Keith, I love you. you are awesome.Thanks for giving back and helping our next generation. I'm just and old FOH A1 in Vegas and have been sending my techs to your videos to help them out. Keep it up, you're making a difference. Peace to you.
Thanks for your videos.. i was a student/graduated in 2006.. been working in live sound, post production & system installations ever since. Thanks to everyone at CRAS for your work! God bless
The method you describe is pretty solid with delay. The phase inversion doesn't have to be exclusive of delay, either. I tend to not invert, because (in outdoor scenarios), I will often try to employ Endfire-type techniques for sonic containment. This requires at least 4 points of sound, though a 4.3 system (tops in each corner, summed sub array in front, rear corners are full range) has worked best so far. Essentially, you measure from the center of the space, first aligning the front tops with the front sub array, then aligning each rear top with its respective subs. Once you have that done, you move to the rear line and you delay the rear stacks until you are 90° out of phase from the front stack. Step back maybe a foot further away from the front stack. If you have adjusted things correctly, you will get a summing of the signals inside your sound space and almost total cancellation outside of if. You may have to fiddle a bit with your times, and the more bands of delay you have, the more profoundly you can create this effect. When I say fiddle, I mean tenths of a millisecond for mids and highs, so lightweight processing isn't going to do the job. It does work, I've done it, however, it's not easy, it's very time consuming, and I would only suggest even thinking about trying this if you own the system, or have a very supportive owner and team who are willing to give you the time to try and fail until you get the hang of this. If you can get it right, you absolutely blow minds because if you are outside of the field, you will question whether the system is even on, and then as soon as you cross the threshold, the sound hits you like a wall.
My question is "aligned from WHERE??!" Measuring from one spot/seat in the room will make for a varried results from other locations of measurement, depending on relation from seating locations vs subwoofS locations, and sub to sub respectively and in assocation??? Seams one sub will be close and one far from any given seating, so how do you balance and for Phase on that?! Seams challenging if impossible to balance for all seating locations in this particular instance. Besides my marbles, what am I missing here???
Its at 0. You need to have a 0 point to define all other mathematical things youre doing. Doesnt matter where zero is. It's like 'at X i measure this' . Ofcourse in this field of work there are sweatspots. Spots where phase cancels out in some frequencies. Standing waves with the room etc etc. But 0 = your absolute 0 point of measurement and aligment. Heck it can be 20cm in front of the subs. it can be 500meters in the distance . Doesnt matter where you put that after that the equations start.
Problem with polarity inversion vs. time delay in this case: what does polarity inversion do to frequencies above and below the crossover? It may be perfectly aligned at the exact crossover frequency, but is it combing out energy anywhere else outside of the crossover frequency? I think Merlijn van Veen has probably the best solution for a situation like this. Do separate measurements with both time delay and polarity inversion and not only check the phase at the crossover, but check about 1/3 octave above and below the crossover as well. (If your x-over is 100hz, 1/3 octave below and above would be roughly 80hz and 125hz, respectively). If polarity inversion causes weird buildup or cancellation of frequencies, then it's not the ideal solution (likewise with time delay).
Keith, you're a badass. Thanks for teaching me the skills to be successful. I'm still trying to nail good mixes fast but it'll come with time, I'm sure
Good video. But more people might have a deeper understanding if he defined his terms. Plot the trace, phase wrap, delay locate. These are all Smaart terms that are simple in concept but students may not know.
Well man., prstty damned amazing. He set up a sub real quick with out employing any music! To me, that’s a bit like a blind man calibrating my new OLED!
Keith, can’t knock you for making a decent video. However, you failed to show your students the full picture when it comes to possible differences between your polarity inversion vs. time alignment. For one, it was by chance (or perhaps by design for the sake of your video) that polarity inverting your subs aligned the phase traces. That’s going to be a shot in the dark for many deployments. A student should see what happens in the crossover region after you’ve applied the two different actions outlined in your video. With the polarity inverted on the subs, what does the measurement look like with the main array added back in? Do you have proper summation above and below the crossover frequency you were focused on? With both systems energized, do you have significant cancellation in the crossover region if you flip polarity back on the subs? How does this compare to the delay method? If the subs remain in relative phase and you time align them to the tops, do you have more or less summation in the crossover region? If you flip polarity on one of the systems in this case, do you have more or less cancellation. You may be misleading your students by not showing them this additional context for measuring and tuning crossover points. You might be “in phase” at that one crossover frequency by overall out of time. You may be giving away “free” energy of the systems aren’t properly time aligned through the crossover region. You should see distinct summation through the crossover region with both systems time aligned and and in proper polarity. If this summation produces undesirable frequency response, apply corrective EQ. This way your system design is benefiting from the “free” summation energy. You may have been there with what you demonstrated with the sub polarity flip, but I feel you need to demonstrate how to validate it with additional measurement.
