I’ve used the Feedback Assistant (FA) for two live performances so far and had the same problem both times. I let the FA control feedback during the performance, and rather than keeping the notches as small as possible, it kept pushing them down and (more destructively) widening them so far that it started cutting out a huge swath between about 100 to 1500 Hz. See the two attached photos. The first was half way through the performance, at intermission. The second was after the performance was over. As you can see, it kept widening the notches, even after I turned on HOLD. You can see at several points the frequencies being push all the way down below the graph. I was lucky in this case that it wasn’t too destructive to the sound; I could notice it, but it still sounded good. During the second gig, the same big swath got pushed down so far that it seriously affected the sound; the majority of it was below the bottom of the graph (didn’t take photos because I was too busy reining them in).
Has anyone else had this problem with the Feedback Assistant during a live performance? Am I just doing it wrong? At this point, I don’t feel confident using the FA in a live performance. I don’t want it constantly widening the notches and pushing them down to that extreme.
Wow, that looks a bit off!!
I use the “Live” setting and have watched it recover and change frequencies as our gigs progress.
It’s been a boon to me, especially in tight locations.
I have to say it’s only been notching and nothing like you’ve encountered.
Not sure what the “fixed” setting does but there’s a lot of energy there it thinks needs cut.
Have you tried changing mics/moving positions/moving PA?
What happens when you disable the feature - can you get any volume at all from your mics before it starts to howl?
Once again guys, please ref to the manual before blaming hardware:
False Positives and best practices:
False positives can occur with any feedback-like sound being played (e.g. guitar feedback, keyboard and synth sounds, flute, operatic vocals etc…).
The FBA cannot know if this is a desired feedback-like sound or not.
The FBA on all channels will always boot with HOLD active to prevent unnecessary filters being added for this reason.
It is best not to leave any FBA actively detecting and adding fixed filters. Instead, if being left unattended, it is best to leave the filter Mode as Live.
Just to clarify- I’m definitely not blaming the hardware - I’m lauding it!
I’ll read the manual when I encounter something that’s causing a problem for me and I can’t figure it out?.
I use the FBA on stage monitors, but switched it off after the first time of trying it out on FOH due to it pulling out frequencies left right and centre. I very rarely need to cut frequencies on FOH, I have one or two problematic frequencies notched and that’s it. But I find it works fine on aux’s, too a point. Have used both the hold and live settings.
The band I play and work with are not loud players on stage which really helps in controlling any feedback issues. Get the basics right and you can pretty much have a stress free performance.
FBA live at my last gig CREATED FEEDBACK that were very hard to understand, I passed a very stressful moment. I strongly advise not to use FBA live except if very necessary and with much care.
Let me explain:
The room was difficult, narrow, rectangular, with lot of reverb. So after the sound balance, I asked the musicians to leave their instruments on stage then pushed levels to detect feedbacks on each output with fixed FBA. Up to this point, great.
Then, because of the room difficulty I decided to activate live FBA, both on PA and monitors. Bad idea.
The musicians were violon, accordion and flute. The FBA is bad to distinguish these instruments from a real feedback, so it started progressively to reduce all mediums everywhere, automatically, without notice.
Consequences:
I grew up main level to compensate this loss in the mediums (I shad not yet realized this issue). Technically, due to the FBA, that was equivalent to increasing high frequencies. That created feedback in highs. After several seconds the feedback was detected probably by FBA. And some random time this nasty feedback came back. I understood nothing. I finally totally cut the monitors, and there was still feedbacks randomly appearing in the Highs only due to the PAs (this room was difficult).
In addition: one musician told me the sound was terrible on their monitor. Yet I didn’t hear issue on the headset monitor. The PA was also degraded but not as much.
Now that I understand what happened it is clear, and know what to do. But during the show I was lost. When I saw the terrible FBA filter, I didn’t dare disable it by fear of more feedbacks. In the end, the auto FBA is very risky because we don’t control it, which can end to increase feedback. I’ll avoid any automatic decision made by the “IA”, too risky.
(As a feature proposition, I would suggest using the IA to make notification/proposition, but let the sound engineeri decide to accept it or not. That would be valuable)
Same thing here, by the end of my first set the FA pulled out so many frequencies I thought I had blown a speaker, the sound was really thin and sounded like crap.
I looked at the FA and it had pulled out most of the low end frequencies, I turned it off and reset everything and the PA came alive again.
I wouldn’t bother using it again until it gets sorted out.
I was soundly chastised for suggesting the probability of exactly these issues with an auto feedback algorithm when the CQ was introduced. Gutting the sonic landscape is the price paid for failure to properly set up and perform professionally by relying on automatic filter placement.
Hugh
CQ18T is, in my opinion, not designed for professionals. If you’re a professional you should use a mixing station more complete and faster to use than CQ18T (except for its small size). I’m not a professional but a musician who support other musicians in small gigs
I didn’t rely on FAB, it was not necessary. I though it could be convenient to avoid potential feedback issues that have occurred in this setup.
So that leads to : this feature has a low quality, compared to the overall high quality level of the CQ18T, and given the targeted users.
In my opinion this feature should be either improved (better algorithm or ability to control it) or removed, to avoid more users to learn by bad experiences.
Audio automation is a seductive enticement that seldom , if ever, fulfills it’s promise! There is no short cut or algorithm that can replace the essential elements of know how to establish an effective SR set-up. It aint rocket science, it is a simple matter of making an effort to gather the essential SR protocol dots and connecting them.
