EQ'ing my new IB

[PeteD]

First post here, but have been lurking for some time. Great forum!

I have played with REW with some manual imported readings and I greatly enjoy the interactive adjustments that can be made to the response curves. Truly a fun and useful program.

Anyway, my BFD should arrive tomorrow. I am not configured to use REW to test my response, yet, but I am going to do some manual measurements and generate some filters to enter manually (for fun and to get my IB to a listenable state), while I finish acquiring the components (sound card, cables, etc.) and climb the learning curve of the program. Anyway, it sounds like manual entry will be with me for the foreseeable future because of the Midi issue with the new BFDs. I am sure there will be many more along the way, but my first questions relate to the general procedure for sub EQ'ing. Let me know if I have the general procedure right.

1. Measure response with BFD in chain (necessary?) and the subwoofer plugged in, other speakers not. During this, the anticiapted bass management settings for the receiver will be active. i.e. I will have my crossover for the mains set at 80Hz and my crossover for my surrounds set at 100 Hz. I assume this will limit the amount of EQ I will generated above my crossover, if it is not engaged during sub testing.
2 Correct my measurements with the .cal file provided on this site for my SPL meter (4050).
3. Import my corrected data into REW, checking off that I used "C weighting", which I did.
4 Set up REW as indicated in the help file, including loading my desired house curve and selecting the BFD as my EQ.
5. Find peaks and then manually tweak and adjust these peaks (I have done this already with preliminary measurements).
6. Enter the filters into my BFD.
7. Measure the response again, with only the sub active still (?). If looks good, continue on. If not, adjust filters in REW again.
8. Plug in one (both?) of front mains and measure response of main and sub together. Adjust filters, as necessary to correct peak or dip at crossover (maybe adjust crossover frequency?) Use readings to adjust SPL of sub and main speaker to achieve similar response, generating smooth curve across frequency plot.

Any corrections or tips on things I may be missing would be greatly appreciated. Of course, I will post curves as I generate and tweak them.

Link for my IB project, if interested:
http://tinyurl.com/yd4fzd
I have already used the REW to generate filters (all cuts) that really helped my curve. Hopefully they will translate to reality. I will be performing 1 Hz measurements from 10 through 100 at least to tweak my response better than 1/6 octave. Painful, yes. Worth it, I hope so!

Thanks,
Pete

[brucek]

1. Measure response with BFD in chain (necessary?) and the subwoofer plugged in, other speakers not. During this, the anticiapted bass management settings for the receiver will be active. i.e. I will have my crossover for the mains set at 80Hz and my crossover for my surrounds set at 100 Hz. I assume this will limit the amount of EQ I will generated above my crossover, if it is not engaged during sub testing.

You don't really need the BFD in the chain, but why not have it there. Simply select the bypass mode. Since you're doing this manually it would be nice not to disturb your receiver/sub levels before and after filter entry, so you can do a proper comparison on the effectiveness of the filters with a before and after plot. That would be hard to do without the BFD in the chain.
The surrounds have no effect here. We equalize the sub first, and then add the mains only in stereo mode to check the crossover interaction.

8. Plug in one (both?) of front mains and measure response of main and sub together. Adjust filters, as necessary to correct peak or dip at crossover (maybe adjust crossover frequency?) Use readings to adjust SPL of sub and main speaker to achieve similar response, generating smooth curve across frequency plot.

I would use both mains, some use one. Feed the receiver from REW with a Y-splitter to the AUX or CD input and you'll have both mains.
The crossover is usually determined by the quality of the mains (usually 80Hz). They've got to be really good to select 60Hz...

I will be performing 1 Hz measurements from 10 through 100 at least to tweak my response better than 1/6 octave. Painful, yes.

I did it for years. Not so bad. But once you see it all take place in a matter of seconds when you get REW finally working, it will make you weep...

brucek

[PeteD]

brucek wrote:

The surrounds have no effect here. We equalize the sub first, and then add the mains only in stereo mode to check the crossover interactionbrucek

Should I send full signal to the sub during initial calibration measurements? i.e. If I use cross over at 80Hz, will I trim some potential LFE response. I saw in the BFD guide Sonnie talking about LFE having 120Hz content.

Thanks for the feedback,
Pete

[brucek]

Should I send full signal to the sub during initial calibration measurements?

No, just set and use the crossover you'll normally use and then equalize.

