The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
feedback destroyer Utility Software
Old 19th April 2014
  #121
Gear Maniac
 

Hi; I'm new here. Everything I've read throughout this conversation is kind of foreign to me. My experience using the dbx unit at the Mains insert is like this:
- First of all, it is fairly worthless in 'on-the-fly' 'live' mode. Feedback never gets close to setting a filter, but THE MUSIC will set filters all OVER the place! I don't see how one would expect it to work - unless someone just points a mic into a horn or something - because of things like the fact that there's so much delay: there's the physical distance across the stage on top of the supressor's OWN digitizing delay, and then my QSC speakers have another one.. The feedback builds so slowly! (I'll bet MONITORS might benefit.) Then there's the fact that the QSCs are so flat and accurate in freq response and dispersion that you'll never get a 'sqawk!' to trigger a filter. Plus, I've already set a few FIXED filters, so the only feedback left is singy-ringy stuff, if anything. I love channel compression though, so I try to push the GBF close to ringy anyway (that's why I also have a 31-band).

So anyway, I set nine-or-so FIXED filters - usually one or two each for vocals and overheads, then the rest for the whole live stage. Then I leave two 'live' filters set to extinguish aforementioned contingencies, and reset in 1min. I gain about.. Mm: maybe 4 or 5dB.
Old 20th April 2014
  #122
Here for the gear
 

Thanks dboomer. Congratulations on your teams' work on those units. All the brands have come quite a ways since the older products. I guess it's unavoidable that there is a trade-off between speed and accidentally setting filters on some program material, but logic in the feedback suppressor can help that to be less of an issue (the faster boxes to respond to feedback were not always those who had the most false filters set on program material). As you probably know first hand, it is quite an art to tune those boxes to perform as well as possible in all of the categories. There are always trade-offs.

An easy way to test how prone a box is to setting filters falsely on music is to play music through the box while feedback suppression is active. Any filters that get set are sure to be false, since there's no feedback present. Each box was tested with the same tracks and same input levels to avoid bias. No box is perfect. There will always be some synth out there that acts just like feedback and tricks the boxes.
Old 20th April 2014
  #123
Here for the gear
 

Hey Joe,

It is hard for the feedback suppressors to kick in while the band is playing, like you have experienced. And the feedback suppression in the new dbx PA2 has been greatly improved over the old boxes, including how prone it is to accidentally setting filters on music. Like I mentioned in a previous post, all of the brands have come a long way in this category since the old products released a decade or more ago.

I would use live filters the same way you do, more as a safety net than anything else. First and foremost, EQing the system will give you a lot, then I would ring out the system like you do and catch a few obvious feedback tones to get a few more dB out of the system for the performance. After that, I'd turn down a little for the performance to avoid that feedback that hangs around with the music but never quite grows all the way. With any feedback destroyer, you'll probably see a few dB increase in GBF but not a lot more. Just the way it goes (due to the nature of the feedback). Good room EQ will go a long way.
Old 20th April 2014
  #124
Lives for gear
 
Wyllys's Avatar
 

The way to "destroy" feedback is to set things up properly in the beginning. To this end one can well employ one of the many "feedback destroyers". This, however , is something of a misnomer. A much better term would be "auto-setting parametric equalizer" as that would more properly describe the practical function of such units.

I have used a half- dozen different brands including Peavey, Shure, Behringer, DBX and Sabine. The DBX was the most troublesome to employ with the user interface buried in a larger feature set and suffered from lack of accessibility. Others also had their weak points, but performed the task of auto- setting PEQ filters well enough.

The Sabine GrapiQ , however, stands head and shoulders above the others for ease of use, feature set and functioality.

With all the units I've used you can leave one or two filters in " dynamic" mode to give you the "exterminator" function to catch any transient spikes. But overall. the best use is to ser them up in advance as a part of setting things up to AVOID feedback rather than deal with it after the fact.
Old 21st April 2014
  #125
Lives for gear
Well there's only so much "pre-setting" you can do with reference to feedback prevention (as opposed to setting for tone). If the mics move or someone starts singing wearing a cowboy hat then things change and adjustments you previously made may no longer be valid. And may even be counter productive.

