Home Theater Forum and Systems banner

Matching 2 SW

Tags
matching sw
2K views 13 replies 2 participants last post by  DomeTrust 
#1 ·
Hello

I have bought a second sub and used REW to adjust the delay on the closer sub (named sub 2). With my minidsp I put in a delay of 3.4ms on “sub 2” and that matches the real change in distance very close.

Both subs are pointing in the same direction and “sub 2” is on the left side and “sub 1” on the right side of the LP.

So the question I have is that the 100%FS peaks are not on the same side so would that be a problem?

Thanks
 

Attachments

See less See more
2
#2 ·
Looking at the side that the 100% peak occurs can be misleading. The SW1 IR is distorted a little due to room effects. This very normal, but makes it very difficult to determine what the best delay is by just looking at the IRs. We are usually pretty safe just using distance, but for some reason that doesn't appear to have worked out well in this case. I say that looking at the initial rise of the 2 IRs. I would expect them to be near to the same timing. Looking more closely, I understand you delayed SW2, but it actually is less delayed on the chart.
If you post an .mdat of the SW1 and SW2 at known distances/delay settings we can then look at the phase relationship of the 2 SWs and find the best timing by shifting the IR timing within REW.

  • Where are you adjusting the delays; Distances for 2 subs in an AVR?, MiniDSP?, Both?
  • Have you or will you be applying EQ? If so, MiniDSP only?
 
#3 ·
Thank you, I will make some new measurement on SW1 and SW2 (it appear some have been lost…).

Some info is that I only have one distance setting on the AVR (Yamaha 677) and I have it is set to 260cm (to make “good” integration with both main speakers) (SW1 is about 220cm from LP) so I then use a MiniDSP to split the LFE to both SW and then put a delay on the output of SW2 (SW2 is about 135cm from LP) to 3.4ms. I will also use the MiniDSP to EQ both SW.

Thanks
 
#4 ·
It seems that I made a mistake with the picture of the 2 SW with no delay and posted one that I made myself in REW (where I moved SW2 in the wrong direction) sorry for that. :sad:

So I manged to find some measurements on SW1 and SW2 without any delay that I post here now but I will do some new ones just to be safe.

Thanks
 

Attachments

#5 ·
Thanks for the file.
This file is still confusing. The posted .mdat shows SW1 IR arriving first and SW2 IR arriving about 3.6 ms delayed. Possibly you just have the labels reversed in the file? If the labeling is just reversed then that would explain the discrepancy. A 3.4 ms delay on the closer SW will indeed align it pretty well with the more distant SW as you indicated.
I adjusted for that 3.4 ms in REW, but unfortunately different room mode impacts on the 2 SWs do not compliment each other. The phase is very different resulting cooperation at some frequencies, but significant differences at others. The end result being that it would take pretty aggressive EQ to arrive at a smooth SPL target. It may be possible to get reasonable results that way however. I took a quick look to see if some different delay timing would result in a smoother SPL, but it doesn't appear that it would.

  • SW2 (per the file label) appears to more negatively impacted by its room position. If it is possible, try it to another location to see if that helps.
  • Another option is to apply some EQ to each SW independently to smooth each response somewhat. That could be done on the MiniDSP output PEQ bank for each SW. That may help with the cooperation of the 2 SW together. If still needed, a common EQ filter set can be applied on the input PEQ bank.
I'll wait to see if a location change is feasible and if I am correct that the labels are just reversed on the SWs, before I post any charts.
 
#8 ·
Thanks for info and the files.
The timing you posted does make sense to me now. I was confused last night both by the early IR signal start on SW1 (I still don't understand why that is) and I also didn't get the reasonable phase tracking results that I now find this morning with the 3.6 ms delay setting on SW2. Oh well, I do get confused easily.

Below are some charts with the 3.6 ms SW2 delay applied.
  • The phase tracking is very good from 42-100 Hz as can be seen in both the phase chart and the SPL chart. There is good cooperation between the 2 SWs. Below 42 Hz the room modes on SW2 disrupts the phase tracking leading to less cooperation. The very low bass is thus rolled-off a little more than rest of the range. If that is a problem requiring heavy EQ then a new location for SW2 is needed. It would be very unusual to have close tracking across the entire range with an asymmetric room setup.
  • The important part is the SPL is smooth enough to PEQ if you are looking for a fairly flat bass response. The null at 57 Hz should left unfiltered. The rest can probably be addressed with reasonable filters.

