Second guessing the engineers

[Jay Barracato]

It seems like second guessing the design engineers is a common thread starter.



Yesterday, I was playing with my HPR181 and 152's, trying to use SMAART to look closer at the crossover region. The HPR's have a built in crossover that is at 100 hz.



I started with the boxes stacked with the front edges aligned. It seemed my best measurement was at 1m on axis with the mic centered between the two boxes. I also tried ground plane at 1m and 2 m. I was getting consistent results with all three positiona. I finally determined a delay of 1.5 ms.



Then I switched over to the pole mount and determined the delay needed. You guessed it: 0 ms.



Lesson learned: If you are going to use a system that is designed to work together, it is probably best to trust the designers.



This is probably biased because I have been listening to these boxes with minimum processing for two years, but I didn't find any eq changes for the top that I thought sounded better than just leaving the processing flat.

 

[TJ Cornish]

Re: Second guessing the engineers

+1 - I'm in that same boat.



I bought the EAW JFL stuff rather than competing gear, because the JFL with UX8800 is a ''system''. I plug it in, dial for delay (if I'm doing a non-standard deployment of some kind), and maybe dial down a room mode. Other than that it's plug and play - no time wasted ''shooting the room''. I spend time fixing the issues that really matter, and the gig is better for it.

 

[Bennett Prescott]

Re: Second guessing the engineers

Then I switched over to the pole mount and determined the delay needed. You guessed it: 0 ms.

Measured where? That will make all the difference.

 

[Brad Weber]

Re: Second guessing the engineers

+1 - I plug it in, dial for delay (if I'm doing a non-standard deployment of some kind), and maybe dial down a room mode. Other than that it's plug and play - no time wasted ''shooting the room''.

Since I'm sure it will get mentioned, a room mode is something you can't fix with speaker or system processing, the best you could do would be to try to not drive it as hard. And since the room will vary it is not part of the 'system' processing. What is nice is that if the system processing is very good then for a simple system the room is about all you should have to address.

 

[Jay Barracato]

Re: Second guessing the engineers

Bennett, I was using the sub as the base for the pole mount, so there was no timing issues with differing distances between the two speakers. I ended up with basically the same results at 1m ground plane or 1m on axis, which I interpreted to be equidistant from the two speaker cones.



My logic is: if the measurement is taken at the same distance from the cones (at the point of an isolateral triangle) at 1m, as you go further and further back from the stack, the distances to each of the cones will approach the same value.



I was not looking at the interaction of two stacks, center clustering or any setup other than the two ways the speakers can be placed in a single stack.

 

[Ivan Beaver]

Re: Second guessing the engineers

+1 - I plug it in, dial for delay (if I'm doing a non-standard deployment of some kind), and maybe dial down a room mode. Other than that it's plug and play - no time wasted ''shooting the room''.

Since I'm sure it will get mentioned, a room mode is something you can't fix with speaker or system processing, the best you could do would be to try to not drive it as hard. And since the room will vary it is not part of the 'system' processing. What is nice is that if the system processing is very good then for a simple system the room is about all you should have to address.

And with room modes-if you do turn down the ''hot spots'', you will ALSO make the quiet ones (the ones not loud enough) quieter.

As usual it is a tradeoff.

 

[TJ Cornish]

Re: Second guessing the engineers

+1 - I plug it in, dial for delay (if I'm doing a non-standard deployment of some kind), and maybe dial down a room mode. Other than that it's plug and play - no time wasted ''shooting the room''.

Since I'm sure it will get mentioned, a room mode is something you can't fix with speaker or system processing, the best you could do would be to try to not drive it as hard. And since the room will vary it is not part of the 'system' processing. What is nice is that if the system processing is very good then for a simple system the room is about all you should have to address.

And with room modes-if you do turn down the ''hot spots'', you will ALSO make the quiet ones (the ones not loud enough) quieter.

As usual it is a tradeoff.

Both things mentioned were exactly my points. If I decide to compromise for the room differently than ''flat'', I don't have to guess how much of what I'm changing is non-linearity of my PA vs. the room - I know what baseline I'm starting from.