I would recommend setting very few ( or no) static filters. Leaving them in dynamic mode will usually be more effective. I would also recommend using a separate FBS for each input rather than on a bus output. That way you only drop filters on the problem input and not on everything else. You also likely end up with more filters available for any single channel as when they are on a bus they tend to divide up over the number of problem mics.
Old 21st April 2014
  #126
Lives for gear
 
loujudson's Avatar
Quote:
Originally Posted by The Mawg View Post
Suggestions for Feedback Destroyer Units
I work for dbx and have run some very vigorous tests on all of the main feedback destroyers (old and new). I won't give specific numbers, but I will give some general guidelines.

Invest in the newer versions of each product!
(edit)
my 2 cents.
Great to hear from someone who knows and has tested! I have two dbx 224s and only use them for conference sound where folks are cluseless and suing headmics or lavaliers. I use it on Music 1 mode, preset the 12 filters and use the seond set for dynamic problems.

Is the 224 really that out of date, or do you think it is okay? It is hardly worth updating it becasue I only do two to four such things in a year.

My live music skills have made it unecessary to use them on music gigs, especially with the newer digital mixers that have excellent filters and EQ...
Old 21st April 2014
  #127
Here for the gear
 

AFS224 still good?

Quote:
Originally Posted by loujudson View Post
I have two dbx 224s and only use them for conference sound where folks are cluseless and suing headmics or lavaliers. I use it on Music 1 mode, preset the 12 filters and use the seond set for dynamic problems.

Is the 224 really that out of date, or do you think it is okay? It is hardly worth updating it becasue I only do two to four such things in a year.
I would say you have a good product on your hands that performs respectably. I have measured its performance as well, and here's what I've seen:

The AFS224 filters are as minimally invasive as modern competitors / dbx products.
The suppression time is a bit longer on average than the new products.
It is better at distinguishing between feedback and music than the DriveRack PA+, but not as good as the DriveRack PA2.
Old 21st April 2014
  #128
Here for the gear
 

Quote:
Originally Posted by dboomer View Post
Well there's only so much "pre-setting" you can do with reference to feedback prevention (as opposed to setting for tone). If the mics move or someone starts singing wearing a cowboy hat then things change and adjustments you previously made may no longer be valid. And may even be counter productive.

I would recommend setting very few ( or no) static filters. Leaving them in dynamic mode will usually be more effective. I would also recommend using a separate FBS for each input rather than on a bus output. That way you only drop filters on the problem input and not on everything else. You also likely end up with more filters available for any single channel as when they are on a bus they tend to divide up over the number of problem mics.
Definitely agree. Just with people moving around, the feedback path (and thus the problem frequencies) can change drastically. What I really like to do is do the proper placement / room EQ, and then set some "setup" filters with the performers in place in front of the mics, and have them move around their general areas to activate any potential problems. Then I switch to "performance" mode to have a couple filters just for safety during the actual performance.
Old 21st April 2014
  #129
Lives for gear
 
loujudson's Avatar
Quote:
Originally Posted by The Mawg View Post
I would say you have a good product on your hands that performs respectably. I have measured its performance as well, and here's what I've seen:

The AFS224 filters are as minimally invasive as modern competitors / dbx products.
The suppression time is a bit longer on average than the new products.
It is better at distinguishing between feedback and music than the DriveRack PA+, but not as good as the DriveRack PA2.
Thanks! They do work well, glad they compare well in tests... I'll keep em!

I have two Behrys from the early 90s I should sell, not used for a dozen years... Craigslist here they come! :-)
Old 21st April 2014
  #130
Here for the gear
 

Easiest thing I use is the Presonus 16.4.2. If you need larger, they have 24 & 32 channel units, or can link units together. They have what's called a "fat channel" that v is nothing more than a semi-parametric eq with notch filtration. Works very well and replaces tons of outboard gear making setups much easier. They are a little spend, $1800-2000 U.S., but no more effects units, eq's, compressors etc. Everything is in the board, plus can be controlled with an iPad and each band member can control their own monitor mix with an iPhone, or iPod touch.

Setup is fast beater all you need is to plug in amps/speakers. You can save scenes of every place you play and record every performance if you want. Best piece of gear for the band that has to mix itself. Presonus has video after video with tutorials on YouTube if you need help..
Old 22nd April 2014
  #131
Lives for gear
 
loujudson's Avatar
A mixer is a different thing from a feedback remover! Even if it is easy, it does not have the auto filter function... but anyway, I used to use Prsonus StudioiLive until the Allen & Heath Qu-16 came out. It is a vastly superior device for my uses!