Have you compared this method to using the "delay locator" method? If I wanted to align the tops to the subs or vice versa - I would want my crosshairs (cursor) to be exactly on the crossover point on the magnitude trace AND I would want the phase traces to align at that position as well. That is why you have the straight line going into to the upper screen with the crosshairs up there as well. They (Rational) made it easy for us to lock it in this way. It seems you are choosing a different point to align the boxes that is not at the down-point on the crossover where they cross. Can you explain further? In my mind when it comes to delay vs. polarity - I usually go with whichever option uses less delay.
Keith, I know this videos been up for a bit. Just wanted to ask how you were sure you weren't working off reflections when measuring the sub? Where was the mic here?
Nice job Kent. Some food for thought... not that I have a conclusive answer or the point is mostly academic... but in this scenario we're looking through the perspective of a continuous signal. What about a transient? Polarity or delay the better choice? Also, important for everyone to understand the alignment is only as good as the design allows the alignment to be consistent over as much audience area as possible. A whole other topic....
Thank you Keith.. Delaying or Polarity inverse would essentially be the same in this case, 8ms delay would be half cycle of 62.5hz which you spotted the alignment is needed. correct me if I am wrong , thanks in advance...
Great video, just a bit confused, is the oscillator still on 5.20seconds, just low that we couldn't hear it? (otherwise surely it's just matching the output time which I know you aren't doing) or is thee a way to monitor the differences without noise?
Keith, very good video! Thank you very much! My question is if the subs are really far from the main.. let say the the main is 10ft behind the subs, then do we have to find the delay time between sub and main first and then phase match them?? And how do we find that delay?
Why would you 180 invert? What about if you have a kick drum being played on PA, on initial transient array speakers will move out while subs will move in. Not ideal, in my opinion...
Fernando Ventura theres a video on youtube to crack smaart v6, in portuguese but its pretty easy to follow just looking at the screen. Ive done it, takes 10 minutes
Open Sound Meter. It's basically Smaart Light but also has some cool features like being able to move captured traces forward and backward in time. It's also free for Mac/PC and around $50 for iPad OS.
I wouldn't invert the polarity at all. In my experience, I got amplitude reduction significantly. Also to me it's much more easier by overplotting the impulse response of each individual driver.
Nice video Keith, thanks for sharing. I like the idea of introducing delay to emphasize the phase traces. Do you have a preferred bunch of settings for the FFT: window shape, window length, averaging? Can you think of any pitfalls that would result in an unreliable phase measurement?
First step to use smaart is to synchronize the reference jumper on it. I didn't see you do that. So I think you screw up Dude! This Is why your lines look so awkard.
@@snapascrew The only thing I heard him said was as long as Smart has a reference of time he doesn’t bother time delaying it. Well I as far as I’m aware of you need to time delay so it knows the distance between the speakers and measurement microphone. I just never seen it done this way. Why reinvent the wheel if is already round?
![How to phase align main+sub in Smaart® [GSwSST34.1]](http://i.ytimg.com/vi/5leb1k4LUA0/mqdefault.jpg)
Keith, I love you. you are awesome.Thanks for giving back and helping our next generation. I'm just and old FOH A1 in Vegas and have been sending my techs to your videos to help them out. Keep it up, you're making a difference. Peace to you.
Thanks for your videos.. i was a student/graduated in 2006.. been working in live sound, post production & system installations ever since. Thanks to everyone at CRAS for your work! God bless
The method you describe is pretty solid with delay. The phase inversion doesn't have to be exclusive of delay, either. I tend to not invert, because (in outdoor scenarios), I will often try to employ Endfire-type techniques for sonic containment.
This requires at least 4 points of sound, though a 4.3 system (tops in each corner, summed sub array in front, rear corners are full range) has worked best so far. Essentially, you measure from the center of the space, first aligning the front tops with the front sub array, then aligning each rear top with its respective subs. Once you have that done, you move to the rear line and you delay the rear stacks until you are 90° out of phase from the front stack. Step back maybe a foot further away from the front stack. If you have adjusted things correctly, you will get a summing of the signals inside your sound space and almost total cancellation outside of if. You may have to fiddle a bit with your times, and the more bands of delay you have, the more profoundly you can create this effect. When I say fiddle, I mean tenths of a millisecond for mids and highs, so lightweight processing isn't going to do the job.
