to confirm that a favorable timing has been chosen.

The redirected bass from channel 1 or 2 is the correct bass response to measure. This is not the LFE (channel 4 bass response. It appears you likely measured the redirected bass correctly so this is just highlighted the correct terminology so as not to confuse other users.

 

[IWC Doppel]

Chart 1 is measured chart 2 is the estimate using he alignment tool before moving the curser to the 60 Hz crossover point and then hitting align phase at curser. Clearly this is wrong ? I id not measure the proposed setting as it look wrong and summing the LFE and C with no LFE match the measured result ?

 

[jtalden]

chart 2 is the estimate using he alignment tool before moving the curser to the 60 Hz crossover point and then hitting align phase at curser. Clearly this is wrong ?

Yes, the cursor must be placed near the XO frequency before selecting; align phase at cursor. In some cases the room effects can still cause the proposed timing to be poor so it is best to review the resulting SPL and phase charts as you did. If the charts look questionable we can change of the cursor position staying within the XO range and try again. This can sometimes result in a better REW proposal.

I id not measure the proposed setting as it look wrong and summing the LFE and C with no LFE match the measured result ?

I wouldn't have measure the system with the -24.83 ms delay change either. There would be no reason to.
I don't understand the question though.

 

[IWC Doppel]

I will do some more measurements today, my question I think you have answered which is the suggested -24.83 is not a sensible recommendation and the black graph should not show a dip. I had assumed that the black line on the sim was the predicted curve when any delay or gain is entered. In tother words you could manually move the curser and see what curve looked best, so for example if I added 2ms delay to the R channel (LFE or reduce the delay to R channel, to keep the C correct ) I could improve the crossover performance.See below, I also assume if I used the trace arithmetic A(R no LFR)+B(R LFR only) then this would be the 'perfect' addition of the curves with phase correct so I could see what I should have if phase is correct ?

 

[jtalden]

Yes, this is all correct.
I recommend the SWgroup delay timing be adjusted to best fit with the sum of the FL+FR. The delay of the CC is then best adjusted to match as well as possible to the FL+FR through the midrange.

 

[IWC Doppel]

Great, I am going to measure the curves at each seating position as well, and them listen to how things sound with the adjustments. For some unexplained reason if I add equal delay tot he subs (say 5ms each they no longer align, I cant get my head round this bit !

 

[jtalden]

For some unexplained reason if I add equal delay tot he subs (say 5ms each they no longer align, I cant get my head round this bit !

I don't understand that either.

 

[IWC Doppel]

My three subs aligned over the three seating positions well slightly wider over a 2m wide seat have delays of 5ms,3ms, 3ms if I measure them again just playing the three subs with say 10ms, 8ms, 8ms they do not align, so the graph that is now reasonably flat when all three are working together falls apart. When aligned it looks like the following

Three subs measured indivually at each seat, 5,3,3

The there subs playing as one in the three seating positions and an average 5,3,3

 

[jtalden]

How did it change when you repeated the second chart measurements with 10,8,8?

 

[IWC Doppel]

I might measure agin to see hoe it is now behaving but here are two much older measurements at 5,3,3 and 9,7,7

 

[jtalden]

That is much more than expected, but not too much difference in practice. Maybe the mic was in a different position or some of the various room influences changed between the 2 measurements.

 

[IWC Doppel]

I spent some time yesterday and here is the differences just measuring he subs in the central seating position, you can see differences even with 1 2ms added by the time you reach 5 quite significant.

 

[JStewart]

A 2ms delay is a 2ms delay. Should measure the same.
Are measurements using the same delays always repeatable?

 

[IWC Doppel]

The only changes are altering the delay for each channel using the Acurus, then press measure again. It’s been troubling me as to why, I have got a mini dsp to try. I did test it with the Acurus set to one channel then drove the subs with the same PEQs as I use in the Acurus. It came pretty close measurement wise and sounded very similar but not quite a good just using the Acurus, I didnt try adding additjal delay via the mini dsp as an alternative

 

[jtalden]

1. Post a copy of the Post-13 mdat containing only measurements 2-5 (as shown in the chart). Delete all the other measurements to reduce file size.
2. Test to answer JStewart's question. Take at least 5 measurements with no changes.
3. Question: Does the Acurus accept delay input directly as ms or are you converting the ms to cm or ft or other unit?