Anyone want to buy a Presonus 24-4-2?
Old 22nd April 2014
  #132
Gear Maniac
 

Hi, Mr. Mawg

Thanks for replying - I've lurked on Gearslutz for a long time, and have learned so much! sorry I didn't reply sooner.

Hey - don't think I was complaining; like I said, I don't see how anyone would expect it to work with a band playing, when a PA is already almost flat - especially when there's compression and expansion twitchin' the gains around.. I think that one can get a good sense of what I'm thinking by noticing the behavior of the feedback, while doing the fixed-filter ring out with the '224 across The Buss: only the first few resonances are distinct and 'alone'; then you start getting multiple weak tones. That's why I've found it effective to bring up one (applicable) mic at a time, and grab their first resonance.

I'm so pleased with my Soundcraft / QSC system - I can do smaller gigs with no EQ at all, no problem (especially if it's in a bigger room.. Well - more like "if the performers aren't against a wall") - but in many situations, I get kick out of seeing 'how much gain I can get' (really 'how much channel compression can I get away with)! No: I'm not 'crushing the life out of them'; just 'making it sound more like a record'.

Anyway, I'm very pleased with The'Ol' AFS. Because of the '224, I most often don't have any 'notching' on the dbx 1231; just get to use it for general shaping/room EQ.
Old 22nd April 2014
  #133
Lives for gear
 

Quote:
Originally Posted by dboomer View Post
Well there's only so much "pre-setting" you can do with reference to feedback prevention (as opposed to setting for tone). If the mics move or someone starts singing wearing a cowboy hat then things change and adjustments you previously made may no longer be valid. And may even be counter productive.
While in the purest technical terms this is true…but in practical terms this is vastly overstated, somebody could get the impression that we can't work without these boxes when nothing is further from the truth.

'Good' sound Engineers setup and tune systems everyday only to have the artist and musicians run around the stage (and even into the audience) without problem. The fact that we never see these boxes in professional tour rigs or in good installs says a lot.

Quote:
I would also recommend using a separate FBS for each input rather than on a bus output. That way you only drop filters on the problem input and not on everything else. You also likely end up with more filters available for any single channel as when they are on a bus they tend to divide up over the number of problem mics.
This is the best advice on the use of these devices, using one of these boxes across a bus makes little sense
Old 22nd April 2014
  #134
Lives for gear
Quote:
Originally Posted by Samc View Post
While in the purest technical terms this is true…but in practical terms this is vastly overstated, somebody could get the impression that we can't work without these boxes when nothing is further from the truth.
It's difficult in a couple of sentences to reply to a topic on a forum like this because there are so many different levels of operation going on here. Plenty of people work without comp/limiters per channel. I don't. Plenty of people work without 4-6 band parametrics per channel. I don't.

Can you work without one? The probability is high? Can you work better with one? The probability is high? Are they more efficient at dealing with feedback than most engineers? The probability is high

When pre-tuning a system there is a big difference between tweaking for tone and tweaking for GBF. When a system is tuned with wider filters (such as a graphic) that is a good way to adjust tone but it will scoop out much more than necessary for eliminating feedback. Trying to do both with a graphic is almost always a compromise.

Best of all worlds, tune your system for tone and sensibly for feedback and install a FBS on each channel that could experience feedback. Set the FBS to be dynamic (it only comes on in the presence of feedback). Now if you get through the performance with zero feedback then thenFBS sit idle outside the circuit. If you get feedback then it goes to work.
Old 22nd April 2014
  #135
Lives for gear
 
Wyllys's Avatar
 

Quote:
Originally Posted by Samc View Post
While in the purest technical terms this is true…but in practical terms this is vastly overstated, somebody could get the impression that we can't work without these boxes when nothing is further from the truth.

'Good' sound Engineers setup and tune systems everyday only to have the artist and musicians run around the stage (and even into the audience) without problem. The fact that we never see these boxes in professional tour rigs or in good installs says a lot.