It does work, I've done it, however, it's not easy, it's very time consuming, and I would only suggest even thinking about trying this if you own the system, or have a very supportive owner and team who are willing to give you the time to try and fail until you get the hang of this. If you can get it right, you absolutely blow minds because if you are outside of the field, you will question whether the system is even on, and then as soon as you cross the threshold, the sound hits you like a wall.
My question is "aligned from WHERE??!" Measuring from one spot/seat in the room will make for a varried results from other locations of measurement, depending on relation from seating locations vs subwoofS locations, and sub to sub respectively and in assocation??? Seams one sub will be close and one far from any given seating, so how do you balance and for Phase on that?! Seams challenging if impossible to balance for all seating locations in this particular instance. Besides my marbles, what am I missing here???
Its at 0. You need to have a 0 point to define all other mathematical things youre doing. Doesnt matter where zero is. It's like 'at X i measure this' . Ofcourse in this field of work there are sweatspots. Spots where phase cancels out in some frequencies. Standing waves with the room etc etc. But 0 = your absolute 0 point of measurement and aligment. Heck it can be 20cm in front of the subs. it can be 500meters in the distance . Doesnt matter where you put that after that the equations start.
Problem with polarity inversion vs. time delay in this case: what does polarity inversion do to frequencies above and below the crossover? It may be perfectly aligned at the exact crossover frequency, but is it combing out energy anywhere else outside of the crossover frequency?
I think Merlijn van Veen has probably the best solution for a situation like this. Do separate measurements with both time delay and polarity inversion and not only check the phase at the crossover, but check about 1/3 octave above and below the crossover as well. (If your x-over is 100hz, 1/3 octave below and above would be roughly 80hz and 125hz, respectively). If polarity inversion causes weird buildup or cancellation of frequencies, then it's not the ideal solution (likewise with time delay).
Lmao, hey buddy.
Keith, you're a badass. Thanks for teaching me the skills to be successful. I'm still trying to nail good mixes fast but it'll come with time, I'm sure
What crossover are you using? I didn't hear that mentioned anywhere.
Thanks for the video! Definitely a good review for the System Engineering practical I'll be taking tomorrow!
Sean Conklin
7:49 absolutely, most of the time there is few various solutions to the problem solving path. Cool video. Thanks
Keith is a great teacher!!! Thank you for sharing!
Good video. But more people might have a deeper understanding if he defined his terms. Plot the trace, phase wrap, delay locate. These are all Smaart terms that are simple in concept but students may not know.
Student here, a little disappointed at how over-my-head this is
From my experience as an engineer, if you don't use smart you're not even an engineer. Its a bit like saying you don't like Digico.
Well man., prstty damned amazing. He set up a sub real quick with out employing any music! To me, that’s a bit like a blind man calibrating my new OLED!
Should you use delay reference on the mains and or subs?
if you are talking about TF function delay in smaart than you should definitely use it prior to capturing first trace of tops
Keith, can’t knock you for making a decent video. However, you failed to show your students the full picture when it comes to possible differences between your polarity inversion vs. time alignment. For one, it was by chance (or perhaps by design for the sake of your video) that polarity inverting your subs aligned the phase traces. That’s going to be a shot in the dark for many deployments. A student should see what happens in the crossover region after you’ve applied the two different actions outlined in your video. With the polarity inverted on the subs, what does the measurement look like with the main array added back in? Do you have proper summation above and below the crossover frequency you were focused on? With both systems energized, do you have significant cancellation in the crossover region if you flip polarity back on the subs? How does this compare to the delay method? If the subs remain in relative phase and you time align them to the tops, do you have more or less summation in the crossover region? If you flip polarity on one of the systems in this case, do you have more or less cancellation. You may be misleading your students by not showing them this additional context for measuring and tuning crossover points. You might be “in phase” at that one crossover frequency by overall out of time. You may be giving away “free” energy of the systems aren’t properly time aligned through the crossover region. You should see distinct summation through the crossover region with both systems time aligned and and in proper polarity. If this summation produces undesirable frequency response, apply corrective EQ. This way your system design is benefiting from the “free” summation energy. You may have been there with what you demonstrated with the sub polarity flip, but I feel you need to demonstrate how to validate it with additional measurement.