 

[IWC Doppel]

1. Hi files attached,
2. I can retest but this has been very repeatable and I have raised questions around this since July
3. The Acurus has delay in MS adjustments for each channel and has three separate sub channels. PEQ's for equalisation are input for each channel separately as well

 

[jtalden]

I loaded the 4 files into one session. I found nothing that explains the SPL difference issue.

I did find more questions however:
All 4 impulses start at about +2 ms. If the SWs delays were changed as noted then they would be offset from each other by the delay amount. Possibly you aligned them after the measurements were made and then saved them that way?

There is a lot of ringing prior to the start of the impulse probably from the effect of an FIR EQ filter applied to the LFE channel. This is not an issue if this is the case.

The ETC is inverted for some reason. The data is also at the very tail of measurement. I have no idea as to why this is, but @John Mulcahy may be interested in this so I point it out to him.

 

[IWC Doppel]

Interesting question regarding the delays I time aligned to the L and simply saved them so not sure about why the delay hasn't changed with timing adjustments. Id be interested in an explanation around the impulse and ETC I haven't used those before

 

[John Mulcahy]

The ETC is inverted for some reason. The data is also at the very tail of measurement. I have no idea as to why this is, but @John Mulcahy may be interested in this so I point it out to him.

Looks normal when I open the measurements.

 

[IWC Doppel]

Not sure this helps but looking at the impulse for L,C,R measured from the central seating position with my 'normal' C, L,R delays it looks like they are quite well timed ?

 

[jtalden]

Interesting question regarding the delays I time aligned to the L and simply saved them so not sure about why the delay hasn't changed with timing adjustments. Id be interested in an explanation around the impulse and ETC I haven't used those before

If you are still interested in pursuing this then:

1. Test to answer JStewart's question. Take at least 5 LFE measurements with no changes.
2. Test to see if acoustic timing is working properly. Take 5 measurements of a single SW at 5 different sets of SW delays as maybe 0, 5, 10, 15, 20 ms. You should hear the chirp in the FL channel tweeter before each LFE sweep.

Are your SWs coming out of standby mode or are they fully powered for all the the test measurements? This is just a thought as I don't know if this can be an issue or not.

The ETC has nothing to do with the SPL and delay problems. It is helpful in room treatment efforts.

The mains timing looks okay.

 

[jtalden]

Looks normal when I open the measurements.

John,
Just additional info FYI in case it pops up again somewhere.
I loaded the 4 files 3 times today - ETCs look normal.
Yesterday, in the first instance after reviewing the data in various ways, I saw the issue with the ETCs. I continued data review in various other charts and later returning to the ETC to see if there was change. After checking the ETC chart 2-3 times it did change to looking normal for the rest of the session. I eventually cleared and reloaded the measurements 2 more times to see if it was a transient problem in both cases the ETC look wrong and I could not get it correct itself by using other REW charts and controls. I cant remember if I restarted REW for sure, but I think I did at least once.

 

[IWC Doppel]

If you are still interested in pursuing this then:

1. Test to answer JStewart's question. Take at least 5 LFE measurements with no changes.
2. Test to see if acoustic timing is working properly. Take 5 measurements of a single SW at 5 different sets of SW delays as maybe 0, 5, 10, 15, 20 ms. You should hear the chirp in the FL channel tweeter before each LFE sweep.

Are your SWs coming out of standby mode or are they fully powered for all the the test measurements? This is just a thought as I don't know if this can be an issue or not.

The ETC has nothing to do with the SPL and delay problems. It is helpful in room treatment efforts.

The mains timing looks okay.

Many, Thanks, I will do so later on today.

The subs are left on between any measurement and are simply turned on individually by a switch they are Mosfet powered 500W 4Ohm amps fed directly from the Acurus ACT4 that has three individual sub connections that have db adjustment, delay in ms, phase 0-180, 4 PEQ filters they then feed each sub amp then they feed each sub. All phase is set to 0, I keep gain adjustment to trim only as the sub amps have volume

 

[IWC Doppel]

Not sure what to check with timing but here are LFE all repeated

My timings are C 0ms, L 1ms, R 1ms, SL 3ms, SR 3ms, Sub 1 5ms, Sub 2 3ms, Sub 3 3ms

Time using the left channel, chirp before and after. LFE 2 Xms means Sub 2 only at Xms delay total, LFE A means all subs played and repeated with timing 5,3,3