Home Theater Forum and Systems banner

1 - 7 of 7 Posts

·
Premium Member
Joined
·
129 Posts
Discussion Starter #1
Dear John, you're incredible with your attention to detail in REW, and it shows. It's a true labor of love and my mind is blown. The help section is in itself a wonderful tutorial on the process of room analysis.

I started a thread long ago but am too lazy to find it so I'm creating a new one ;-).

I'd like to put in a reminder nudge for a variable window support. It would be wonderful to see REW confirm the true psychoacoustic response, as well as see the same response in REW as I see in Acourate's test convolution. For example, when I do a fixed window of 20 ms left side and 500 ms right side in REW, and no matter what smoothing, I see considerable frequency response difference compared to Acourate which uses a wide window at low frequencies, narrowing down to only a few ms. at 20 kHz. Acourate can be set up so at 1000 Hz the window is either 10 or 15 ms. and I think it starts near 500 ms. at 20 Hz so you can get an idea what I'm talking about. For some future release of REW, if possible.
 

·
REW Author
Joined
·
6,711 Posts
That's already on the request list Bob. The implementation of the windowing itself is pretty straightforward, a little more challenging is how to convey the variable-with-frequency time domain width of the window on the impulse response plots and provide suitable controls for the windowing, that needs a bit more thought.
 

·
Premium Member
Joined
·
129 Posts
Discussion Starter #4
That's already on the request list Bob. The implementation of the windowing itself is pretty straightforward, a little more challenging is how to convey the variable-with-frequency time domain width of the window on the impulse response plots and provide suitable controls for the windowing, that needs a bit more thought.
Good luck with that, John. I would be satisfied with a single vertical indicator in the impulse timeline indicating where the center of the variable window would be placed. I think it would be too much (impossible) to show the actual width of the window at every location. Then of course there is the definition of the left and right window size and type, which in your implementation, can be different shapes and types. I doubt that Acourate uses more than a single Hann or similar window around the impulse and I can't get my head around how it knows what frequency it is when it is working in the time domain. And if the window is centered on the impulse and symmetrical, does this mean he wastes 250 ms on the left side of the impulse to get an effective 250 ms. on the right side, which would be a waste? That's why I'm impressed by your ability to specify different window lengths on left and right side of the impulse. So little to do, so much time :).
 

·
Premium Member
Joined
·
129 Posts
Discussion Starter #7 (Edited)
Here's a small feature request. An example: I want to compare two impulse responses. The first is an impulse response of my left front speaker made when I was using all-analog processing, and so the latency is almost non-existent. The other is an impulse response of the same speaker made using Acourate Convolver, so the latency is very high. The first impulse occurs at about 9 ms, and the other impulse occurs at about 900 ms!

It would be nice to compare the look of the two impulses next to each other, offsetting one very slightly in time on the same scale so they don't overlap. I can offset an impulse in the impulse tab using the shift IR function. Then I can go into the overlays window and overlap the two shifted impulses. It would be nice to make it easier to shift each impulse in the overlays window and not have to jump back and forth until I get a satisfactory look. Attached is an example of the two overlaid impulses in overlays, after correction outside of the overlays window.
 

Attachments

1 - 7 of 7 Posts
Top