Which limiters are good enough?

Re: Which limiters are good enough?

As far as the whole "limiting an out of control engineer," I wouldn't do that either. A truly ignorant engineer will just push further into those limiters you're trying to set to keep him quiet, and you'll end up sending a higher power average to your speakers, and could do more damage than just letting him suck out loud. If anything, turn him down on the DSP and let him run out of headroom on the console.



Evan

+1000 Evan. Everyone in JV read this. It should be "Rule Number One" when dealing with a less than stellar engineer.
 
Re: Which limiters are good enough?

As far as the whole "limiting an out of control engineer," I wouldn't do that either. A truly ignorant engineer will just push further into those limiters you're trying to set to keep him quiet, and you'll end up sending a higher power average to your speakers, and could do more damage than just letting him suck out loud. If anything, turn him down on the DSP and let him run out of headroom on the console.



Evan

Back in the 80's I had a regular Sat night gig (not installed-I brought it in each day). The venue (Wilmars Park in Brandywine Md) had a large outdoor area and a indoor room (maybe 1000 heads) as well. If the weather was good-we would go outside. If bad-we went inside.

I used the same rig either way. Indoors was not a problem. But outside (doing Metal) it was not quite enough. Sounded good-but didn't have enough level for what the typical engineer wanted. But I was not being paid any more-so they got the same rig-deal with it.

A common problem with the "engineers (I use that term loosely) was that they would just keep pushing the system and bury the mix in the main limiter. It would huff and puff-but didn't blow up. :)

When the next band would come on-I would simply drop the level a good bit and "stay under the limiter". Then the first band would come up and start yelling at me because I "screwed them" and didn't let them be as loud. I couldn't get it through their thick ego heads that what I did was DROP the level for the next band that I was mixing.

When the mix wasn't being buried in the mix-there was more dynamic range and there was more impact to the sound-even if it was actually quieter.
 
Re: Which limiters are good enough?

BTW, I'd second Evan's recommendation of the DCX2496 as a budget loudspeaker processor. It includes a combined RMS and peak limiter.

Nick
There's nothing in the manual about dual limiting. On each output, it appears to have a typical peak limiter with a threshold and a release time.

I've used these in a number of churches. They're a great little processor for those on a budget but I wouldn't trust the limiter to enable running hard, unless each limiter is set really conservatively. I measured a few once and the outputs were matched to within +-0.1 dB. But I have heard of units that have died. Something to do with the main board being close enough to the case to sometimes short.
 
Re: Which limiters are good enough?

Hi Michael,

There's nothing in the manual about dual limiting. On each output, it appears to have a typical peak limiter with a threshold and a release time.

FWIW, my comment was based on a review I did quite a few years ago. I don't remember many of the details now, but I did post to this thread in another forum on the subject. No guarantees.

Nick
 
Re: Which limiters are good enough?

Would that protect my speakers as good as or better than the Dbx Driverack limiters?

I think it is more inportant how the limiter is setup than the "quality" of the limiter itself. As far as protection just about anything that can do 20:1 or better is going to perform just about like anything else with maybe the exception of a slight adjustment to the setup.

If you are asking about the sound quality of a limiter I would say that they all suck when they kick in while doing speaker protection duty.
 
Re: Which limiters are good enough?

If you are asking about the sound quality of a limiter I would say that they all suck when they kick in while doing speaker protection duty.
And when you are pushing the system hard enough to be close to having to use speaker protection, then the sound quality is the last thing you should be interested in. If it sounds like total crap-then that can be a good thing. At least you will know where you are-and it is time to turn it down-or it may get turned WAY down-for you-by killing the speakers.

I remember once a band telling me (around 11pm on a Sat night) to TURN IT UP-and if you blow the speakers-we will pay for them! Yeah right. I bet-----------------. And if I blow the speakers-then how are you going to play the rest of the show? Like I can get replacements around midnight.

I ran it as hard as I dared-and didn't blow anything. Of course this was 30yrs ago-and limiters were not as common as now (at least on the club scene). I did not have any in the system-except my hand on the master level.
 
