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

Sure, catch you later.

Here is some food for thought regarding the previous link:
530712d1455300384-making-two-towers-25-driver-full-range-line-array-optimum.png

After playing several movies over the last few days I'm convinced a bit of reverb on the ambient channels may create even more atmosphere. I only briefly played with it a long time ago and it showed promise. I'll have to revisit that idea. But how? Probably using convolution to be able to optimise it. Just a tiny bit can go a long way. I don't want it to be too obvious. You shouldn't even notice it, except when it's turned off.

The above picture will be a good guideline...
The start point for me being this graph of what I have now:
postHaas.jpg
 
No measurements yet, just some trials. I used a convolution reverb to try what it could bring instead of using my strange (but working) L-R and L-R with a bit of C thrown in.
It did not convince me this is the way to go. Not saying it was a bad result. I did enjoy the resulting sound. But it had a shimmer that was prominent on all music played. At first I liked it very much, as it did feel like being in another place. But as the effect was the same (or similar enough) on all songs it gets old fast. It worked to dominate my room result, but not as well as I hoped it would.
In short: my old home brew setup did better to keep things more natural. There's an advantage for the reverb in staging though. Even though I was able to control the added effect I heard somewhat, using a convolution of a real room will also give you the quirks of that room. Not something detrimental when you are making mixes, as you can use what works to compliment the sound you recorded but not that great for faking a listening environment.
My next experiment will probably be with Valhalla Room, I downloaded the demo and as it's praised a lot as a versatile reverb, it will be affordable enough to get if I like it enough. If not, I can just revert back to what I cooked up and forget about this experiment.
As Valhalla room is a synthesized result, it will be different on each recording used. The demo cuts out intermediately so I can actually hear any difference compared to running without ambience.
The amount of choices within Valhalla Room is a bit overwhelming. But the same can be said about convolution based reverbs of coarse. But the convolution based reverb didn't convince me enough to go for a bigger convolution based package. Something like that would make more sense if I were actually making mixes.
Hope to be able to test the next reverb soon. In a comparison with headphones I could get it very close to the preferred sound I got using the convolution based solution. But I expect it to have less coloring overall. It's remarkable how much the reverb tail can have an influence the overall sound and tonality.
I'm only trying to hide my room here, by "inserting" the fake queues of a bigger one.

To bad I cannot try the Harman Logic 7 variant without hardware.
 
Last edited:
I'm sure you've done your research, but did you know Voxengo, one of your favorite company, has one too?
8-ch convolution (impulse reverb) plugin (VST) - Voxengo Pristine Space - Voxengo

Liquidsonics has Reverberate, which has a lot of control buttons as well!

After that, prices jump into the stratosphere with things like Altiverb, Vienna Suite, etc...

But, I was looking things up and found a library of wav impulses called Samplicity, that looks nice.

Here: Samplicity - Simply excellent reverb impulse response libraries
 
Yes, I did notice that.. I stopped at SIR 2. For the reasons I mentioned above.
I'm not as much looking to add a room, I'm trying to get rid of my own room :).

By using a synthetic reverb I'm probably only enhancing what is already there, not adding new stuff. That's what worked with my own recipe and I'm hoping to take it a step further with some gentle and short reverb.

The convolution based reverbs do have their place, I'm sure. Just not in enhancing what's there. They add too much new sound of their own.
That's why I mentioned Logic 7 from Harman, from what I've read it works very well. But it takes what I'm doing even further by analyzing phase and steering based on that. That might even be too much. But I do think it will sound pleasant.

I'm not looking for a huge sound effect, I'm trying to mask my tiny room. With adding as little as possible. Does that make sense?
 
Member
Joined 2009
Paid Member
By using a synthetic reverb I'm probably only enhancing what is already there, not adding new stuff. That's what worked with my own recipe and I'm hoping to take it a step further with some gentle and short reverb.

