Home Theater Forum and Systems banner

REW windowing

Tags
windowing
4K views 5 replies 3 participants last post by  bcodemz 
#1 ·
Hi,

I'm trying to use REW to do some quasi-anechoic measurements. With ARTA and HolmImpulse the gating is straightforward. However, on REW I can't figure out what to do.

In IR windows, there is left window, window ref time, and right window. I simply want to apply a 5.5ms gate. If I put 0 for left window and 5.5ms for right window, I get a downward sloping curve that vastly underrepresents the treble output. If I use the default 125ms for left window and 5.5ms for right window, I get something comparable to ARTA and HolmImpulse but still doesn't seem quite right. And certainly the frequency resolution isn't 7Hz.

What should the settings be? What is this left window parameter?

Also, how do I measure the step response of the speaker for time alignment purposes? I can't see anything useful in the step response from the sine wave sweep.

Thanks
 
#2 ·
The step response trace is available on the Impulse graph, that is the step response of the system that was measured.

The windows are relative to the chosen reference point, which is typically the peak of the impulse response. Left and right can be specified independently as a symmetric window is often not the best choice for analysing the IR. If you set left to 0 and the reference is at the peak you will typically be cutting the HF part of the IR in half, hence the odd appearance if you do that. A common use of a narrow window is to eliminate reflections after the direct sound, they occur after the peak of the direct sound so the right window is the one that should be reduced in that case. The response is fairly insensitive to the left window width unless it is made so wide as to incorporate some of the distortion components that appear at negative time before the direct sound's peak or the measurement is of a low bandwidth system (e.g. a subwoofer) which has a correspondingly broad IR peak. You can see the effect of window choices by looking at the Window and Windowed IR traces on the Impulse graph.
 
#3 ·
Thank you so much for the explanation. I figured it out now.

As for the step response, I do see the step response, but it is not anything useful. A step response from a sweep wouldn't tell me if my drivers are time aligned or not since the tweeter doesn't start playing until the middle of the sweep. Is there another type of test tone that can better measure the time alignment of drivers?

Thanks
 
#5 ·
A step response from a sweep wouldn't tell me if my drivers are time aligned or not since the tweeter doesn't start playing until the middle of the sweep.
the analysis presented by REW is not a simple recording of the sweep, it is a characterisation of the behaviour of the system using a sweep to provide the signal. You can use the step response to understand how different drivers are playing together. For simply looking at time alignment then using a fixed timing reference (loopback or audio) and comparing IR peak positions is sufficient if measuring devices of similar bandwidth. It gets more complicated when you compare e.g. a woofer and a tweeter.
 
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