Re: Which limiters are good enough?

Hi Michael,

FWIW, my comment was based on a review I did quite a few years ago. I don't remember many of the details now, but I did post to this thread in another forum on the subject. No guarantees.

Nick
Hi Nick,

I took a look at the images on the "other forum." I must confess I don't understand quite how to interpret the diagrams. It's not what I would expect to see from any limiter. But I often miss the obvious. :)

I've attached a diagram showing an simple example of the trends I expect from a dual RMS/peak limiter.
(a) Shows a sine wave going from quiet, to loud (above limiting thresholds), to quiet
(b) Shows (a) after going through the RMS part of the limiter. The limiter begins to engage at time t0 but takes some time to settle. When the input signal drops at time t2, the limiter takes some time to release.
(c) shows the waveform from (b) after passing through the peak limiter. This example has no look-ahead. The peak limiter takes a short time to engage and settle (quickly after t0) and then by time t1, the RMS signal falls below the peak limiting and the RMS behaviour continues.

Of course I've made some assumptions here including the use of RMS limiting first, then peak limiting to mop up anything the RMS limiter doesn't grab fast enough. (Some limiters run the peak and RMS stages in parallel and the combined gain is a function of the gains from both computations.)

Best,
Michael
 

Attachments

  • limiter_example.gif
    limiter_example.gif
    12.1 KB · Views: 0
Last edited:
Which limiters can do more than cook?

And when you are pushing the system hard enough to be close to having to use speaker protection, then the sound quality is the last thing you should be interested in. If it sounds like total crap-then that can be a good thing. At least you will know where you are-and it is time to turn it down-or it may get turned WAY down-for you-by killing the speakers.

Ivan! Your stories fill me with awe. I don't know what kept you at this, but I'd have turned tail and run after one of your gigs. :)

On limiters...

What if you really could find a beautiful woman that can cook?

An example.

The following traces were made with short term sine bursts of (4) wavelengths each at a resolution of (24) steps per octave. Continuous sine sweeps or broadband anything (pink noise, etc.) will make a limiter respond far more aggressively than it will with music. I also use music as a stimulus to prove limiter behavior and sine bursts are so close that the difference is immaterial.

The blue traces are how the low pass of a normal limiter responds that is setup for sub use. Peak threshold was 0dBu (equivalent to -3dBu RMS due to the crest factor of sine waves) and the stimulus input to the limiter varied from -10dBu (black trace only) to +15dBu. You'll notice a bit of a soft knee as well.

The blue traces show the (best case) ugliness that happens as the output is squashed by the limiter - notice how the sub is now throwing mud into the mix at the highest levels. This is what everybody tolerates on BOTH ends of each passband. Except for me, of course, as I prefer the pretty ladies. :)

The black traces are possible through side-chain control of the limiter's behavior. There are many more happy ramifications to this.

Ivan.png
 
Re: Which limiters are good enough?

Hi Michael,

I took a look at the images on the "other forum." I must confess I don't understand quite how to interpret the diagrams. It's not what I would expect to see from any limiter. But I often miss the obvious. :)

It's not really what I expected to see either!

I've attached a diagram showing an simple example of the trends I expect from a dual RMS/peak limiter.
(a) Shows a sine wave going from quiet, to loud (above limiting thresholds), to quiet
(b) Shows (a) after going through the RMS part of the limiter. The limiter begins to engage at time t0 but takes some time to settle. When the input signal drops at time t2, the limiter takes some time to release.
(c) shows the waveform from (b) after passing through the peak limiter. This example has no look-ahead. The peak limiter takes a short time to engage and settle (quickly after t0) and then by time t1, the RMS signal falls below the peak limiting and the RMS behaviour continues.

Of course I've made some assumptions here including the use of RMS limiting first, then peak limiting to mop up anything the RMS limiter doesn't grab fast enough. (Some limiters run the peak and RMS stages in parallel and the combined gain is a function of the gains from both computations.)

