The making of: The Two Towers (a 25 driver Full Range line array)


Paid Member
2003-04-26 10:25 pm
Yes! :) And it seems to have some very interesting properties. I kept my neighbour up and banging my floor at 00.30 - I forgot the time and level - ouch.. not good. Only have one ch yet playing mono ... but the energy - wow - even on low volume ;) - I'll report some more soon in my thread...

  • Like
Reactions: 1 users


2013-01-11 2:58 pm
Well, I've considered parallel/series to be better than series/parallel, at one point in this thread.

What about doing both (see the red line)?
  • Like
Reactions: 1 user
It seems there are tiny differences when real world impedance curves are used...

4 drivers, each having it's own impedance curve (first 4 out of my library of Scan 10F curves) with 10 watt applied, voltage curve shown.

And now the version with the proposed connection in-between:


So it does change something. Limiting/averaging the voltage deviation among drivers?
With this little number of drivers, you could probably group/pair them to take some benefit of this.

I did not expect to see any differences to be honest.


Re-arranging the drivers makes the difference even smaller....
Clearly they become 2 averaged pairs of impedance traces this way.
Last edited:
  • Like
Reactions: 1 user
Here's the parallel - series use case (switched the drivers back to get the lowest voltage variation):


So that one extra wire in the series parallel makes them look pretty much identical.
In that case, the parallel - series or series - parallel usage will be identical?

The way I looked at it before (after noticing tiny differences between both schematics) was: the drivers in series make the impedance rise.
Putting the series drivers in parallel drops that to the nominal level again.

Beginning with parallel would make the impedance drop first, and back up to the nominal level after the series connection.

Somehow, that last use case seemed more favorable to me, but I wouldn't know if that holds true in any practical way.
I wasn't convinced enough to rewire my arrays to parallel - series. But with 4 drivers I'd measure all impedance traces and find out how to
group/pair them with X-sim :D.

For completeness, the 4 impedance curves used (to show there are no large deviations):
Last edited:
  • Like
Reactions: 2 users
It isn't in the same room ;). I have a Xeon workstation with the 3060 TI and it does all the rendering of the picture prior to sending it to the TV set.
I know it's not for all people but I like it that way. I'm using MadVR to get the picture quality I like. It is the high quality video renderer within JRiver.
It does take a powerful PC and ditto graphics card but to me it's worth it. As I do use the card's potential for 3D design programs anyway.
No contest sending the picture straight to the TV or trough MadVR. All processing on the TV is bypassed as the graphics card with the right settings
does a better job. I've been using this for as long as I have JRiver (since JRiver version 17) as I've always had a workstation with high end graphics card
available. It is a remarkable difference to me, first on a Plasma screen, now on OLED.

One can use it to upscale SD, HD etc to 4K and even mapping options for HDR... use what you want/need.

I know the trend is to stream everything these days but in video I'm after the same quality as in audio, for fun and entertainment.

For those with deep pockets, MadVR has the Envy.

It is a stand alone video processor that pretty much does the same thing that we HT nuts are doing with the madVR software.
It features a RTX 2080 Super graphics card in the Pro version, the more expensive Extreme edition will be upgraded to a RTX 3080.
The RTX 2080 super is pretty much on par with the newer RTX 3060 TI graphics card I just got. ;)
The biggest difference is the capture side of the stand alone machine, so it can function in between 2 HDMI devices.


2013-01-11 2:58 pm
View attachment 1071579
What about doing both (see the red line)?
One other advantage I could see is that each driver will better dampened by the lower driving impedance. @wesayso can we somehow simulate this?
Edit: I realise now that this extra dampening would only happen for differences between the drivers/acoustic impedance of each driver. But as we have always have non ideal, not perfectly equal drivers the end effect could be a positive one :)
Last edited:
There's a part of this thread talking about amps that would be nice to power line arrays.

Few would be able to find the one you are using now (and back from a smoking death!), but I thought I could share an option that would work very well.
Back in my hometown, I dug out my old Bryston amp. That thing is a beast, both in physical weight and audio reproduction weight.
It has massive grunt, but it is also very melodic. It can power a PA system, or a nice home setup. It punches way above other much more expensive amps, and the old ones can be had for not too much money these days.

Mine is about 40 years old.

And the cool thing... it is still supported by the company. They just offered a program to revamp the inside with new components for a very reasonable amount of money. Supporting 40-year-old products... that's my kind of company!

So, if anyone is serious about powering line arrays (or any other decent system), have a look at the classified section for a Bryston amp. If you can get your hands on one, you won't be disappointed.

Now, I need to figure out a way to ship this one from Canada to Asia... and arrive in one piece!

  • Like
Reactions: 1 user