AI can STOP live microphone feedback

[@dslynx]

Finally! A use for AI that isn't completely garbage. Of course, this is all good till it starts changing the audio and producing words completely different from what you are saying. ;)

[@devinlsheets_alphasound]

Soon we’ll be adding heresy filters for churches, so you can select between denominations like Baptist or Lutheran, etc.

[@Djhg2000]

That 0 latency claim is what sketches me out a bit. It can be negligible, like 10 ns, but it's not going to be 0.

[@devinlsheets_alphasound]

What we mean by zero latency is that there’s no difference between running or bypassing the plug-in

[@duanesheets5484]

Waves lists many of their plug ins as having 0 latency.

[@Ozoingo2]

Zero processing latency and zero transport latency are two very different things. These don't add latency to the processing chain but you still have to factor in the AD and DA conversions coming in and out of the computer. Stuff like soundgrid, DVS or even the Yamaha dm3s USB interface can bring this down to around 6ms roundtrip in my experience with other plugins

[@duanesheets5484]

He’s not taking about hardware. He’s taking about the plug in. You choose whatever hardware you want and get the latency you get.

[@Djhg2000]

​@@devinlsheets_alphasound Well, there's always _some_ delay when using a plugin. Even something as simple as a gain plugin is not zero latency, it's just really really short. The same goes for the mixers themselves. In the case of analog mixers stem from things like input capacitances and inductances, and a simple capacitor based lowpass filter will also introduce somewhat variable latency in the order of up to a quarter (IIRC?) wavelength of the Nyquist frequency. For digital mixers it's similar, like the quantization time of the ADC and DAC and the delay of whatever fixed function logic/FPGA to process the filters, with the added buffer length of the pipeline. Probably faster than the analog mixer, but still not zero. From my understanding this software uses a real time filter and a non-realtime (but fast) AI algorithm to tune that filter. The filter would then still introduce latency, even if it's short and predictable with fixed function logic. I'm not talking about hardware latency or the inherent latency of Windows subsystems, the actual filter inevitably has some delay no matter how you do it. Like I said the delay can be negligible, but it's not zero.

[@dvuckovic]

Can it be inserted across a group, ie a panel discussion?

[@devinlsheets_alphasound]

Yes, but make sure no more than about three people are ever talking at the same time.

[@SonuSharma-bg8bb]

Hello brother make a complete video on this topic, how will we set it up, what will be the connection, then make a complete video on this so that it becomes easy for people to solve the whistle problem and how can we use it in any other mixer other than DM3

[@duanesheets5484]

What whistle problem?

[@MichaelUrban-j1k]

Can't wait to test mine out tomorrow

[@mx676]

VSThost under Windows for audio? What about linux with a low latency kernel?

[@MFKitten]

Ideally it would be an ARM based processor, and it should only be running the code and nothing else. And it should be rackmount. Hopefully someone can make that happen!

[@손상표-h9q]

How do you set it up? For example, how do you connect equipment? Do you have a guide video?

[@duanesheets5484]

Alpha Sound

[@DavidSinanan]

Very cool. yeah, that's very cool. 0ms with the only 0.5 line rtt? There's so many applications for that. Making me dream of setting up a mixer for my living room (totally unrealistic for an apartment...lol). I love watching your tech.

[@dvuckovic]

Does it help with bleed into the vocal mic?

[@devinlsheets_alphasound]

It definitely helps with certain types of bleed, such as room reverb and some amount of background noise, but it isn't specifically designed to remove other forms of bleed from things like drums or whatnot.

[@dvuckovic]

@@devinlsheets_alphasound guess that's the next plugin then :)

[@Sound68]

But what happen if we increase the volume more after de feedback working. Will it remove more feekback

[@pindopurba]

can you please share the specifications for the computer that you are using?

[@devinlsheets_alphasound]

We don't publish the full specs of our recommended machines because they aren't very meaningful, since most of the magic happens at the OS configuration level. For example, that unit in the video has an N95 processor, but outperforms even some i9 based systems because of how we have set it up.

[@djrenault]

@@devinlsheets_alphasound a little bit of transparency goes a long way here. we don't believe in magic with computers. how is windows configured for more performance on an n95 than an i9 with stock windows? even if you post a full detailed guide on how to optimize windows for this, many people will still buy the premade kit, and it would be a win-win for you. i'm not interested in buying something shrouded in secrecy and magic sauce, but i could be interested in this plugin if the method of optimizing the computer were available beforehand so i could see if it's something i want to do.

[@Super_Mario_BiancoNero]

Is Mac compatibility in the works?

[@devinlsheets_alphasound]

yes

[@Super_Mario_BiancoNero]

@@devinlsheets_alphasound Is it possible to subscribe to a notification / newsletter etc?

[@oliverheywardmusic]

Is there a hardware option you have tested that could run 2 or maybe 3 instances that’s not such a jump in cost?