This is the best advice on the use of these devices, using one of these boxes across a bus makes little sense
Quote:
Originally Posted by dboomer View Post
It's difficult in a couple of sentences to reply to a topic on a forum like this because there are so many different levels of operation going on here. Plenty of people work without comp/limiters per channel. I don't. Plenty of people work without 4-6 band parametrics per channel. I don't.

Can you work without one? The probability is high? Can you work better with one? The probability is high? Are they more efficient at dealing with feedback than most engineers? The probability is high

When pre-tuning a system there is a big difference between tweaking for tone and tweaking for GBF. When a system is tuned with wider filters (such as a graphic) that is a good way to adjust tone but it will scoop out much more than necessary for eliminating feedback. Trying to do both with a graphic is almost always a compromise.

Best of all worlds, tune your system for tone and sensibly for feedback and install a FBS on each channel that could experience feedback. Set the FBS to be dynamic (it only comes on in the presence of feedback). Now if you get through the performance with zero feedback then thenFBS sit idle outside the circuit. If you get feedback then it goes to work.

Yes, it's true that there are many levels of application for sound reinforcement. I'm going to go with Sam on this.

Properly set up, mains OR monitors should be set up to be stable and yield maximum GBF prior to performance. Once things start it's too late. The one thing you can't pre-set is the "talent". No matter what or how much you do, you can't idiot-proof your stage.

Otherwise, do your best to set things up to handle the job. With the GraphiQ's, I have the use of a 31 band graphic (with adjustable Q) and 12 filters/channel of selectable parametric, fixed or dynamic filters. This makes for a ton of options and flexibility in set up, control and access as well as being able to take advantage of using a couple of dynamic filters as "feedback destroyers".

But again I would classify such devices properly used as "Automatic Parametric Equalizers" rather than "Feedback Destroyers". You should not have to destroy feedback if you have properly set up your system. There will always be transients and a dynamic filter can be handy to catch them. But major and continued instances of feedback just should not be in the picture if you've done your preparation properly.

Ounce of prevention, pound of cure.
Old 22nd April 2014
  #136
Lives for gear
Quote:
Originally Posted by Wyllys View Post
The one thing you can't pre-set is the "talent". No matter what or how much you do, you can't idiot-proof your stage.
I keep trying ... but they keep making better idiots

Quote:
But again I would classify such devices properly used as "Automatic Parametric Equalizers" rather than "Feedback Destroyers".
Which is what they are (or maybe automatic notch filters). They simply (well not simply) measure for problems and then drop a very narrow filter directly on top of it.
Old 22nd April 2014
  #137
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by dboomer View Post

Quote:
But again I would classify such devices properly used as "Automatic Parametric Equalizers" rather than "Feedback Destroyers".
Which is what they are (or maybe automatic notch filters). They simply (well not simply) measure for problems and then drop a very narrow filter directly on top of it.
While this is perhaps an overly esoteric point, addressing a system's tendency to feedback with EQ is just fixing the "symptom" caused by speaker sound being picked up by the mic and being recirculated multiple times through that electro-acoustic loop. This repeated summation of a delayed version of the original creates a comb filter response transfer function with alternating bumps and dips. Feedback rings generally occur centered on these bumps in the transfer function where that bump level increase when added to the forward path frequency response is greater than unity. (This is why feedback only occurs at a small handful of spot frequencies and doesn't just re-emerge nearby at only a slightly higher/lower frequency when you notch one out.)

Speaking hypothetically, If one were to identify the time delay between the loudspeaker and the mic picking up the signal and subtracting a similarly delayed version of the speaker signal from the mic input, one could net out the feedback sound from the desired mic pickup sound.

In theory this could effectively squash any trace of feedback, and instead of punching holes in the frequency response the system would be restored to nominal flatness, or as flat as it could be in an ideal world.

This would not be simple to tweak on the fly but it would probably not be that hard to connect a microprocessor or DSP to the system during set up to allow it to ping the system with clicks or tone bursts to learn the system delays, and tweak acoustic path frequency response.

I have not performed a proof of concept, so feel free to dismiss this an old man talking smack... but it sure seems like it should work (to me) and in theory deliver superior performance to current feedback abatement approaches, that IMO only address the symptom and not the problem.

I don't have any present plans to pursue this, I am kept occupied and amused working in a different product area and market.