You'll have much less cpu task to perform so it's going to be 'lighter' on processing time and you'll probably experience less latency for treatments.

If you are on the short side of reverbs, try disable the 'cue' or 'tail' of reverb and play only with ER level/density/spread/time (if allowed by plug in) you should have some very convincing results this way without the 'blurring' and sometime 'washback' effect tail of synthesized reverbs brings on.

A nice one if you can use vst plug ins is the Oxford reverb clone (you originaly find in Sony all digital 'Oxford' big desks). Very nice and natural on percussive material (=good transient response= good quality from my experience).
 
I've only briefly played with the non impulse reverb algorithms. I decided to upgrade to my JRiver 21 release because all this time I stayed at version 19. I had run version 20 briefly but somehow it never sounded as clear or natural as version 19 to me.
After making sure all the buffer settings were right I re-did my entire processing.
This time recording raw impulses of left and right channels, exporting a minimum phase version of those and re-importing them to REW. Pré EQ is based on those minimum phase impulses and a 1/6 octave window in REW.
After that I re-recorded the left and right pulses and let DRC-FIR have it's way with them.
So far so good. I need a couple of days to see it everything works as planned, but judging from the chills down my spine and goose bumps while listening I'd say I'm on my way again. Hours fly by though. Completely lost track of time.

:)
 
After playing several movies over the last few days I'm convinced a bit of reverb on the ambient channels may create even more atmosphere. I only briefly played with it a long time ago and it showed promise. I'll have to revisit that idea. But how? Probably using convolution to be able to optimise it. Just a tiny bit can go a long way. I don't want it to be too obvious. You shouldn't even notice it, except when it's turned off.

The above picture will be a good guideline...
The start point for me being this graph of what I have now:
postHaas.jpg

As a suggestion, I wouldn't rely too much on a full range ETC as an ambient gauge. Its too frequency biased. Instead, consider 1/1 octave banded ETC.

For a proper ambient effect, the bandwidth must be considered, both in magnitude and temporally.

I also have another thread where I am talking about this sort of thing: https://www.gearslutz.com/board/studio-building-acoustics/963767-artificial-haas-kicker-tail.html
 

Attachments

  • Banded ETC.jpg
    Banded ETC.jpg
    486.6 KB · Views: 441
Last edited:
Thanks Jim, I know the usual IR shows only about ~4 to 8 KHz+ info. I'm pretty used to use the filters on that tab to find out what's really going on :). I want to test APL_TDA with the ambient channels engaged one of these days. The ambience signal will show up as bumps above the noise floor.

A while ago I was having a rough time with one particular harsh guitar note in Opeth's song Windowpane. I must have checked my processing at least 4 times before I found out it was the delayed Hass Kicker that was adding the harshness to that note. At the time I didn't expected it to be that strong in coloring the sound.
That's why the convolution reverb didn't do what I wanted. I want to hide my room more than adding a different one. I can't hide everything but our mind does help if the balance and level works out.
Sadly I can't delay the kicker much further than ~21 ms, as my room still has a strong reflection pulse there. As can be seen in the TDA plot:
TDA_3D.jpg

Yes, I do know (and see) there are reflections before that too :). It will remain a living room. I'm actually surprised how well "hiding" the room works...
 
Last edited:
Basically a termination of the reflection free zone(*). In the plot I posted you can see the main pulse (going to 100%) and a later set of strong pulses at ~20 ms and a gradual decay after that.
Those signals come from the diffused ambient speakers from back left and back right. My back wall is at ~6.3 ms after the main pulse. This can be seen in the APL_TDA plot. It has a damping poster and a couple of pillows to damp as much as I can to get the signal about 20 dB beneath the main pulse.
By introducing a "fake" broadband reflection at ~20 to 24 ms your mind will "believe" it is listening in a bigger room (most effective with eyes closed).
This creates a bit of ambience of a much bigger space than I actually have.

