Home Theater Systems - Electronics and Forum - HomeTheaterShack - View Single Post - Attempts at using waterfall graphs with frequency response to set filters
View Single Post
Old 09-21-07, 06:18 AM   #1 (Link)
 
3ll3d00d
Shackster
Alias: Matt
Loc: UK
User: #951
Since: Jun 2006
Posts: 15
3ll3d00d is offline
Attempts at using waterfall graphs with frequency response to set filters


Hi

I read the recent thread on waterfall graphs (and using them to check that you have set filters accurately to deal with resonances) and thought I'd give it a try myself. The results were somewhat confusing so I thought I'd post them up to see what advice/comments people have.

The room is ~14' * 12' * 10', the frequency response is shown below. The blue line is the raw response and the red is the corrected response. The filters required are fairly straightforward and the REW suggestions work pretty well. The main issue is that peak at ~37Hz which REW suggested a -12dB cut @ 40-7 (IIRC) with bandwidth of 6/60.

frequency_response.jpg

The raw waterfall graph just confirms that there is a big hump at 37Hz

raw_waterfall.jpg

However when I generated the waterfall for the corrected response, 2 things stood out....

- the area from 31-45 is still ringing away for a substantial (albeit reduced) period of time
- there is a slightly odd looking ringing at 120Hz

corrected_waterfall.jpg

On the 1st issue I attempted various tweaks last night including minor bandwidth adjustments, moving the centre of the filter up and down a few steps and a short attempt at covering the hump with multiple filters. None of those attempts really reduced the decay and some of them increased it substantially (up to nearly 1s in one case) or introduced odd residual effects further down the frequency scale (a ringing from 600ms onto about 900ms down at about 20Hz). As a result I switched back to the suggested filter.

Am I right in thinking that that hump is actually 2 seperate resonances that need to be dealt? 1 at ~33Hz and 1 at ~38Hz and hence I need 2 narrower filters in place of 1 big cut? Scrolling through the slices seems to say that to me, for example this slice = 13 graph.

corrected_waterfall_slice_13.jpg

The 2nd issue (120Hz ringing) I have no idea about

Note that I haven't done any extended listening with these filters, I was doing it as an almost academic exercise to see if I can objectively tame the decay time first and only then live with it for a while.

All thoughts/comments welcome.

Thanks
Matt


Forum Rules Reply With Quote