 

[Bennett Prescott]

Re: Second guessing the engineers

I see, the pole mount on the sub is a little over a foot further back. I suppose you need two settings, unless you always use the speaker pole mounted on the sub. Measuring equidistant between the sub and the top is great, people's ears aren't there but if they are far away enough it's close enough. If you tune it at ear level then you're aiming a lobe at crossover up into the air, whether it catches enough of your audience is one of those ''it depends'' things.

 

[Jay Barracato]

Re: Second guessing the engineers

This was an exercise in using the software as much as determining an actual field value for the speakers. I am not sure if or how the boxes are time aligned as a three way system, but it appears as if the design was based around using the pole mount.



I am trying things out. By all means keep poking at my measurements. Defending them will make me really think about what the software is showing me.



In practice I almost always use the pole, but I guess it would be reasonable rather than aligning the lips of the speakers when stacking them to move the top back so the ploe mount lines up, even if the pole is not there.

 

[Jeff Babcock]

Re: Second guessing the engineers

I recall that Evan once posted Smaart measurements of Mackie HD1531 over HD1801 and found the same - they were nicely aligned for that configuration.

 

[Bennett Prescott]

Re: Second guessing the engineers

I recall that Evan once posted Smaart measurements of Mackie HD1531 over HD1801 and found the same - they were nicely aligned for that configuration.

It is far more likely that, given spread crossover points, any top will need about 1.5-2ms of delay to line up properly. If they had put it into the top by default, you can bet people would be screaming bloody murder about 2ms when they found out.



I added 1.8 ms to some sidefill tops on Saturday. I had them flush with the front loaded subwoofers they were sitting on top of. All the processing inside the boxes is analog, so the only place latency can come from is the crossover points I was using... 65Hz and 102Hz.

 

[Jay Barracato]

Re: Second guessing the engineers

Bennett,



Does this latency show up as a difference between the electrical crossover and the acoutic crossover?

 

[Evan Kirkendall]

Re: Second guessing the engineers

I recall that Evan once posted Smaart measurements of Mackie HD1531 over HD1801 and found the same - they were nicely aligned for that configuration.

It is far more likely that, given spread crossover points, any top will need about 1.5-2ms of delay to line up properly. If they had put it into the top by default, you can bet people would be screaming bloody murder about 2ms when they found out.

Bennett,

The delay is actually on the High pass output of the subwoofer! Pretty genius, eh?

Jay,

Got any screenshots of your work?







Evan

 

[Bennett Prescott]

Re: Second guessing the engineers

Does this latency show up as a difference between the electrical crossover and the acoutic crossover?

Basically, yes. To make the acoustic crossover correct you spread the electrical crossover. The tops are seeing phase shift at 100+Hz, while the sub sees the phase shift of a 65Hz xo which is substantially longer, time wise, due to the wavelengths involved.



Evan,

Clever indeed! Keep the box low latency for SOS, monitor, and fill use. Do all the important processing in the only place most people will ever need it.

 

[Jay Barracato]

Re: Second guessing the engineers

I will have to go dig through what is still on my laptop computer. I did not save a lot of traces as I was really just fiddling.



Putting the top delay on the thru put of the sub is a neat idea, but on the QSC's you have a choice of using the crossover built into the sub or built into the top.



I was actually running separate signals and using the high pass in the top.

 

[Silas Pradetto]

Re: Second guessing the engineers

I recall that Evan once posted Smaart measurements of Mackie HD1531 over HD1801 and found the same - they were nicely aligned for that configuration.

It is far more likely that, given spread crossover points, any top will need about 1.5-2ms of delay to line up properly. If they had put it into the top by default, you can bet people would be screaming bloody murder about 2ms when they found out.

Bennett,

The delay is actually on the High pass output of the subwoofer! Pretty genius, eh?

Jay,

Got any screenshots of your work?







Evan

Also, IIRC, the HD top has FIR filtering which requires a bit of latency, while the sub does not. So they have to be compensating for that somewhere, because the latency built into the top is probably more than what's required for alignment, and if the sub is putting more delay on it...well, someone had to plan it out.