How to apply an Linkwitz filter to a sub?

[Chuck Gerlach]

I am trying to apply a HP LR filter to one of my subs, and a bit confused on how to go about it. I found how to "select" LR filters by changing the speaker type in the EQ Tab to driver and then select the filter I want.. So after I do that and then apply a HP filter in the filters tab, is it now an LR 12dB as I requested?

 

[JStewart]

I found how to "select" LR filters by changing the speaker type in the EQ Tab to driver and then select the filter I want..

So this is only affecting the generation of the target curve that the actual response can be Eq'd to to match.

So after I do that and then apply a HP filter in the filters tab, is it now an LR 12dB as I requested?

The HP filter selected on the filters tab is a butterworth type.
I verified this by:
1. Creating a flat measurement and setting it at 100dB SPL
2. Creating a target using LR 12dB
3. Comparing the created target curve to the "predicted" response. You can see the slopes are different.

If the target curve is changed to use a butterworth filter the target and predicted results are the same.

What is it you trying to do or see?

 

[Chuck Gerlach]

What is it you trying to do or see?

I have a 32" Ascendo sub in the rear of the room and 4 Ascendo 18's in the front. Currently, I run them both full range [the 32 goes out to about 1000Hz]. I want to experiment using a LP on the 32 and HP on the 18s - just out of curiosity.

 

[John Mulcahy]

REW is not an equaliser, so it can't affect the signals going to your subs. You can see the predicted effect of filters on measurements you have made by choosing the Generic - Extended equaliser and choosing the LP and HP filter shapes and slopes you want to experiment with, looking at the predicted trace for the result. You can also use the two crossover filter entries in the Generic - generic equaliser (the last two filters).

 

[JStewart]

You can also use the two crossover filter entries in the Generic - generic equaliser (the last two filters).

Thanks. Had no idea they were there.

 

[JStewart]

I have a 32" Ascendo sub in the rear of the room and 4 Ascendo 18's in the front. Currently, I run them both full range [the 32 goes out to about 1000Hz]. I want to experiment using a LP on the 32 and HP on the 18s - just out of curiosity.

Now that we know where to find the XO filters on the EQ Panel, you should be able to model this in REW and also check phase matching between the two and adjust delays or possibly add an allpass filter (if Trinnov supports them), if needed. Happy to try and help with the process, if you need any.

 

[Chuck Gerlach]

REW is not an equaliser, so it can't affect the signals going to your subs. You can see the predicted effect of filters on measurements you have made by choosing the Generic - Extended equaliser and choosing the LP and HP filter shapes and slopes you want to experiment with, looking at the predicted trace for the result. You can also use the two crossover filter entries in the Generic - generic equaliser (the last two filters).

I am aware that REW is not an equalizer. I am trying to simulate the results of using an LP/HP filter [specifically LR] to then be able to evaluate things like phase alignment, Group Delay, etc. I can use an iterative approach [various x-over, slope type, etc] using my Trinnov and then measure with REW but that will take a bit longer than forever.

 

[Chuck Gerlach]

Now that we know where to find the XO filters on the EQ Panel, you should be able to model this in REW and also check phase matching between the two and adjust delays or possibly add an allpass filter (if Trinnov supports them), if needed. Happy to try and help with the process, if you need any.

I still don't understand how to see the effects of applying LR filters nor how to apply them.

 

[JStewart]

So to do what you're looking for you'll first need measurements with the chosen XO filters applied. To do this n REW use the EQ window, apply the filters you wish, such as L-R high pass to the front sub group measurement. Then on the panel on the right find and press "generate measurement from predicted" control. Do the same for the Ascendo measurement with the Low Pass filter applied.

Now you'll have a measurement of each with the respective XO filter applied from which you can use trace arithmetic to combine them (A+B), or use the alignment tool to see the effects of delay changes on phase, or any other manipulations or REW charts.

An important note will be that the initial measurements must be taken with a timing reference for the relative phase data to be correct so that any math done with the measurements is correct.

 

[Chuck Gerlach]

So to do what you're looking for you'll first need measurements with the chosen XO filters applied. To do this n REW use the EQ window, apply the filters you wish, such as L-R high pass to the front sub group measurement. Then on the panel on the right find and press "generate measurement from predicted" control. Do the same for the Ascendo measurement with the Low Pass filter applied.

Now you'll have a measurement of each with the respective XO filter applied from which you can use trace arithmetic to combine them (A+B), or use the alignment tool to see the effects of delay changes on phase, or any other manipulations or REW charts.

An important note will be that the initial measurements must be taken with a timing reference for the relative phase data to be correct so that any math done with the measurements is correct.

Thanks. But how do I apply a specific filter type [e.g. Linkwitz Riley].

 

[ddude003]

This is the only mention I could find in REW help...

Equaliser Selection

Some of the miniDSPs also have an L-T...

And this suggests that there is more than one way to skin that cat...

Linkwitz–Riley filter - Wikipedia

en.wikipedia.org

 

[John Mulcahy]

The requirement is for HP and LP with L-R shape, not L-T. If a filter is to be applied there needs to be something in the replay path that is capable of doing it, either an equaliser or a device which offers EQ.

 

[JStewart]

Thanks. But how do I apply a specific filter type [e.g. Linkwitz Riley].

Reading @John Mulcahy ‘s post I’m wondering if I’m misunderstanding. So here’s what I thought I understood...

You want to take a measurement of your front sub group and a measurement of your large sub and apply XOs to the measurements in REW so that you can look at results in REW for a number of iterations without having to make an acoustic measurement for each iteration you wish to check. Then once you see a result that you wish to try, you’ll put the XOs into your Trinnov and I’m assuming evaluate the results afterward by listening or perhaps measuring to see if the result matches REW’s.

Here are some screenshots using measurements of my left and center, assuming I've got it right. And it's a requirement that any measurements used in this process be taken with a "timing reference", otherwise it doesn't work.

Starting with these two measurements from two speakers in my room.

Here I've opened the EQ for the left, clicked on the EQ Filters button on the top to open the Filters Panel and then scrolled all the way to the right of the Filters Panel to access the two available crossover filters. In this case filters 11 and 12 are all the way to the right of the panel and I'm using #11 in the example. As I understand it, this is the only way to access crossover filters. Once the filter is applied use the "Generate measurement from predicted" control that I've circled horribly in blue. Next ...

Returning to the measurements or All SPL tab you can see the measurement we created in blue from the original in red. I renamed the blue measurement.

Here's now also with the 2nd measurement lowpassed. And next were going to using trace arithmetic to combine them. This will provide the same result as if the XO's had been put in the Trinnov in the first place and then a measurement was taken.

And here is the result in Green. You can use the "Green" to look with any of the charts in REW.

Or better still, you can use the "Alignment" Tool to manipulate the high and low passed measurements to see the effects of delay and gain. Here, I've added 30ms of delay to the lowpassed measurement. You can see the droop it created at the XO (in black) and some messy phase between the two (in the bottom pane).

Here I've put the delay back to "0" for the lowpassed the phase aligns, which it should because I have have appropriate delays set in the processor already. Anyway, the point is you can use the delay and/gain sliders to find the results that give you the best support (phase alignment) at the XO. There's a button there to save the measurement and you can then use it to look at resulting group delay, decay, spectrogram, etc