The CQs are audio tools that obviously are targeted for entry level SR beginners however the CQ 20b also is loaded with a powerful FPGA 24/96K processor that is scaled back for the limited number of I/Os that it has internally available. I have a well known established 50+ year history of high level professional audio activity however I plan to deploy a CQ20b this summer for several solo gigs if the internal I/Os are acceptable and the ancillary recorded playback system is reliable and effective. The tiny footprint desk along with my Flea 47 tube mic and EX10s will all fit in the trunk of my Camry and will be a breeze to set up. The fact that I have a DX32 loaded with primes and D-Live I/Os along with a SQ5 and all sorts of ancillary gear does not preclude my need to deploy for certain gigs a small CQ20b.
Truth told I will not need or use any of the CQs Automated features, It’s supposed to be a pretty great audio tool without any of those features for the solo gigs I am often called on to perform. Either it will perform satisfactorily or it will not: I will find out this summer!
Hugh
I would not write off the FBA, like any tool it needs to be used offline to allow yourself to familiarize yourself with its foibles. It cannot be used without the usual anti-feedback precautions that should be taken at any gig.
Proper placement of microphones in relationship to speakers, preferably a choice off mics, Cardiod and SuperCardiod to allow for awkward FoH or monitor speakers. Once all that is done, maybe you could use the FBA.
I find its the same with Auto Gain. It takes too long to lett every mic be set by Auto Gain. I set them manually. If we had a long time to set up, it would not be a problem but so often we have just 1 hour to set up.
@ChienMort: The difference with the auto-gain is that you can easily balance a gain decrease with a level adjustment and it’s fine, while the Live (auto) FBA going bad ruins your sound: put a violin and a flute, it goes crazy no matter your audio setup and mic placement. It’s hard to practice this alone offline, because you need a band and it depends on the instrument. And anyway there’s nothing you can really learn if it goes nuts, because the automation is not configurable. The only thing you learn, the hard way, is: don’t use it. It might sometimes be useful if you are sure that there is no incompatible instrument on stage (violin, flute, etc.): that requires either a list of compatible/incompatible instrument from A&H, or to carefuly check compatibility with the band on stage before the show. In my opinion anyway the risk is too high I won’t use that feature again. That might change with more control over it and/or a better detection algorithm in future updates.
Fixed FBA however is useful, when the filters are fixed and configured before the show.
@Hugh, the CQ form factor + sound quality (IMO) + easy record&playback is indeed great
Actually I see one case when today’s Auto Live FBA could be useful :
I’m playing with my small band, without a sound engineer, on a tiny “stage” (a pub, a shitty room…) without space for optimal placement and with little time for sound check. All the band’s instruments are compatible with the CQ’s live FBA: that was checked in advance during rehearsals.
In that case, which does occur more often than wanted, the Live FBA would indeed be useful, so that musicians can keep playing in case of feedback.
Hi.
I’m relatively new to the the CQ-18T and have used it for a few gigs. I’ve experienced exactly the same issues as described at the beginning of this thread.
Being the guitarist in an acoustic duo, I don’t have hands free to make adjustments mid-song, so the FBA is (was) an enticing feature. I still like the idea of having a safety net to cut out a temporary problematic frequency mid-song, but I understand from experience and the comments above that leaving the FBA in Live mode potentially causes more problems than it solves. Is there a way of limiting the number of live filters (maybe to 4) to get some benefit without allowing the FBA to go ‘rogue’?
I know everybody goes on about mic placement etc, but we do try to take this into account and we still get problems. We are running 2 mics and an acoustic into the CQ-18T which feeds into one channel of a Bose L1 Pro16. We normally set up behind the Bose and don’t get above halfway on the master volume without being plagued by feedback, so I think our problems must be from reflected sound. Our audiences are always telling us we’re not loud enough, so if anybody’s got any suggestions of things to try, I’d be grateful.
Cheers,
Roger.
I know “live” mode is designed to be left on during an event. I would advise not using this mode for all the reasons already stated.
However, if you are going to use this mode, I would suggest that you turn the “Live Recovery” setting to as “fast” as possible. This will ensure that cut frequencies are “released” and returned to normal as quickly as possible which should help prevent all of the frequencies from being removed and not added back. If the release of a frequency causes feedback again, the system should automatically remove it again. If this becomes a problem, you can lengthen the time little by little until it’s not a problem, or even better just cut the offending frequency manually using normal feedback prevention techniques.
That being said, I don’t have one of these units and I have no idea how fast “fast” really is. It could still be too slow to be of real use. I’d be curious to know how fast it is if anyone wants to test it.
While I don’t use “feedback filters” but they all will do what you have found. They have no way of knowing if what it “hears” is feedback or not during a show.
If you are going to use it, set up your system open up the mic channels, turn on the FBA push up the system level until it starts feeding back, let the FBA grab a few frequencies and then lock it so it does not keep notching out everything when the show starts.
Many thanks for the tip. First impressions indicate that setting the recovery time to as fast as possible does improve things dramatically. I haven’t had a chance to try it out at a gig yet though.
I agree that FBA is completely unusable on this mixer. I don’t agree that this is a fundamental problem with the concept of feedback detection. It’s a problem with its implementation in the CQ series. I’ve used Soundcraft’s AFS and it’s genuinely useful. I think the difference is simple – the Soundcraft waits for the feedback spike to be seriously problematic before cutting it. This makes it a very useful tool while ringing out a room.
The CQ jumps the gun too quickly. I’m hoping that a future update fixes the sensitivity, which hopefully reduces the false positives.