Yeah, the LFE does indeed extend out to 120Hz in the Dolby spec, but my understanding is that the sound engineers roll that off themselves to integrate it more smoothly into the standard 80Hz LPF (low pass filter) crossover provided by all home processors. The actual spec'd information when given to the sound engineer goes all the way to 120Hz and then is digitally cut off. This would be horrible to listen to if sent to a sub, so I guess it's their creativity to roll it off (and a smart move).

Since the 'small setting' redirected bass is normally rolled off from 80Hz (when that's the crossover you choose) at 24dB per octave, I suspect that the LFE channel is given similar roll off characteristics. At 24dB per octave, the amount of signal left at 120hZ is very, very weak. (see the graph above of the perfect sub).

Myself, I wouldn't be concerned about any differences in setup because of the LFE channel.

brucek

[PeteD]

If my mains are crossed over at 80Hz and my other surrounds are crossed at 100Hz, what should I use for my "cutoff" in REW?

I did some quick and dirty EQ with my BFD using some old full scale measurements (no crossover) to get the feel of using it and got some pretty good results for my first attempt.

[brucek]

If my mains are crossed over at 80Hz and my other surrounds are crossed at 100Hz, what should I use for my "cutoff" in REW?

Yeah, that's a good question. I would use the mains crossover of 80Hz, since you'll be in the stereo mode when you equalize. I don't really know any other way to do it.

I've thought about this myself, since I have my mains and surrounds crossed at 60Hz and my center and rear back crossed at 80Hz.

I use an REW target of 60Hz since I equalize with my processor in stereo mode, that's the crossover used for that mode with the two mains. But once I have filtered with my sub only and then added my mains to check the interaction, I also turn on 6 channel stereo to see how it looks with the mixed crossovers. Seems fine....

brucek

[PeteD]

I have noticed that the plot of my manually measured data looks different than in the Excel sheet when imported into REW, with C-weighting checked as "Yes". My imported values have already been corrected, but are higher when imported into REW. Is this because of the "C" weighting?

My curve slopes off in the Excel spreadsheet, but looks more level in REW. Is REW more accurate for some reason beyond the SPL meter correction factors? Do the SPL correction factors in the Excel sheet take the "C" weighting into account?

Here is my plot in Excel:


Here are these very same numbers from Excel plotted in REW versus my house curve:


Obviously, things look better in REW, which makes sense, since I created the filters in REW. Why the difference? Am I doing something wrong?

How does my curve look for sub only? I plan to tweak 25 Hz up a few dB...

[brucek]

I have noticed that the plot of my manually measured data looks different than in the Excel sheet when imported into REW

Yes, because if you're importing Radio Shack meter readings that you have already compensated for in the Excel chart and then subsequently apply C-Weighting in REW, the results will be incorrect.
Say no to C-weighting when you import Excel compensated readings. In fact that's the only way to do it, because if you had imported raw RS meter readings, C-Weight in REW wouldn't have been accurate because REW doesn't apply the calibration data.

Why the difference? Am I doing something wrong?

The main reason is writen above, but also notice that you have altered the vertical axis of the Excel graph. It should be from 45dB to 105dB to match REW. Also in the horizontal axis of REW you're using 15Hz to 200Hz, where the Excel graph is from 10Hz to 160Hz. You're comparing apples and oranges.

Make the change to the C-weight problem and make your graph axis the same and your graphs will be identical.

How does my curve look for sub only?

Perfect. You're done.

brucek

[PeteD]

OK, good point about the aixs. That does exaggerate things. The different actual values is what really threw me off. Plus the values were not off by the SPL correction values, so I figured something else was going on.

I will redo my filters, not applying C-weighting on import, because I have double corrected my values to generate my filters.

So, in the future, if collecting real time readings through REW, I must use the correct cal file for my SPL meter - so that REW corrects them properly?

I will post updated graphs after I redo my filters with the correct data.

Thanks!
Pete

[brucek]

So, in the future, if collecting real time readings through REW, I must use the correct cal file for my SPL meter - so that REW corrects them properly?

Yeah, when using REW to take a measurement, you need the correct calibration file installed and you also need the C-weight box checked to take care of any compensation that is outside the area covered by the calibration file (i.e. above 100Hz). The calibration files only go to about 100Hz, since after that the RS meters are fairly close to the C-Weight standard.

brucek