The signal used is L-R and R-L, band limited, delayed and attenuated. I mix in a bit of center as that helps keep the tonal balance the same throughout the stage.

Jim has a Haas kicker that's done with clever reflection/diffusion in the room. And lately he's experimenting with a reverberant field created by separate speakers in adjoining rooms! His TDA plot shows his original Haas Kicker pretty good:
524831d1452732108-group-delay-questions-analysis-apl-tda-35ms-3d.png

A clear example of the Haas Kicker at 24 ms by clever use of the sound in the room. His later results are even better, more broadband.

(*) not that I don't have reflections, but I did try to get the level down below at least 20 dB in the first 20 ms.
My room without Haas Kicker:
preHaas.jpg

(note the reflections in the first ~20 ms being 20 dB below the initial peak)

And the same room with Haas Kicker:
postHaas.jpg

(a big "reflection" introduced at ~20 ms, gradual decay after that)

But note Jim's post, these filtered IR's show only part of what's going on. You want a wide band reflection, though I prefer it rolled off above ~3 to 5 KHz.
Mine is about 200 Hz 12 dB/oct to 3.5 KHz, 12 dB/oct FIR corrected for bandwidth, not timing. The ambient speakers are aimed to reflect of nearby surfaces. You want the Haas Kicker to be a diffused uncorrelated signal.

Short answer: a mind trick within the Haas limit... :)
 

ra7

Member
Joined 2009
Paid Member
Hmmm... interesting. So you introduce this through the surround channels or through the main speakers?

This is what surround channels do in a home theatre setting, right? Delayed, band limited sounds from different arrival locations can transport you to a different space. Gun shots and bullet ricochets and so on.
 
Yes, that's the ambient speakers doing their job (they only double as surround speakers here ;) *). No alteration in the main channel signal. The kicker has to come from the back/side at about ~120 degree if zero is straight forward.

At no point do you really "hear" these speakers, unlike movie effects. But they do add atmosphere and a sense of space. They actually enhance the stage in front of you.
You only notice what they do if they are turned off! But you need to get rid of all early reflections, having them as low as you can manage. 20 dB down or more was my goal in a living room.

*= in a HT situation you could aim the back speakers to the listening position. The effects have the needed signal alteration to make it believable. In ambient channels you'd want to have a diffused sound, not direct sound from the speaker. I aim the ambient channels to the outside where they reflect of walls and accessories to diffuse them. A bit of EQ/FIR is needed to keep the FR in check. As said, it's also not a pure stereo signal they get.
Lately I'm experimenting with adding some short reverb tail. To fill in what a small room does not (re)produce.

My first queue for this was in Car Audio:
werewolf said:
"rear fill" encompasses many possibilities :

- simple, attenuated L and R
- delayed (beyond Precedence, or Haas, Effect), attenuated, bandlimited L and R
- delayed, attenuated, bandlimited difference signal, L-R

there's a world of difference between these options. Some will confuse the front stage, others will enhance the front stage.
Source thread: http://www.diymobileaudio.com/forum/system-design-help-me-choose-equipment-my-car/9806-rear-fill-do-you-use.html
This "werewolf" gentleman, later posting as "Lycan" has thought me a lot about the concept. It's actually a trick that was used in Studio Control Rooms until they noticed it actually enhances the signal more than make/leave it pure to judge. Not that bad at all in a listening room but you wouldn't want a Control room make the mix sound better than it is :). In a Car or small room it can make the space seem larger than it is.
werewolf said:
Big difference between a concert hall and car: size of the acoustic space!

And it's NOT just a relative difference ... there's a real, tangible threshold involved, as defined by the Haas Effect. In a car, there's a real good chance that the natural reflections will not exceed the precedence effect (about 20msec, if memory serves) ... meaning that natural reflections will only tend to smear and confuse the front stage. In a concert hall, there's a real good chance that the natural reflections will exceed the precedence threshold.