JR

PS: There may be some canned DSP software for echo cancellation in telephone systems that is close enough to this mechanism, at least for recognizing the delay using signal correlation metrics.
Old 23rd April 2014
  #138
Gear Maniac
 

Dude… I think you're right. What a cool place this is.
Old 23rd April 2014
  #139
Gear Head
 

Basic Knowledge

Quote:
Originally Posted by dboomer View Post
I always wonder when i see answers like this.

First ... gain structure has zero to do with feedback.

So assuming you "set up your PA correctly", as soon as a single microphone moves even inches the frequencies that feedback can and\ cannot occur change. So how do you set a PA up correctly to avoid this?
Obviously, Gain structure has a GREAT DEAL to do with Feedback!!!

Everything starts from Gain Structure. If it's not carefully 'treated' it and give you more issues than what you think it's not possible.

To put it simply, to have 'good' Gain Structure, your mix will benefit as you will have more control over all the channels. It will be a lot easier to manipulate your mix and you can ride the fader for whichever needs to shine.

It works the same in the MON world and it works the same from Rock to Opera.

Learn your Gain Structure first.
Old 23rd April 2014
  #140
Lives for gear
 
Wyllys's Avatar
 

Quote:
Originally Posted by Simoniz View Post
Obviously, Gain structure has a GREAT DEAL to do with Feedback!!!

Everything starts from Gain Structure. If it's not carefully 'treated' it and give you more issues than what you think it's not possible.

To put it simply, to have 'good' Gain Structure, your mix will benefit as you will have more control over all the channels. It will be a lot easier to manipulate your mix and you can ride the fader for whichever needs to shine.

It works the same in the MON world and it works the same from Rock to Opera.

Learn your Gain Structure first.
Sorry, but strictly speaking how and where you get your gain (gain structure) is not the determinant. Feedback occurs when the amount of sound at the mic from the system is high enough to begin a self- reinforcing loop.

This has nothing to do with how you stage your gain, only how much total gain there is. Divide it up how you will, it's the end result...total gain...that matters.
Old 23rd April 2014
  #141
Here for the gear
 

Truly Flattening a Feedback System Response

Quote:
Originally Posted by JohnRoberts View Post
Speaking hypothetically, If one were to identify the time delay between the loudspeaker and the mic picking up the signal and subtracting a similarly delayed version of the speaker signal from the mic input, one could net out the feedback sound from the desired mic pickup sound.

In theory this could effectively squash any trace of feedback, and instead of punching holes in the frequency response the system would be restored to nominal flatness, or as flat as it could be in an ideal world.

This would not be simple to tweak on the fly but it would probably not be that hard to connect a microprocessor or DSP to the system during set up to allow it to ping the system with clicks or tone bursts to learn the system delays, and tweak acoustic path frequency response.
You're right in all the theory, however it is verytricky to reduce this to practice. The location of the feedback peaks and dips is so highly dependent on phase/delay, that the most minor change will shift the location of them significantly. If you could be guaranteed your mics and performers would stand perfectly still, it might work. But even small changes in temperature can change the speed of sound enough to cause a feedback loop time to change enough to shift the location of the peaks and dips significantly.

Also think that if the first peak shifts by .1 Hz, the second peak will shift by .2 Hz, the 10th peak will shift by 1 Hz, etc. So the higher the frequency, the less likely you are to properly cancel out the peak there, because any estimation error of the feedback delay time is compounded so many times. If you had a super short feedback loop time, you'd have so few peaks that you could probably pull it off ok.

Another huge issue is that you don't usually just have one ideal feedback path. You'll get multiple reflections off the walls, multiple mics picking up your signal, etc. greatly complicating the whole issue.
Old 23rd April 2014
  #142
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by The Mawg View Post
You're right in all the theory, however it is verytricky to reduce this to practice. The location of the feedback peaks and dips is so highly dependent on phase/delay, that the most minor change will shift the location of them significantly.
While it is not wise for me to argue about a hypothetical, the location of the peaks changing with things like speed of sound in the acoustic path also impact the viability of all the existing notch filter mitigation approaches.

Just like the EQ does not need to be perfect to damp feedback building up, any delay cancellation only needs to remove enough energy that feedback does not sustain itself.