Have you compared this method to using the "delay locator" method? If I wanted to align the tops to the subs or vice versa - I would want my crosshairs (cursor) to be exactly on the crossover point on the magnitude trace AND I would want the phase traces to align at that position as well. That is why you have the straight line going into to the upper screen with the crosshairs up there as well. They (Rational) made it easy for us to lock it in this way.
It seems you are choosing a different point to align the boxes that is not at the down-point on the crossover where they cross. Can you explain further?
In my mind when it comes to delay vs. polarity - I usually go with whichever option uses less delay.
Keith, I know this videos been up for a bit. Just wanted to ask how you were sure you weren't working off reflections when measuring the sub? Where was the mic here?
Nice job Kent. Some food for thought... not that I have a conclusive answer or the point is mostly academic... but in this scenario we're looking through the perspective of a continuous signal. What about a transient? Polarity or delay the better choice? Also, important for everyone to understand the alignment is only as good as the design allows the alignment to be consistent over as much audience area as possible. A whole other topic....
done deal easy pici.....Thanks Keith
So, if you had NOT done the 180-polarity invert, how many ms of delay would you have needed to get the same result?
Perry Mike I believe he switched back to normal polarity before adding delay.
Sorry for taking so long to respond I think it was like 8ms
Thank you Keith.. Delaying or Polarity inverse would essentially be the same in this case, 8ms delay would be half cycle of 62.5hz which you spotted the alignment is needed. correct me if I am wrong , thanks in advance...
WOW So it really has been a while since ive watched audio related video and didnt understand a thing!
I love the way you explained it, please do a Video about PA basics
Great video, just a bit confused, is the oscillator still on 5.20seconds, just low that we couldn't hear it? (otherwise surely it's just matching the output time which I know you aren't doing) or is thee a way to monitor the differences without noise?
Keith, very good video! Thank you very much! My question is if the subs are really far from the main.. let say the the main is 10ft behind the subs, then do we have to find the delay time between sub and main first and then phase match them?? And how do we find that delay?
Why would you 180 invert?
What about if you have a kick drum being played on PA, on initial transient array speakers will move out while subs will move in. Not ideal, in my opinion...
First you need to understand the concept of phase and what is happening with phase response on the crossover point and overlap ...
Kieth you were a great teacher student 2011
Great video. Thanks for the video. I have Smaart v6 how to alignment sub on smaart v6???
the implication here though is that delay versus polarity inversion have ultimately the same result?
only in some instances and distances
So if you don't have Smaart then what other option do you have?
Room EQ Wizard is free. But it does not have this live feature :/
Fernando Ventura theres a video on youtube to crack smaart v6, in portuguese but its pretty easy to follow just looking at the screen. Ive done it, takes 10 minutes
Sat live
Open Sound Meter. It's basically Smaart Light but also has some cool features like being able to move captured traces forward and backward in time. It's also free for Mac/PC and around $50 for iPad OS.
I wouldn't invert the polarity at all. In my experience, I got amplitude reduction significantly. Also to me it's much more easier by overplotting the impulse response of each individual driver.
hello there, the array is on on the booth sides LR?
no only one side at a time
Nice video Keith, thanks for sharing. I like the idea of introducing delay to emphasize the phase traces. Do you have a preferred bunch of settings for the FFT: window shape, window length, averaging? Can you think of any pitfalls that would result in an unreliable phase measurement?
Nathan Lively can help with that
I need to know problems with left and right subs
power alley
Great example! Next time maybe use different trace colors ;)
And if you're more than a full wavelength out?
Use delay. Time align the crossover. This is where the controversy he is speaking about.
He talked at the end about adding a random amount to get it in the ballpark...like 25ms on longer distances.
We need another video Keith
Why did you need all those people in the beginning??? Mad awkward
Out of all these comments here doesn't look like Keith has ever never replied to one single comment? Correct me if I'm wrong? 😢😮
Thank you
Thought this was going to be a Circle Jerks song ; )
Thankyou
Great video. Thanks for sharing...
666 thumbs up
69 thumbs down
Nobody TOUCH those buttons!
First step to use smaart is to synchronize the reference jumper on it. I didn't see you do that. So I think you screw up Dude! This Is why your lines look so awkard.
He mentions in the video why he did not do that. A couple of times.
@@snapascrew The only thing I heard him said was as long as Smart has a reference of time he doesn’t bother time delaying it. Well I as far as I’m aware of you need to time delay so it knows the distance between the speakers and measurement microphone. I just never seen it done this way. Why reinvent the wheel if is already round?