Re: Phase align subs to mains
Sorry, busy week-end, no time for forums.
No you didn't. What you describe requires an analyzer of some type.
No it doesn't, not in a known system. Even when the system is unknown, some assumptions can be made about the response. These assumptions might not always be accurate, and I accept that there is a margin for error here. However, your typical front-loaded sub climbs about 3dB/oct from 100 Hz, and your typical top drops 10dB/oct below 100 Hz. In a smallish system it often stops there, there is no extra sub power, and the crossover frequency and the acoustic crossover are equal and around 80-100 Hz. End of story.
If we add some subs and don't do anything else, and we're getting 6 dB extra from the subs, we are shifting the acoustical crossover by an amount that can be calculated from the slope of the low-pass and the response of the sub. Assuming a 24dB/oct slope, acoustically we have around 21 dB/oct. If we have not changed the high pass of the tops, we'll match up somwhere where the tops are slightly louder, so our frequency shift is (6dB-1.5dB)/21dB/oct= 250 cents = 1.15 times the crossover frequency (roughly).
Now assuming that we have some extra subs, 10dB worth, and the subs themselves are lowpassed at 6dB/oct @ 65Hz to flatten the response curve, so they'll roll off at 3dB/oct above 100 Hz, the shift is going to be less, but likely there are going to be more phase issues, but an actual acoustic crossover is still relatively easy to calculate.
Basically, and this is a very easy method: With a 24dB/oct lowpass, the frequency will shift one semitone for each 2 dB added
One important factor that you forgot is that a loudspeaker itself acts as a bandpass filter, on top of whatever you send to it electronically from the DSP. So while you think all that you need to know is what is happening electronically from the output of the crossover to figure out mathematically what the frequency is where the two intersect, there is more to it than that. Apart from using a very high end speaker modeling software package, the ONLY other way to find this frequency is through the use of an analyzer. That is the ONLY way in which you can actually see the accumulated effects on both passbands, and see exactly where the *acoustic* transition is between the two.
I'm not forgetting the loudspeaker, and remember I was also referring to not using bandpass subs for some of this exercise, with a bandpass sub, like a 4th or 6th order job, and a totally unknown system where one doesn't know anything about the system, things certainly get very difficult to calculate. If it gets bad enough, one just have to A/B the bottom and tops untill one finds a frequency that is close. Still, in most real life situations, one will have a low-pass filter on the bandpass subs that is still defining the crossover frequency, and therefore also what happens in the quarter to half octave above the crossover frequency, bandpass subs generally mean steeper cut-off, therefore less shift.
Remember that systems interact most where they are equal in level. It is at this frequency that you want the two parts of the system to be aligned.
Yes, that is the acoustical crossover frequency we are actually discussing at this point.
The main problem with bandpass subs, filtering and shaping the response curve and generally doing anything, is the shift of phase that occurs, not finding the actual frequency. Typically, anything we do that shapes the upwards sub response and anything that moves the acoustic crossover upwards in frequency will make the phase of the subs lag further behind (become more negative, have a downwards slope on the curve), which in terms of alignment will mean we have to delay the tops more to get the phases to match up. This is of course a good thing in a way, it means we can start off with a delay that reflects the differences in physical distance (which can be reasonably accurately eyeballed) and then increase from there when listening.
Will the result be perfect? Of course not, you can only tune perfect for one location, but as with any other method, walking the venue and making sure there are no really bad spots with the setting that is chosen is important. In many ways, adjusting the delay between tops and subs is an exercise in steering the lobe at crossover frequency.
Tuning any system, no matter what the tools are, one really needs to know what one is doing and what one tries to achieve, and the methods for getting there.
It might for instance become obvious that the time alignment between tops and subs is not right when phases match, so a phase flip is in order. I believe that the subs shoud always be positive phase, so I'll elect to flip the tops if a phase flip seems appropriate.
I guess the question is, if you're already clearly having to set up and use an analyzer, why squint and fight with a single input RTA display when you can just use a transfer function? It's not like they were invented to make life more difficult. You can get a lot more information and look at it in more ways much faster than screwing around with the bars of an RTA. Certainly I understand what you're saying, that it's possible to get a rough sub/top alignment with just an RTA and some looking at the real world to make sure you're being realistic. None of us understand why you'd want to do this unless you just can't justify the cost of a real TF analyzer, in which case you still have to have all the skills that would allow you to use a TF analyzer in the first place to do this alignment method. Most people who can't afford Smaart don't need it and don't know how to use it. Clearly you know what's going on, so why screw around in the dark ages wasting your time?
I don't own a Smaart license, partly because it is fairly expensive for my needs, and partly because it doesn't really appeal to me as a tool.
I see people putting up loads of Smart traces on the forums without seemingly having a good explanation for the correlation between the real physics going on and the traces. I have heard very dull systems that looks good in Smaart and really awesome systems that doesn't look that great. Smaart and other advanced tools might convince operators to try and fix what doesn't need to get fixed. Like the purist preamps from the 70's and 80's that might just have a RIAA stage and a volume control, I believe less is more when it comes to tuning a system. I do sometimes play around with REW and have Blue Cat Frequency Analyst Multi running a comparative trace, and I have to admit that I've even sometimes been thinking that maybe Smaart could confirm or explain things I think I'm hearing.
However, most of the time I'm using either my own system that I know well or a venue system where I have no access, so I'm not in the typical pro Smaart user category at all.
~