Admittedly multiple paths from multiple speakers and mics would be very difficult to manage on the fly, during set-up a dedicated test signal could help parse out the dominant return paths. If a path isn't dominant, it won't cause feedback until path gain is pushed silly high
Quote:
If you could be guaranteed your mics and performers would stand perfectly still, it might work. But even small changes in temperature can change the speed of sound enough to cause a feedback loop time to change enough to shift the location of the peaks and dips significantly.
Just like it does now for fixed notch filters.
Quote:
Also think that if the first peak shifts by .1 Hz, the second peak will shift by .2 Hz, the 10th peak will shift by 1 Hz, etc. So the higher the frequency, the less likely you are to properly cancel out the peak there, because any estimation error of the feedback delay time is compounded so many times. If you had a super short feedback loop time, you'd have so few peaks that you could probably pull it off ok.
If the delay time is long or short, it will affect the depth of the cancellation and energy remaining to support feedback. Indeed any delay time error is more significant the shorter the signal wavelength. Good point, this suggests that while dialing it in the "ping" should involve HF components. The ping could probably be a HF tone burst, maybe a click (while a very short HF burst will sound like a click).

I am not so arrogant to expect bench testing will not reveal problems and perhaps new strategies. No plan survives contact with the enemy.
Quote:
Another huge issue is that you don't usually just have one ideal feedback path. You'll get multiple reflections off the walls, multiple mics picking up your signal, etc. greatly complicating the whole issue.
Yes, but.... If there were a lot of similar strength acoustic paths, there would not be sharp comb filtering with strong bumps and narrow combs, instead much more diffuse combing with shallow combs and more, but weaker bumps. Also less feedback problems.

I suspect there will still only be a small handful of dominant problem acoustic paths. When "Pinging" out the system the DSP would ID the loudest return and subtract that out, leaving the next significant return to ping out until getting a fair null. I suspect a two or three delay routine takes less DSP processing horse power than a stack of notch filters. Ideally these delayed cancellation signals could benefit from some EQ also to better mimic the acoustic path.

Sorry to feed this veer... maybe I need to build one just to check this out for myself. At this point it is just mental masturbation.

I really do not care to play in that market. Been there done that...


JR
Old 23rd April 2014
  #143
Gear Head
 

Probably my words had created this misunderstandings. For sure, I understand how feedback occurs or how it would affect the neighboring frequencies when you try to 'cut' them... Anyway, feedback can never be totally cut if gain is always increasing...

What I meant was that, mixing plays a big part in the MON world. It is always a common mistake that newbies or even experienced "engineers" would make when they mix - well, they thought giving the musicians or artistes loud level would help them hear themselves better. While it's true, this 'more me' and eventually, 'more everything', will quickly build up and it will go on and on and on.... it never ends.... And you would end up 'cutting' the feedback frequencies and c'mon, whatever you 'cut', you're also affecting the neighbouring ones thus, this will never end as well. The MON mix will sound weird and whoever is playing to the mix will ask for more and more...

Clarity and making way for the instruments or vocals that the musician wants is definitely the way to practice as opposed to pushing the fader higher or turning the never-ending knob.... And to first achieve this, you must have proper Gain Structure and maintain it throughout the show. I know very well that there is no rules or a fixed formula for getting a proper Gain Structure, so it's really your experience-acquiring that would get you 'there'. You don't wanna have too much gain such that you can only move a cm of the fader(unless you have some micpre that sounds good when slightly peak or clip - whatever the case, let's move on), or you have too little gain that you need to saw another 100mm to 'up' the level or if you have a knob that turns 720deg....

When 'proper' Gain Structure has been grasped, you can better manipulate your inputs and outputs. Your mix(as mentioned in my previous post) will be easier to sit in for FOH and or MON.

Phase would definitely help in reducing 'feedback' as well... (Don't wanna elaborate on this)

Back to the original question of employing a 'Feedback Destroyer' for a live show. Yes, it works but it's a machine that would coldly 'cut' the 'peaking frequencies. The AFS uses a lot of notch filters as to preserve the best original signal. However, its main job is just to 'cut' feedback and the MON mix will sound, more than often, FLAT and LIFELESS, with no presence and clarity, sometimes hollow or muddy. The 'more me' and 'more everything' will eventually occur. Well, unless the musicians or artistes are not really interested in how it sounds and just wanna quickly get it done with, this 'disaster' will still prevail.