It's tricky showing both RMS and peak thresholds on a single picture. The thresholds are different, and the signal can at any time be "over" one or both limits. I'm sure there are any number of ways to do it but, if either detector says "over", the gain would typically be divided (by a value derived from the attack time and the period of the update cycle). Plus, if the peak detector is "over", the peak might be clipped off. The RMS detector must operate over a rolling window of some length (perhaps derived from the attack time).

For example, here's a 1kHz sine wave (in grey) hitting a limiter with an attack time of about 20ms where the peak limit is 6dB above the RMS limit:

limit.png

Nick
 
Re: Which limiters are good enough?

When the mix wasn't being buried in the mix-there was more dynamic range and there was more impact to the sound-even if it was actually quieter.

Such true wisdom. I recently had a show specd for 1500 so I brought enough for around 2500. 5500 plus showed up. Everyone had a good time and no one complained that it was not loud enough. We ran the PA well within its limits. Clarity trumps volume nearly every time.
 
Last edited:
Re: Which limiters are good enough?

Very good analogy. The best solution to prevent blowing up gear is to make sure you have enough that you're not driving it to the edge. If I EVER see a clip light during a show, that means I didn't bring enough.

I think this is mostly true for experienced operators and finely tuned rigs. That said, I've seen rigs that are so poorly set up, they hit clip lights because one passband is just out of control and/or there is a wild EQ setting somewhere.

I seem to remember a certain someone who blew up only the mids of their rig, mid show. I think this is a good example of where the rig probably had enough gas, but something just wasn't right, and therefore gave up the smoke. Or maybe the mids were just "weaksauce."

I run DR260's on most of my rigs, and have the peak limiter in place just in case things go horribly wrong. But they aren't there to kick in during a show. i.e., bad cable, someone drops a mic, etc. My power amps are matched so that out of clip, the RMS power run for extended periods should not be a problem. You'd have to try REALLY hard to push it to the limit long enough to blow it.

I've always wondered about the DR260s limiters. They have that PeakStop+ and overshoot control. In someways it's sort of like an RMS limiter, though not exactly...

If you set the threshold for X and the overshoot for +6, then the limiter is supposed to start engaging at X and never let the signal go over X+6. Theoretically, if you set X for the Vrms level - 3dB you want your speakers to run, you shouldn't cause much overheating and still have 6 dB of dynamic range (not much, but OK for heavy music program - most recorded music I see today has less than <6 dB of crest factor).

Maybe it would limit the system too much. That said, with the standard 2x RMS power amp sizing, I don't think it would. I think most power amps can squeeze out 2 or 3 dB over their RMS rating, putting you at 5-6 dB over RMS for total "swing."

Power amp sizing (for maximum performance) is anything but straightforward IMO. I think it largely has to do with the type of music - and even then it's not simple. People never seem to wrap their brains around the fact that with a lot of live music you are looking at a 10db crest factor (easily) so to extract every last bit of juice out of a speaker you really need amps that can do the peak rating of the speaker.

I'm not even sure what the peak ratings of the speakers mean exactly - and I think in some cases with a very dynamic signal having an RMS below the rating of the speaker, but with crests above the peak rating - you'd still probably be OK without a limiter assuming your amps are sized to do the work.

That said, by in large, in my experience, if you have amps that are 2x rms, you will likely never blow a driver sans limiters if you avoid, or just occasionally tickle, a clip light.
 
Re: Which limiters are good enough?

It's tricky showing both RMS and peak thresholds on a single picture. The thresholds are different, and the signal can at any time be "over" one or both limits. I'm sure there are any number of ways to do it but, if either detector says "over", the gain would typically be divided (by a value derived from the attack time and the period of the update cycle). Plus, if the peak detector is "over", the peak might be clipped off. The RMS detector must operate over a rolling window of some length (perhaps derived from the attack time).

For example, here's a 1kHz sine wave (in grey) hitting a limiter with an attack time of about 20ms where the peak limit is 6dB above the RMS limit:

limit.png

Nick
I agree. What you've shown here is a zoomed in version of my (c) diagram around time t0 and t1, albeit with 0ms attack on the peak limiter. (I shouldn't have shown any attack time; easier to show a hard limit.)
 
Last edited: