Can someone explain this feedback problem?

Our mixer (soundcraft digital) has feedback suppression built-in but I've been afraid to use it in an actual gig.


You should try it. The technology has been around since at least the early 90's when I used a rack full of Sabine SL-820 FBX Solo, which worked pretty well.

It's been a long time, but I believe the old Sabines gave you an option to choose how many fixed and how many dynamic filters were used. The process was to increase gain so until the system was on the verge of feedback. During this stage the fixed filters keyed in on problem frequencies and were locked down. I don't remember this taking very long. After this you return the gain to normal. Later during the show the dynamic filters were free in case a new problem frequency developed.

I have also used Bose L1s which have built in Feedback suppression that works surprisingly well.

Feedback is hazardous to your hearing. I recommend wearing hearing protection when your ringing out a system.
 
  • Like
Reactions: ktedrow
You should try it. The technology has been around since at least the early 90's when I used a rack full of Sabine SL-820 FBX Solo, which worked pretty well.

It's been a long time, but I believe the old Sabines gave you an option to choose how many fixed and how many dynamic filters were used. The process was to increase gain so until the system was on the verge of feedback. During this stage the fixed filters keyed in on problem frequencies and were locked down. I don't remember this taking very long. After this you return the gain to normal. Later during the show the dynamic filters were free in case a new problem frequency developed.

Yep, I've been thinking I need to try it. The soundcraft implementation has fixed and dynamic filters also, pretty much the same. I just need to buckle down and do it. Maybe practice at home to make sure I know what I'm doing.
 
Ensuring the subs are not overly hyped may be something to consider. If the subs are crossed over at 100hz and feeding back at 185hz, they are way too loud.

I did have the sub volume knob turned up much higher than usual (3:00). I felt like that was needed to get enough lows at this venue, I assumed because it was outdoors.

I also assumed that the feedback was through the mains, but I never tested the theory. So maybe could've been the sub.
 
  • Like
Reactions: Wasnex
It would be well worth your time and money to tweak the pa with some sort of pink noise generator and software to set your basic eq settings. This should just about eliminate the feedback and ringing. Did it in one of my bands and it made a crappy pa, not too bad...
 
I did have the sub volume knob turned up much higher than usual (3:00). I felt like that was needed to get enough lows at this venue, I assumed because it was outdoors.

I also assumed that the feedback was through the mains, but I never tested the theory. So maybe could've been the sub.


Not that everyone agrees with this approach, but I typically run my subs off an aux. I set the aux level so the sub and mains have nominally flat response. If I need more or less lows I use the channel EQ instead of messing with the sub level. But I usually try to setup the mains and subs to have some semblance of phase integration at the crossover frequency. Chaning the level of the subs, changes the effective crossover frequency and introduces phase response errors. I only assign channels to the sub tha benefit from extended low frequency response. Candidates include bass, bass drum, left channel of the keys, and possibly floor toms.

Another thing I tend to do is run a high pass filter (HPF) on all channels that do not require extend low frequency response. If the board has variable HPFs I will run them on all channels including the bass and bass drum. If the subwoofer only works down to 45hz, then an HPF slightly below 45hz can tighten up the low end and allow the system to play louder.
 
  • Like
Reactions: LilyT and AGCurry
It would be well worth your time and money to tweak the pa with some sort of pink noise generator and software to set your basic eq settings. This should just about eliminate the feedback and ringing. Did it in one of my bands and it made a crappy pa, not too bad...

IMHO, tuning the system for flat response does very little to eliminate feedback. Some programs do give you graphic display so you can quickly identify problem frequencies and notch them out. But when you apply a deep notch to control feedback, you are intentionally modifying system response so it is not flat.
 
  • Like
Reactions: Mugre and ktedrow
I only assign channels to the sub tha benefit from extended low frequency response. Candidates include bass, bass drum, left channel of the keys, and possibly floor toms.

This is a good idea. I have the problem that my sub and mains aren't matched, so I don't know how to ensure a smooth crossover of content that goes to both (bass for example, maybe the only example). Actually, I have one idea, use the mixer to control the cutoff/slope of sub signal and main signal to ensure a smooth crossover. Both the sub and mains have the capability to disable their LP and HP filters, so this should work in theory.
 
Last edited:
I like this idea but our sub (cheapo alto model) isn't matched to the mains, and I don't know the sub cutoff freq or slope (not published). I know the cutoff freq for the mains (switchable between OFF, 100, or 125Hz), but I don't know the slope. For this reason I've been using the sub's built-in crossover to create the main signal (main HPF switched off) thinking this will ensure a smooth crossover. Then use ears to get right mix of mains and sub. The problem is that this puts all signal content into the sub, including vocal mics.

I'm open to ideas for how to resolve this. Our mixer does have spectral analyzer built in, so maybe with a noise generator I can figure out what the sub and main cutoff slopes looks like. But I feel like this is hard to do accurately (dependency on mic, mic position, room etc), so I haven't really pursued it.

Maybe I should create separate sub/main signals, as you suggest, and adjust main eq through the board until.... what? Until a recording into both sounds good? Until my bass sounds good (hey, this is all that matters, right)? But with two variables (eq, sub vs main volume) I'm not sure how to do this. I'm at a loss here.

I don't have sophisticated test equipment, so I do the best I can. I typically run systems that have the ability to apply delay to the outputs. So I set up a tone generator at the crossover frequency and then A, B the mains and subs adjusting for equal loudness. Next if the mains and subs are collocated, I invert the polarity of the mains and start delaying them for the maximum null. When you find the amount of delay that provides the most cancellation, flip the polarity back and your done. You can either use your ears or place a microphone midway between the subwoofer and woofer of your mid/high pack. Solo up the mic on your mixer and monitor the level on the meter. I think using a mic and the meter on your board is slightly more accurate.

Sometimes you wind up with more delay than you expect given the wavelength at the crossover. Ideally you're looking for delay less than 1/2 wavelength at the crossover frequency (see attached frequency wavelength chart). Basically your dealing with the difference in physical offset between the drivers and the different phase response of the drivers. The phase differences between the drivers is unknown, but I believe the mains will lead the subwoofers unless you get into active port frequencies. If you have to delay more than expected, it suggest the subs need to be delayed to the mains instead of delaying the mains to the subs. I believe this can occur if you run the mains low enough for their port to start becoming active. When the port starts resonating it introduces some phase shift that must be accounted for. I think ideally you should run the crossover a bit higher so the port is not active, and then delay the mains to the sub...but I am not an expert on these matters, so hopefully one of the engineers will do some myth-busting if I stated something wrong or poorly.

Now if you speakers are not collocated, you delay the nearest speaker back to the farthest speaker. For example, if your subs are center packed you delay the subs back so they align better with the mains. When the speakers are not collocated, keep in mind you can only dial in perfect phase response at one location, so you need to choose where the best sound will occur. I usually stand in the prime listening spot and use a tape measure to determine the physical offset between the subs and mains, this gets me in the ballpark before I do the alignment.

The idea is to try and start the alignment so the drivers are phase aligned within 1 wavelength if possible. Why? Because when you are tuning for the null it's impossible to tell if the speakers are 180, 540, 900 or 1260 degrees out of phase...it all sounds exactly the same as 180 degrees since your only using one frequency.

Ideally when you flip the polarity the system is perfectly aligned and the phase shift is 0 at the crossover. Unfortunately there is no guarantee you hit the mark if you're doing this without test equipment. If you miss the mark, phase may be off by 360, 720, 1080 etc at the crossover frequency. In other words the signal is in phase, but time displaced by 1, 2, or 3 wavelengths.

I believe any resultant problems occur above and below the crossover frequency, because the high pass and low pass signals will not be perfectly in phase throughout the crossover region. If you miss but are relatively close (say your off by one wavelength), the result will be some minor non-linearity in the frequency response (comb filtering), so it's nothing to get overly worked up about.

Like I said, do the best you can with the tools you have available. If it doesn't sound right, think about it and try again. It get's easier with practice.
 

Attachments

  • frequency wave length chart 2013.pdf
    103 KB · Views: 80
Last edited:
  • Like
Reactions: 31HZ and mikewalker
Whoa, this is gonna take a while to digest. A lot of new concepts for me. Gotta gig tomorrow so it might be a couple of days. But I will study and get back.

Thanks for all of your help wasnex.

Quick Q, what does collocated mean?

Our mixer does have a delay effect. Is that sufficient to do what you're suggesting?
 
  • Like
Reactions: Wasnex
Whoa, this is gonna take a while to digest. A lot of new concepts for me. Gotta gig tomorrow so it might be a couple of days. But I will study and get back.

Thanks for all of your help wasnex.

Quick Q, what does collocated mean?

Our mixer does have a delay effect. Is that sufficient to do what you're suggesting?

Collocated means in the same location.

If the mixer has a digital delay effect it probably won't work. If the outputs have delay it will.
 
<To be honest, we've been having feedback problems at most of our venues. I just thought that this one, being outdoors, would be feedback-free.>

Okay, FWIW, after I stopped playing professionally I started doing sound in clubs and wound up providing service at the biggest venues and concerts in the western US. The clue to your problem is your observation that you've been "having feedback problems at most of our venues". Something, probably many things are no doubt going on with your system that are quite simply incomprehensible to you at this point, and in all likelihood both with the gear you are using, and how you are using it. We all go through the learning curve. The question is, how much time do you have to sort all this out while trashing your gigs with bad sound and feedback? I suggest you hire a sound man / gal who actually knows what they're doing [not easy to find sometimes] to go through your system to see how its engineered, and then to get it up and running properly (that is, if this is indeed possible with your setup). There's been some great advise posted on this thread from some highly knowledgeable people. The problem is, just reading your OP, there's no telling what the problems really are. If you have technical questions (or any questions, for that matter), Agedhorse is a source second to none. Good luck.
 
As a general rule keep your speakers away from walls, the lower frequencies will tend to travel along the wall & it's exacerbated when they are close to the wall. Use the high pass filter on the drummer's mic channel & if you have a digital desk with RTA use it to find the offending frequencies and cut the graphic EQ to compensate for the resonant frequency although in this case you know what that frequency is.
Also turn the sub down a little and don't try and balance it from close up move 10 to 15 feet away to get a balanced level. If you listen close up you will always set it too loud.
Another thing you can try is if the sub has control over the crossover frequency lower it to around 120 Hz.
Hopefully this should help.
 
  • Like
Reactions: ktedrow
I've had feedback problems since the beginning of time until I bought this. It's basically an automatic notch filter. It takes all of the guessing out of setting up a equalizer. It's a dbx AFS² Feedback Eliminator.

AFS2_Front_full_width.jpg


Behringer makes a less expensive version but it gets confused about what is feedback and what is not feedback. I'd set it up at the beginning of a practice or gig and everything was fine. As the show went on I noticed that the vocals kept getting thinner sounding as we went on. One day I reset the Behringer unit after the vocals started sounding crappy and then the vocals were fine again for a while. The dbx unit does not have this issue. You set it up once and let it do all of the work.
 
Last edited:
  • Like
Reactions: Nevada Pete
I have the same problem with my drummer's vocal mic. We only have three mics in the system and I can still get feedback from him. It's just a matter of compensating for the "known knowns," as a former defense secretary liked to say.
 
Interesting. I would expect overall loop gain to be the only thing that matters for feedback. I do have the mic preamp gains up pretty high, to where I see clip light often, although I never hear any clipping in the vocals. In fact I felt the vocal quality was better with the higher gain, which is why I set the gains high. Maybe this is something I should look into.

Many good suggestions. One of the keys to good sound is proper gain staging - the relationships between all the places you can affect levels throughout the signal chain. Overall gain is not the only thing that matters, in fact, having channel preamps up that high is very likely to produce more feedback than balancing the gains at other places in the signal chain. A good approach to start with is to set mains and channel levels to unity or nominal gain (usually marked on the faders) and then only bring the channel preamp up as far needed to get the level where it should be. Minimizing the preamp gain will make the mics less sensitive and substantially reduce the chances of feedback. Modify from there for especially loud or soft inputs, mic technique, room considerations, etc. I also second using HPF on every channel whose source doesn't use lower frequencies.
 
  • Like
Reactions: AGCurry
Our mixer (soundcraft digital) has feedback suppression built-in but I've been afraid to use it in an actual gig. I tried it out at home and it's nerve-wracking being on the edge of feedback for so long while ringing things out. And I'm not very proficient with it.

If I was convinced that feedback suppression is really effective I would give it a more serious shot. But my experience at home was that it didn't help much because there wasn't one or two frequencies that stood out as being worse than other frequencies. But maybe this was an artifact of ringing out in a living room (small, carpets and drapes) as opposed to a venue. I also worry that with a lot of notches the vocal quality suffers.
Give that feedback suppression a try at a low stress gig. I use my UI-16 suppression quite often, especially with mic’d Kick drum. It works as advertised and seems to be just an automatic notch filter.
Also, you can try to mic that kick drum from the other side ie: if you’re mixing a hole in the front try mic’ing the rear head close to the batter. If you’re already doing that, try mic’ing the front then. Kick drums can cause all kinds of difficulties. Keep us posted.
 
  • Like
Reactions: ktedrow