House Curves and Setting Speaker Levels

[adogand6kids]

I need you all to check my logic on a couple of things regarding implementing a housecurve. Once I have set my BFD filters to produce the house curve, I do not believe I can simply use subwoofer pink noise test tones to balance my sub against my mains. The housecurve makes one end of my frequency response louder(that is it measures louder) than the other. The pink noise test tone is made up of all the frequencies in the bass spectrum, so when it plays, it will measure at the SPL of the loudest part of the spectrum. If I use my test tones to set my sub level equal to my mains, I will be setting the mains to match the SPL of the low end of the bass response (because this is playing the loudest), when in reality I want it to match the high end of the bass response (where it crosses over with my mains). For example, if I have a house curve that offsets the bass response by 10 decibels from 30 Hz to 80 Hz and use test tones to set the Sub pink noise SPL to match the main speaker pink noise SPL (say 75 dB), then I will have the sub playing at 75 dB at 30 Hz, but playing at 65 dB at the crossover frequency (10 dB lower following the housecurve). The mains will be playing at 75 dB. Am I correct?

So, what do you do? Do you just set the Sub level using test tone pink noise and ofset the sub and mains by however many decibels are included in your housecurve? Or do you use REW and continue to measure and adjust sub level until it looks matched at the crossover point?

Second question. I have heard people saying that you can use your BFD presets to have a couple different sub responses. One preset can be used for a flat response and another can be used for a housecurve. That way, you can just switch from one preset to another to change your room response for movies or music. But, if what I said above is correct about matching levels, then if you set levels using a flat response you will be down 10 dB at the crossover when you switch to a house curve. And, if you set levels using a house curve, then you will be up 10 dB over the whole bass spectrum when you switch it over to the flat response. Again, am I right in my thinking here? If so, then it makes the process of switching between flat and housecurve more complicated since you would have to re-balance the subs and mains everytime you switched.

[brucek]

The pink noise test tone is made up of all the frequencies in the bass spectrum, so when it plays, it will measure at the SPL of the loudest part of the spectrum

But this would also be the case if you had no equalization and you had a large peak in your sub response. The sub versus mains speaker level setting using the processors test tones is a rough setting that can be accomplished much better with REW whether you have equalized with or without a housecurve. It is much more accurate.

Your second point is quite valid. If you've been aggressive at cutting at higher frequencies around the crossover (for whatever reason), and you have two programs where one has a lower output around the crossover, then you will have a step problem between the sub and mains. It's something to watch out for..........

brucek

[terry j]

Wade, I don't use REW for ht at all, and all this stuff about LF out etc makes my eyes glaze over because I have no reality on it, but I would imagine integrating the sub is the same as in the stereo case, and what I do is run the sweep up past the x/over point.

I would imagine you run up to 200 hz in any case, so set your sub, then put your mains in ( which you are recommended to do in any case in order to check around the x/o point) and do a sweep up to 200.

You can check the smoothness of the transition point quite easily then, and it bypasses the 'unfair weighting' of the signal strength that you and Bruce spoke of, and of which 'I know nutting' ( sgt schultz voice)

[Wayne A. Pflughaupt]

adogand6kids wrote:

The pink noise test tone is made up of all the frequencies in the bass spectrum, so when it plays, it will measure at the SPL of the loudest part of the spectrum.

Yes and no. The meter is only going to display what it “sees” as the hottest frequency, whatever that may be. You noted that you have created an upward slope from 80 to 30 Hz. Using the old correction values for the Shack meter (I think they have been somewhat discredited in recent years, but we’ll use them here anyway for the sake of argument), the meter starts to roll out below 50 Hz. So with a house curve, assuming a picture-perfect one (big assumption), the meter would be reflecting what it “sees” at the highest level. That would be 50 Hz, since equalizing is dropping everything above that point, and the meter’s response is dropping everything below that point. The meter is down 3 dB at 31.5 Hz. Your house curve may or may not be fully making up for that – you’d have to have a greater than 3 dB increase at 31.5 Hz compared to 50 Hz for it to register on the meter as the indicated SPL reading. Otherwise, the meter is reflecting 50 Hz with a house curve.

But even this isn’t in stone. The problem with pink noise is that it’s something of a random signal. You’ll notice on an SPL meter that the reading will jump around a lot compared to say, a steady-state sine wave tone. With a broadband pink noise signal, all frequencies will be jumping around like that. You can see on a real time analyzer - at any given second, one frequency will be hotter than another (assuming a baseline of flat response).

If I use my test tones to set my sub level equal to my mains, I will be setting the mains to match the SPL of the low end of the bass response (because this is playing the loudest), when in reality I want it to match the high end of the bass response (where it crosses over with my mains). Am I correct?

It should be obvious at this point that with an SPL meter, it’s impossible to do a clinically accurate level setting. If that’s your goal, your best bet would be to use a real time analyzer. As Terry mentioned, REW can work, but it would be better for this to use coarse smoothing – between 1/3 and 1-octave.

Not to sound rude, Wade, but IMO you’re really overanalyzing all of this. At the end of the day a house curve is 100% subjective. Personally I’d just toss the meter and set the levels by ear. If the bass sounds weak once you pop in a DVD or CD, then turn the sub up until it sounds right. If it seems like the upper bass frequencies are too strong, then go back in to the EQ settings and tame them as needed. And so on.

Second question. I have heard people saying that you can use your BFD presets to have a couple different sub responses. One preset can be used for a flat response and another can be used for a housecurve. But, if what I said above is correct about matching levels, then if you set levels using a flat response you will be down 10 dB at the crossover when you switch to a house curve. And, if you set levels using a house curve, then you will be up 10 dB over the whole bass spectrum when you switch it over to the flat response. Again, am I right in my thinking here?

That could be a problem since the BFD has no provision for adjusting signal levels, much less per-preset. Your best bet would be to EQ your “flat” curve along the regular target line, as you did the house curve along that target line. Make sure both target lines are at the same level where they peak, and you should be set. If that doesn’t do it, your only option (and it’s admittedly a lame one) is to tweak equalizer filters as defacto level adjustments.

Regards,
Wayne

[Blaser]

I have a question: Is it possible to set a house curve target line with REW, by entering the amount of boost at the desired frequency?

Thanks!

Blaser

[brucek]

Is it possible to set a house curve target line with REW, by entering the amount of boost at the desired frequency?

Yes, that's how you set up the house curve. See the REW HELP files

brucek

[Blaser]

Got it, thank you!!

[Ayreonaut]

Wayne A. Pflughaupt wrote:

At the end of the day a house curve is 100% subjective. Personally I’d just toss the meter and set the levels by ear. If the bass sounds weak once you pop in a DVD or CD, then turn the sub up until it sounds right. If it seems like the upper bass frequencies are too strong, then go back in to the EQ settings and tame them as needed. And so on.

Right on. I was reading this thread and thinking the exact same thing. House curve is about what you want to hear.

For whatever it's worth, I think pink noise works just fine.
You can use it as a starting point and adjust to taste.

You are the boss. Don't let your graphs push you around.