So if you delay the rear fill by ~20msec or slightly more, you can create an apparent acoustic space that is much larger ... without confusing or ruining the front stage. This is a good thing Yes, we face the orchestra in a concert hall ... a classic argument against dumb rear fill ... but we also sense that the concert hall is bigger than about 25 cubic feet ... a very good argument in favor of smart rear fill

How to best accomplish it, is perhaps a discussion for another thread. You need a processor with time alignment, and the ability to bandlimit the rears. The difference signal can be formed either actively before the amp (trivial with balanced signals, not even hard with transformers and single-ended signals), or speaker-level after the amp (the classic Hafler matrix was the old, dumb example ... with no time delay, and even only crude attenuation and bandlimiting). But here's an interesting link for further reading:

http://www.mlssa.com/surround/surround3.htm

Of course, you may want to consider the options mentioned above ... Dolby Pro Logic II (invented by Fosgate with an automotive acoustic space in mind), DTS Neo:6, and the new Harman Logic7 processor for the Car (which, by the way, is starting to get some very encouraging reviews). All of these options create a multi-channel signal from a 2-channel source, and make very intelligent use of speakers behind your head

I guess my main point is this : all "rear-fill" is not created equal It's important to distinguish what type, before we summarily embrace or dismiss the entire category.
 
Last edited:
A while ago I was having a rough time with one particular harsh guitar note in Opeth's song Windowpane. I must have checked my processing at least 4 times before I found out it was the delayed Hass Kicker that was adding the harshness to that note. At the time I didn't expected it to be that strong in coloring the sound. .

I have found that one must be careful of peaks in a specular return. Unlike a diffuse return where the energy at any given frequency can be spread out in time, the specular type arrives all at once. I usually go to 1/24th smoothing and take a look at a time restricted IR as to see the Kicker only. Then adjust my panels around to avoid peaks that may be hidden at 1/6th smoothing.

Whether it be via Kicker or otherwise, specular energy can cause harshness if not smooth in its response, and at the right magnitude. I think you do have more latitude with a 20-24ms specular return (kicker) than say a 5-10ms early return in terms of how quickly it can poison the fruit (cause unwanted side effects such as coloration/harshness).
 
Of coarse JRiver makes it easy to implement this at home. You can add/subtract, delay attenuate and EQ/FIR. I'll see if I can dig up the thread that explains the basics.

Here's a nice compilation a member made:
Rear fill Pro's and possiblities - Car Audio | DiyMobileAudio.com | Car Stereo Forum

There's lots more actually, but as werewolf isn't a member name anymore it's harder to find, though google still does a reasonable job.
 
I have found that one must be careful of peaks in a specular return. Unlike a diffuse return where the energy at any given frequency can be spread out in time, the specular type arrives all at once. I usually go to 1/24th smoothing and take a look at a time restricted IR as to see the Kicker only. Then adjust my panels around to avoid peaks that may be hidden at 1/6th smoothing.

Whether it be via Kicker or otherwise, specular energy can cause harshness if not smooth in its response, and at the right magnitude. I think you do have more latitude with a 20-24ms specular return (kicker) than say a 5-10ms early return in terms of how quickly it can poison the fruit (cause unwanted side effects such as coloration/harshness).

True! Which is why I was surprised a signal that late (at least 20 ms) could still have such a strong effect. I'd swear it was coming from the main speakers.
My ambient speakers are behind pillows without a direct path to the listening position. Ideally I'd like them somewhere else but I like to keep my relationship healthy too :rolleyes:.

