Can we export a measurement including the effect of the filters?

[mojozoom]

Here's what I'm trying to do:

I'm using REW to apply all-pass filters, and it works absolutely amazing for a single 2nd order all-pass filter.

The problem I have is that I'm cascading two 2nd order analog filters and they apply one at a time in the real analog world, whereas in REW the two all-pass filters appear to be combined into a single filter. So to properly simulate what I'm physically constructing I need to run the REW sweep, apply an all-pass filter to the data, then apply another all-pass filter to that result. Does that make sense?

So I need a way to save the results of the predicted phase measurement. Any ideas? Thanks!

 

[John Mulcahy]

Can you explain how you are using REW to apply all-pass filters? Not really following what you are doing at the moment.

That aside, the easiest way to generate measurements you can export that include the effects of filtering is to import the impulse response of the filtering you are applying and use the trace arithmetic features on the All SPL graph to multiply the measurement by the filter IR, the result can be exported. You will need to make sure the measurement and filter impulse responses are properly time aligned (no delay) to have a valid phase result.

 

[mojozoom]

Hi John, thanks for the reply.

I'm trying to simulate the effect of two AP filters in series, which would have me taking a set of data, applying one AP filter, somehow turning that result into a standalone measurement, and applying an AP filter to that. I think I just assumed that there's be an export involved in the process.

My application is phase alignment of car audio drivers. I've written some of it down here: https://drive.google.com/file/d/0B_rlEHWsb1tXbjBqN0w3SnBhNVk/view?usp=sharing

I followed your description and my results came out to the shape I expected, but off by 180 degrees, so I expect they weren't properly time aligned as you said. That's not really an issue for me, as I'm focused entirely on curve profile now and can alter the location of the curve in time with TA later as needed.

Thanks again!

 

[mojozoom]

John,

I was kind of struggling in getting REW to allow me to apply another AP filter to the "A times B" data. Even though I would highlight the "A times B" tab, when I clicked on EQ it would take me to the EQ for the impulse filters I had imported. Once I deleted the tab with the impulse filters everything worked as expected though.

I'm also starting to think that I may be incorrect in my assumption that REW is combining the two AP filters into into one big digital biquad. I'm thinking maybe the capacitors I'm using may just be out of whack.

Aahhhh, the joys of analog devices.....

 

[mojozoom]

Just to wrap this up, I've figured it out and it's entirely on me. REW does exactly what it should and can easily be used to model two AP filters in series by plopping the values for both into the EQ filters window at the same time.

The trick in the analog circuit implementation is that I have to apply them to the signal chain in order of the lowest Q first. I had mistakenly though just the opposite and that was my downfall. With the highest Q filter first the lower Q filters that follow won't apply at all, and it really can drive a person crazy.

John thanks for everything you've done for us. REW is not only the most amazing piece of software from a functionality standpoint, but the amount of detail and care in the layout and organization of the screens, fonts, etc. is almost impossible to find in software today. Really nice stuff!

 

[John Mulcahy]

Glad you worked it out, and thanks very much for the compliments. Still lots of room for improvement though, not sure I'll ever get to the end of the ToDo list