Green Text Line Plot Diagram


Text Green Blue White Line


Text Blue White Line Green


For next steps I suggest first confirm/adjust the XO timing (AVR Sub distance setting) to assure the XO range SPL is maximized. Finally, create the needed PEQ filter set for the 2 SWs. Do this as mono, i.e., using the same filter set for both SWs.
 
#9 ·
Thank you for the analysis and suggestions, I am going to follow your advice and test some other SW locations (may take some time) and in the meantime do some EQ on the current setup and have that as fallback setup under testing.

I was thinking of doing a house curve on the EQ or is it better to do a flat EQ and then raise the low end with “low shelf” filter?

When searching for a new location what should be my main thing to look at in the graphs?

Thank you :smile:
 
#10 ·
Thank you for the analysis and suggestions, I am going to follow your advice and test some other SW locations (may take some time) and in the meantime do some EQ on the current setup and have that as fallback setup under testing.
That sounds like a very reasonable plan to me.

I was thinking of doing a house curve on the EQ or is it better to do a flat EQ and then raise the low end with “low shelf” filter?
'Better', is the sound you prefer. I would definitely recommend experimenting with the SWs level after you have finished the EQ effort. Just change the SW gain to find the level you prefer. I EQ flat 20-100 Hz, but have then been known to occasionally increase the SWs 1-2 dB as a last step after listening to a new house curve. Others like to continue an upward slope in the bass range down to 20 Hz. There is no correct option. Just search for 'house curves' for background info.
See the info Here in the sticky thread that contains info on house curves as well as other good info. I just posted some of the curves I have tried Here , but there is lots of info elsewhere as well.

A 3-6 dB shelf filter at the bottom end for 2 SWs in a reasonable size room would not be expected to be an issue. Many people do this. Just be mindful not to overdrive the capacity of your SWs.

When searching for a new location what should be my main thing to look at in the graphs?
It's not easy to say what is best as there are several competing effects. SW2 has the most room mode effects, but the output at the low end is still much better than SW1. I guess the best way may be to use the current SW1+SW2 SPL measurement as the control. It should be the same as my calculated SW1+SW2(at 3.6ms) chart in post 8. For new trial locations just adjust the delay need according to tape measurements as you originally did and sweep test the 2 SW together. Compare the new SPL trace to the control. That will tell the whole story. Just pick the locations that best approach the house curve you intend to use.
There are more efficient ways if there a more than 2-3 options for locations. One method (assuming the 2 SWs are identical):
  • Sweep measure either SW in each location (it doesn't matter which SW is used as they are identical).
  • Record the distance from the LP with each measurement.
  • To compare a potential pair of locations, e.g., location #2 (2.5m) and location #5 (1.3m) just subtract the LP distances measured (=1.2m). Use that as the distance offset for the closer SW (#5 in this case). [We can convert to 3.48 ms if we like and use that number since that is value that would be needed in the MiniDSP.]
  • In REW manually shift the IR of #5 by the negative of that number, i.e., -1.2 m or -3.48 ms. [In REW we must use a negative number to delay the IR.]
  • Use REW 'Trace Arithmetic, A + B' to calculate the SPL of 2 SWs set at those 2 locations. Do that for each location combination and see which SPL curve best approaches the house curve.

A variation of this would have one of the SWs set at the LP and the mic is then moved to the potential SW locations for measurements. The analysis is very similar. That's another option I haven't tried. I suspect that there may be minor discrepancies this way. Possibly they remain trivial.

If you provide a file of sweep measurements and the LP distances associated with them, I can help with the REW analysis if needed. There is no need to measure locations close to each other as the wavelength is very large at the low end. I would expect the locations should all be at least 0.8 m distant from each other to be significant, but I haven't used this method and considering room mode effects I may be wrong about that.
 
#12 ·
So I have done some testing and for now SW1(XTZ 12.16) is moved and SW2( XTZ12.17) is in the same place but inverted. So I think I got good phase tracking like this and then added some EQ Rew did 4 filters and I added a 5th filter to boost a little at 33Hz, all in the .mdat.

Here are also one more .mdat that show SW2 at another location that have good SPL response but was difficult to match up with SW1, a null appeared at ca 40Hz.

Thanks
 

Attachments

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