After playing with convolution reverb, even mixing it at 100% to the ambient channels I had to re-think how much the coloration is or can be from the fake room (and real room), even if it's down by a significant amount.
The convolution reverb option did not work out that well, the (L-R) based signal worked better by itself. It didn't work mainly because every song had the same added effect with a convolved real room tail. I already had that in real life! (lol)
The less I notice the ambient addition, (and forget about my room) the better. Though I love the feel it gives to the music. The added ambience is noticable.
 

ra7

Member
Joined 2009
Paid Member
I think you will find this interesting, though you may have already seen it:
WATSON-Stereo_Expansion_Loudspeakers

Why is it important for it to be coming from the rear? Won't decorrelated sounds more than 60 degrees away from the nearest main speaker suffice? I think there are some guidelines for this in Toole's book.

One time, I mistakenly delayed the lower frequencies, say below 500 Hz, on the main speakers. Even that added a sense of depth and ambiance. It was strange. I didn't pursue it further.
 
Yes, I know that experiment. He did more experiments with extra speakers which I've linked here before:
Surround stereo system
He uses a processor to generate the ambient channels. I just mix and match my own signal.

The Watson seemed to start as a sort of cross talk elimination but got another function in the end. I doubt it's still in use.

I tried pointing my ambient channels up, like Linkwitz does in the link I posted, I actually build them with that kind of placement in mind. The ceiling reflection got in the way though. It was way too strong. I settled on aiming them outwards. Creating a diffuse, distant sound.

The key why this Haas kicker works is how we perceive room sounds in daily life (and the way we are trained to ignore most of it).

I've read Toole's work and interpretation and I don't think that's the same as what we talk about here. His rooms are infinite better than my room too!
I've read somewhere he experimented with Haas Kickers in an ambient space and dismissed the idea.
I have no idea what he used or tried though, but eventually Harman made Logic 7. Wasn't David Griesinger involved there at the basics which led to Logic 7? Did you see the link Jim put up? http://www.davidgriesinger.com/paris_talk.pdf (page 8 resembles what I have, it actually works in more than one seat but the sweet spot is best, even my 9 year old kid knows that very well! No way I get to sit there with movies anymore :). Luckily I kept an eye on the sound in the off axis seats)
Logic 7® multichannel surround-sound technology has set a new standard for acoustic precision and authenticity, contributing to a completely new surround-sound experience in automobiles. Originally developed over the course of many years for professional recording studios by a team of Lexicon engineers under the leadership of acoustics guru David Griesinger, Logic 7 processing has been the first choice of discerning music lovers for their homes since the introduction of the Harman Kardon AVR 7000 audio/video receiver in 2000. Logic 7 processing became available for cars in 2001. This unique technology has become indispensable, particularly in the automotive industry, because it delivers authentic, three-dimensional listening pleasure to all seats in every car.
I guess I remembered it right. From: Harman Kardon Innovation - Logic 7 Processing

Logic 7 is a more advanced steering mechanism than Pro Logic and a (pretty strong) competitor to Pro Logic IIx I guess. But even those are not using exactly the same ideas as the original Haas kicker. I wouldn't know how they fare though. I know Logic 7 has a pretty good reputation, even for 2 channel music while others, like Neo:6 and Pro Logic IIx seem to disappoint. But the Haas Kicker isn't meant to be surround sound. Just enhancing the 2 channel stage and imaging is the primary goal. Logic 7 actually steers sounds to the back channel.

For me, after trying this ambient stuff it has become a permanent addition to my setup. I have used it with movies too, with a down-mix to 2 channels. Worked very well but after taking an hour of my time recently, I could convert my setup to use 4.0, out of a 5.1 signal as I still don't have a center or sub(s).

Even though it's easy enough to try, you've got to work with the room. The level of early reflections needs to be down sufficiently (20 dB+). The signal of the ambient channels needs to be diffuse but smooth in frequency response. All in all it takes work to get the most out of it. I'm sold, I'm only playing with reverb due to what I remember reading from David Griesinger.

It's sole goal: make me forget my room and enjoy music! Works very well for that, as you can read in the review I posted of my system as written by Jan Fekkes (link in the first post of this thread). A more believable experience is what we are after right?
 