Of course, there are a lot more ways to increase the level and have better clarity but I shall end here as it would be another topic totally.
Old 25th April 2014
  #144
Here for the gear
 

Yeah, like has been mentioned somewhere in here, there's definitely a trade-off between tonality and getting maximum gain before feedback. Ideal for max GBF is a perfectly flat response, which doesn't sound very good to the ears. One thing I don't think a lot of people realize is that by pushing the gains until they almost feed back, they are already significantly coloring their mix with some pretty bad comb filter-ish peaks. In a lot of cases the feedback destroyer, by notching some of those peaks, is making your overall response more closely match what it would have been without any feedback present. Best way to not color your mix is to not push your gains to the limit, if it can be helped, because having a high feedback gain is like automatically adding a bunch of peaking filters to your system.
Old 25th April 2014
  #145
Here for the gear
 

Quote:
Originally Posted by JohnRoberts View Post
While it is not wise for me to argue about a hypothetical, the location of the peaks changing with things like speed of sound in the acoustic path also impact the viability of all the existing notch filter mitigation approaches.
[edit]
Sorry to feed this veer... maybe I need to build one just to check this out for myself.
JR
True, the same problem affects existing feedback destroyers. They typically allow their filters to move around slightly in order to compensate for this. However, once you switch to performance mode, usually those setup filters get fixed and immovable, and so they definitely could run into this problem.

It would be much harder for the delay compensator thing to accurately account for this. Can't ping the room during a live performance as easily, and also the fact that it's bound to not be accurate enough to cause high-frequency filtering in the correct locations. I didn't realize this until I tried building one myself.
Old 25th April 2014
  #146
Lives for gear
Mawg-

The other side of this I frequently see is the unnecessary and over use of 31 band GEQs filters to combat feedback which often cut way more useable material in the name of stopping feedback when much narrower filters would have done the job.
Old 25th April 2014
  #147
Lives for gear
To JR's proposal ...

The cancellation does not need to be complete to be useful. Simply every dB of cancellation that could be achieved would normally result in an equal improvement in GBF.
Old 25th April 2014
  #148
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by dboomer View Post
To JR's proposal ...

The cancellation does not need to be complete to be useful. Simply every dB of cancellation that could be achieved would normally result in an equal improvement in GBF.
Yup...

Another thing, while I haven't tested this on the bench I suspect the Q of the comb filter (or narrowness of peaks and dips) depends on how close the gain is to unity. So EQ will generally be an imperfect match for the coloration changes caused by delay re-circulation.

This feedback transfer function is not unlike the behavior of feeding the output of an audio delay line back into it's input. When that is recirculated with more than unity gain we get the same feedback-like result. When you get a delay line right on the hairy edge of running away, you get the hollow sounding coloration of a live performance stage on the edge of feeding back.

JR

PS: I think this may be called a transversal filter but I don't have enough book larn'in to be confident about nomenclature.
Old 25th April 2014
  #149
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by The Mawg View Post
True, the same problem affects existing feedback destroyers. They typically allow their filters to move around slightly in order to compensate for this. However, once you switch to performance mode, usually those setup filters get fixed and immovable, and so they definitely could run into this problem.

It would be much harder for the delay compensator thing to accurately account for this. Can't ping the room during a live performance as easily, and also the fact that it's bound to not be accurate enough to cause high-frequency filtering in the correct locations. I didn't realize this until I tried building one myself.
I mentioned earlier about the possible application of echo cancellation technology used in telephony. Sophisticated DSP routines can perform correlation measurements between the signal being sent to the speakers, and the raw sound received by the mic. Another approach could be a simpler periodic test of the depth of cancellation, with the nominal delay, and adjacent longer/shorter streams. If the longer or shorter audio stream cancels deeper, adapt and tweak the delay after a definite pattern of shorter/longer delay is determined with confidence.

Of course the more diffuse the feedback signal the more complex any real time attempts to correlate on the fly.

If this was easy somebody would have already done it (or it doesn't work).

JR

[edit] I guess a digital console with internal signal delay capability might support manually experimenting with this. While I wouldn't try to do this manually in the field. {/edit}
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump