I was agreeing with you. But not where you see that phase and delay are different, well yes of course they are but changing either of them for one sub relative to others changes the result. That wasn't the point was it? I thought the point was which one to prefer for practical results?
Last edited:
Yes, you are correct. The point is to get the best possible result with multiple subs.
What suprised me that many of the happy "Geddes method" users were not messing with phase or delay. They just moved the subs around and used some PEQ. Now that's interesting and completely new to me as I bothered with the delay all this time.
Just to clarify: I didn't imply phase and delay are different. At the en of the day they serve the same purpose. AFAIK delay may offer more accurate settings while it's not frequency dependent.
What suprised me that many of the happy "Geddes method" users were not messing with phase or delay. They just moved the subs around and used some PEQ. Now that's interesting and completely new to me as I bothered with the delay all this time.
Just to clarify: I didn't imply phase and delay are different. At the en of the day they serve the same purpose. AFAIK delay may offer more accurate settings while it's not frequency dependent.
That is going to be the more popular method, but yours is to be preferred.
Recently I stumbled across a post describing a utility for calculating filter settings given a measurement of one sub and the other subs, so clearly there is an interest.
PEQ are not effective on sharp dips, but broad ones are OK. Just to be clear phase and delay are not the same thing. WIth delay the phase rotates more and more as the frequency goes up, I assume that "phase change" means the same phase at all frequencies. AT any rate, the most I have ever done with phase is to reverse the polarity. I have never used phase or delay, it hasn't been required. And I don't move my subs around. I fix them in place and measure each one then flatten each as much as possible and sum. Then try to flatten this result with polarity and PEQ. A sharp dip will be a problem and may need special attention, but I have not seen that in practice.
To experience what I mean about phase changing everything consider this. I have a system where several subs add to a peak. I use a PEQ to tame that peak. Now change the center freq of the PEQ slightly and you will see larger changes at the sides of the PEQ than at its peak. That's because the phase changes of the PEQ are greatest at the sides of the peak rather than at the peak itself.
To experience what I mean about phase changing everything consider this. I have a system where several subs add to a peak. I use a PEQ to tame that peak. Now change the center freq of the PEQ slightly and you will see larger changes at the sides of the PEQ than at its peak. That's because the phase changes of the PEQ are greatest at the sides of the peak rather than at the peak itself.
Last edited:
Corners are not perfectly correlated, they are partially so. Hence two subs in opposite corners is only partially reduced in effectiveness, a third would be much more so, and a fourth corner adds almost nothing. Hence, corners do work, just not as well as using a more random placement (if that is possible.) Don't let details interfere with the main goal of a smooth LF response. If corners is all you have then use them rather than not.
Since I only have one big stereo amp for my four subs. I simply place them four different places and only reduce the one rise in FR that seems controlled by the room.
I dont change phase in relation to my mains, since I understand that phase under 100hz is a non issue and my mains run down to around 60hz.
I dont change phase in relation to my mains, since I understand that phase under 100hz is a non issue and my mains run down to around 60hz.
Absolute phase is irrelevant, but the phase difference between subs is critical. For example, two subs in corners will cancel one set of modes, but amplify the other. Switch the phase of one and the exact opposite will occur, so clearly relative phase is a big effect.
How the subs interact with the mains depends entirely on the relative locations. It's all too complex to imagine what will happen, which is why one must have some measurements. What I hear people do all the time is to use a crossover between the mains and subs. This is clearly NOT what I do or recommend. They must all play together in the overlap region.
Last edited:
Yeah - I know
I use a filter too, but only a second order to protect my mains from the very low frequencies under 60hz. Then I cross my subs at 90hz with a steeper filter, to supress the unwanted harmonics from the subdrivers.So my mains and subs should overlap nicely. For years I actually simply crossed at 90hz with a 48db LR between subs and mains. Still worked better than what I heard most places
I will try and post some measurements, so that it's not just talk
A software capable of automated integration of 4 subs is about to be released any day now. It is called Dirac Live Bass Management.
Thank you, Earl, for clarification. I read many posts with successful outcomes and was wondering how long do you need to make a proper integrations? Some say a couple of hours, others say months ...
Reading at your posts this can be done very quickly. Am I missing something?
Thank you, Earl, for clarification. I read many posts with successful outcomes and was wondering how long do you need to make a proper integrations? Some say a couple of hours, others say months ...
Reading at your posts this can be done very quickly. Am I missing something?
It took me about 4 hours once I had the measurement system setup (that takes a lot of time.) I measure each sub and the mains and use EQ to smooth each independently. I was doing this in real time looking at the traces. Then I sum them up. I try flipping phases to get to the smoothest response without a lot of deep holes. Then I try to correct the sum with overall EQ (all sources.) If this doesn't get me where I want I might go back and tweak each individual sub.
I think that maximum integration, both subjectively and objectively, comes from a large overlap of mains and subs. If I had small woofers in the mains (mine are all 15") then I might use a 1st order HP on the mains. In my installation there is no high pass on the mains, they are full range, but with limited LF response because they are sealed boxes. Generally each sub has a LP filter, of course, but not all at the same frequency. The individual measurements will indicate where each is best HP'd.
And let's talk about what the final curve should be for a moment. What we all tend to think of as good bass is that we hear in a large venue. In a large venue there are no modes to worry about and the FR is flat. BUT, and this is the main BUT!, large venues have long reverberation times. A longer signal will always sound louder to an extent. Thus, in a small room we have two problems. One is the discrete modes and the huge FR variations - multiple subs help this out. But the reverb time is still much lower and so the bass signals decay much faster. To yield the same perception of bass we need to boost the LFs to compensate for the small room. I use about +3 dB at about 40 Hz. In my room < 40 is a mess anyways. You just cannot dampen that very lowest mode which in my room is just below 30 Hz. This gives me a nature boost taking the system down to about 20 Hz.
Last edited:
Thanks for the very full answer Earl, I hadn't considered the need for bass boost like that. I have a major room mode at 40Hz which requires cutting, but otherwise not too bad, I have a couple of U frame subs which I've positioned for pretty smooth response, and have just recently installed a IB sub in the front corner of the room which on it's own is giving good results too. All in all, I've been lucky, I know it isn't always so easy, but then I'm only after one sweet spot.
Not too hard to do above the Schroeder frequency FS, but very difficult in the modal region. Above Fs this power level is not really position dependent, but below that it is very position dependent. But then everything depends entirely on the specific room! Hence there really is no global answer to the question.
Above Fs one can calculate the room power with a simple equation relating the source DI, reverb time (the dissipation) and source power. I don't have that equation off hand but most books on room acoustics will have it. It is a simple energy balance equation.
Does this make sense? If a speaker is designed to be put into a corner, it could be measured in that corner. A suitable measurement scheme would have to be developed to take in the space properly. Outside this space could be zeroed in the process of establishing DI.