Last edited:
Yes, I know that experiment. He did more experiments with extra speakers which I've linked here before:
Surround stereo system
He uses a processor to generate the ambient channels. I just mix and match my own signal.

The Watson seemed to start as a sort of cross talk elimination but got another function in the end. I doubt it's still in use.

I tried pointing my ambient channels up, like Linkwitz does in the link I posted, I actually build them with that kind of placement in mind. The ceiling reflection got in the way though. It was way too strong. I settled on aiming them outwards. Creating a diffuse, distant sound.

The key why this Haas kicker works is how we perceive room sounds in daily life (and the way we are trained to ignore most of it).

I've read Toole's work and interpretation and I don't think that's the same as what we talk about here. His rooms are infinite better than my room too!
I've read somewhere he experimented with Haas Kickers in an ambient space and dismissed the idea.
I have no idea what he used or tried though, but eventually Harman made Logic 7. Wasn't David Griesinger involved there at the basics which led to Logic 7? Did you see the link Jim put up? http://www.davidgriesinger.com/paris_talk.pdf (page 8 resembles what I have, it actually works in more than one seat but the sweet spot is best, even my 9 year old kid knows that very well! No way I get to sit there with movies anymore :). Luckily I kept an eye on the sound in the off axis seats)

I guess I remembered it right. From: Harman Kardon Innovation - Logic 7 Processing

Logic 7 is a more advanced steering mechanism than Pro Logic and a (pretty strong) competitor to Pro Logic IIx I guess. But even those are not using exactly the same ideas as the original Haas kicker. I wouldn't know how they fare though. I know Logic 7 has a pretty good reputation, even for 2 channel music while others, like Neo:6 and Pro Logic IIx seem to disappoint.

For me, after trying this ambient stuff it has become a permanent addition to my setup. I have used it with movies too, with a down-mix to 2 channels. Worked very well but after taking an hour of my time recently, I could convert my setup to use 4.0, out of a 5.1 signal as I still don't have a center or sub(s).

Even though it's easy enough to try, you've got to work with the room. The level of early reflections needs to be down sufficiently (20 dB+). The signal of the ambient channels needs to be diffuse but smooth in frequency response. All in all it takes work to get the most out of it. I'm sold, I'm only playing with reverb due to what I remember reading from David Griesinger.

It's sole goal: make me forget my room and enjoy music! Works very well for that, as you can read in the review I posted from Jan Fekkes (link in the first post of this thread). A more believable experience is what we are after right?

Nice post!

I want to make something explicit that you implied. Whether were talking about ambient surround or Haas Kickers, at the end of the day, if you can hear (directly) what you have added, then you have done it wrong (as it applies to 2 channel stereo reproduction).

Another way to say this is that if your kickers/ambient content doesn't enhance what the 2 channel output is doing on its own, then it is done wrong. By enhance, I dont mean added to. While that is what we are doing from a technical point of view, the perceptual impression should be a deepening of whats already there, not a sense of something being added to it.
 
Nice post!

I want to make something explicit that you implied. Whether were talking about ambient surround or Haas Kickers, at the end of the day, if you can hear (directly) what you have added, then you have done it wrong (as it applies to 2 channel stereo reproduction).

Another way to say this is that if your kickers/ambient content doesn't enhance what the 2 channel output is doing on its own, then it is done wrong. By enhance, I dont mean added to. While that is what we are doing from a technical point of view, the perceptual impression should be a deepening of whats already there, not a sense of something being added to it.

Exactly! I wholeheartedly agree to the above. I'd definitely say it's a worthy addition to enhance the 2 channel experience. Without making it into a roller coaster ride. More smiles per minute is the goal without it being obvious. I often go to the side of my couch to check if they are indeed on. Though I should notice if they aren't. Just a habit I guess, I want to know if they are indeed playing... I can't help myself...

I have played music using the back channels only, then turn on the mains. For about 2 seconds I can kid myself I still hear the back speakers. After that? Forget about it...