My graph please help...

[yswong]

First graph is the measurement.
Second is the corrected graph. Few huge dips. Is that bad?

[brucek]

You have taken a measurement with the sub and mains together. You need to disconnect the mains and do a sub only measure and then apply the filters. Once you've corrected the subs response, then you add the mains and do a measure to see the effects at the crossover (if any).

You really only have two problems to deal with for equalization. The peaks at ~42Hz and ~56Hz require filters to bring them down to the target level. It's a waste to use filters after ~100Hz for those narrow dips.

BTW, use a vertical scale of 45dB to 105dB for your next graphs...
brucek

[yswong]

hi brucek

This measurement is only with the sub. I connected the BFD directly to the sub. It did not go through my AV. I'll try to measure the response with my mains too later on. Thanks a lot

[brucek]

I connected the BFD directly to the sub. It did not go through my AV

Ah, OK. It's really not the best idea to do that since you won't have the bass management crossed sub signal that you would normally be using.
Feed the receiver AUX or CD input and put the receiver in stereo mode with the mains set to small using your regular crossover and have the mains disconnected and give it another try.....

brucek

[yswong]

ok. I'll try again and post the graph again. Thanks!

[yswong]

Here's my new graphs going through my AV. I measured them with my sub in a 2 different positions.

Graph 1 - Position 1
graph 2 - position 2
graph 3 - filtered graph 1
graph 4 - filtered graph 2

I'm still deciding which graph to position. Those filtered graphs are what REW suggested. I didn't modified anything. Graph 1 few dips. Graph 2 is OK but the response between 30 to 40 Hz are too low.(I with they are higher) I wonder if I can bring them to target response . If yes, how do i go about doing it? I'm still learning how to play with the filters. Brucek, you were saying using the filters above 100Hz is a waste, so should i just turned those filters off in EQ filter?

[brucek]

I think the graph #2 looks pretty good (purple dotted).

As mentioned before, best to use a vertical scale of 45dB to 105dB for your next graphs.

After you've taken a measurement and you want to be more aggressive with the filtering, you can raise the level of the measured line to include a bit more above the target line. Use the Trace offset thumbwheel in the left panel pulldown called Trace Adjustment. You can click the Add offset to data button to permanently add the data from the thumbwheel and zero the offset after that's done. Once that's done, the filters will be more aggressive since REW sees more of the signal above the target. (Alternately, you can achieve the same thing by lowering the target after a measure, but I prefer the data offset).

This can be iterative. Raise or lower the data measure line as described above, and then create the filters and Optimize them. Then, if you don't like the result, clear the filters and peaks and repeat after you've adjusted the measure line again until you're satisfied.....

Tip: click the up/down arrow symbol in the EQ filters panel to sort your filters from low to high. Makes it easier.

you were saying using the filters above 100Hz is a waste, so should i just turned those filters off in EQ filter?

Usually, there's not a lot of overall effect of filters above 100Hz after the mains are added in, so you simply don't load them into the BFD. You can turn them to None in the EQ filters panel or if you're entering them by hand, simply don't load the ones above 100Hz. reserve any filters at this frequency and above until you've added the mains, and then they may be useful for tweaking around the crossover..

brucek

[yswong]

thanks for the great tips brucek. I tried to be more aggressive and using the trace offset and it turned out that only one filter is below 100Hz. From the graph obviously there are a lot of peaks and dips below 100Hz and somehow REW sets the rest of the filter above 100Hz. Does it work if I set the Find Peaks setting between 20 to 100Hz? I tried it and now all the filters are below 100Hz. Does this do any good?

More questions, is it a good idea to add more filters manually since some filters are unuse? Would that screw up the whole thing?

[brucek]

Does it work if I set the Find Peaks setting between 20 to 100Hz? I tried it and now all the filters are below 100Hz.

Yep, that will work fine to limit the filter assignments.

is it a good idea to add more filters manually since some filters are unuse? Would that screw up the whole thing?

If extra filters are needed, there's no reason not to add more or even alter the suggested ones a bit. Adding filters to get a ruler flat outcome is not required though. The less filters the better. You won't notice a few dB up or down, so don't get crazy about making the result perfect.

Always remember to 'Optimise PK Gain & Q' and 'Adjust PK Gain' buttons after each change or addition to a filter.

Remember to take a look at your waterfall after a measure to see how your filters are doing with respect to decay. It's useful to overlay the before and after waterfall (and make use of transparency feature) to see the filters effectiveness.

brucek

[yswong]

I'm learning all these stuff. Do we have an article around here that explains waterfall? I've seen those graphs but I dont really know how to read them. Thanks.

[brucek]

Do we have an article around here that explains waterfall? I've seen those graphs but I dont really know how to read them

There is some information on waterfalls in the REW Help files, and I suppose the internet could google up a mess of great stuff.

Basically it adds the third dimension of time to your response graph.

The waterfall is derived from the impulse response by shifting the impulse response window to the right by a proportion of the time range to generate each succeeding slice. It has to be generated before you can observe it, so click the generate waterfall for each measure you want to see..

The slice slider shows you slices of time from the first response at zero time and then time is increased as you move to slice 30. So if you have the Time Range (ms) set to 300 ms and the resolution Window set to 300ms, each slice is 10msec after time zero. This shows what the microphone hears as time moves on from the initial sweep. If you have a resonance that tends to decay very slowly, you'll see it in the waterfall. This is what we're trying to reduce with the filters.

Create a waterfall of your measure and set the slider to zero. Then slowly move it out to 30. This is the decay of the signal in the room. It will tell you quite a bit where the room resonances are.

I don't have great pictures of this, but directly below is a waterfall of a perfect world system with no real resonance. See the signal decay to zero at about slice 13 (130ms) - beautiful. You see it's three dimensional. The regular horizontal and vertical is there, but now added is time in msec after the sweep.

Below that is a not so perfect system where you see that between about 20Hz-30Hz, there is room mode that is still ringing in the room way after the full 30 slices (300ms) have passed.. and you'll hear it too. Hopefully a filter will reduce this.

In the pic below that I show where you can overlay the before and after waterfalls and look at the results.................

Perfect world
perfect waterfall.jpg

Actual world with a bad resonance
real waterfall.jpg

Overlay the two
overlay.jpg

brucek

[Ivaols]

Thank you Brucek, usefull for me to