[@duanesheets5484]

Read the FAQ

[@nandakishore588]

is there any guideline of how much db of extra gain we can get by using this plugin ? From this video, i am assuming that this works only on mics which are on the edge of feedback or intermittent feedback , not on something that is already howling. does this mean i can get only 3-4 db of extra gain by using this plugin ?

[@duanesheets5484]

6-10 db depending on system.

[@theNitefly]

if you were going to build a dedicated piece of hardware to run this, what is the buffer size the algorithm needs (not USB buffer size)?

[@devinlsheets_alphasound]

It’s complicated, but in some sense the algorithm is buffer independent. It could operate with a buffer of 1 sample if conditions were ideal. Send us an email if you want to get into the weeds, although we can’t talk too much about it without NDAs and such.

[@cheater00]

pretty cool, as a club tech the next thing i want to see is a documented log of your device working in a hot club for a year without crashing or having to be rebooted

[@Peter-p7e]

😁 1 reset /20min

[@devinlsheets_alphasound]

Buy one and let us know how it goes. Should be fine

[@cheater00]

@@devinlsheets_alphasound i can't use unproven hardware in a live setting. this is something you need to understand. whether your box will work without supervision for a year at a time is more important than any feature it contains. you have to run these tests and report them to people. no one will risk their reputation as an audio engineer for you.

[@devinlsheets_alphasound]

@@cheater00 we’ve been using them in unmanaged installs and for our regular shows for over six months now with no issue.

[@cheater00]

@@devinlsheets_alphasound ok, please document this in detail, installation date, what the equipment rack looks like, whether it's exposed to high ambient, what the incidence of issues was, etc. that'll help you a lot as a brand.

[@toehser]

How is this better than a Shure dfr11eq or a Sabine? Fairly smart automatic feedback eliminators have been available for decades and decades... I see a dfr 11 EQ on eBay used for 40 bucks right now..

[@devinlsheets_alphasound]

Those sound terrible lol

[@toehser]

Honestly I think feedback extermination should be one device per microphone that does not involve computers or plugins, I have used three or four dedicated feedback devices and always they are placed between the microphone preamp stage and the first affect your amplification or mixing stages. I have been a harmonica player in saxophone player for a long time and standing in front of PA systems are amplifiers with a harmonica and a microphone has led me to a lot of experience with electronic feedback extermination. The more I think about it the more I don't even think this product is a good idea in the first place..... The first line of defense for feedback is always the pickup pattern of the microphone and the angle of the speakers. Automatic EQ types of things are just like the fire door in case of emergency. Certainly not something that needs improvement with AI and windows plugins...

[@devinlsheets_alphasound]

@@toehser that’s fine you don’t have to buy it.

[@leopisaq]

This is all great, and I'm here for it. But we still need the latency to go just a little bit shorter in order to make it usable for live band performances. Can't wait :)

[@devinlsheets_alphasound]

5ms is fine for most people, but for IEMs you may need even less, and that’s where native Dante and other solutions come into play, but they’re much more expensive.

[@BartKrap]

Based on how you set it up, it should be possible on pretty much all digital mixers to not have this inserted after the tap point of the monitor mix, making the latency irrelevent for that. And 5ms is not noticable in most situations.

[@Audio_Simon]

If you have more inputs can it handle multiple mic channels?

[@duanesheets5484]

FAQ

[@Audio_Simon]

​​@@duanesheets5484 I suppose this (below) is the answer, but still seems unclear to me. I saw mention of not running more than 4 on their more expensive system. So I guess 4 is the answer. HOW MANY INSTANCES OF THE PLUGIN CAN I RUN ON MY COMPUTER? As many as your machine CPU will handle, given a certain buffer size.

[@amirharony3859]

how many channels can u run with that at the same time?

[@duanesheets5484]

FAQ

[@jean-lucbattista2492]

Make a deal with behringer! One way or another every digital console will have this soon

[@SousaphoneMusic]

so how does it work?

[@slofty]

@@maxmaix _Technically the truth._

[@devinlsheets_alphasound]

magic

[@robendj]

This is awesome!

[@ps_sanjayofficial5973]

this is your profession ?

[@devinlsheets_alphasound]

It is now!

[@MrZetor]

Fascinating! Could this zero-latency AI-processing be used to transform 'wet' farts to sound as 'dry'? Or even better, make them completely silent. Asking for a friend. He said he'd pay good money for a miniaturized setup (mic, AI HW/SW, loudspeaker) that would fit in a small bag, and make farts totally silent. The system would have to work in very large and echoey spaces where lots of people are tightly packed, though.

[@Ahmikariane]

Oooh Blessed IA that will end let's us all without work.... It's very hilarious. In the future club owners won't need to have sound men as employees.... Just a computer and press start button Not to get in to debate on how this crap of technologies will do in a more complex situation. Latencies and all that implies to use and rely upon technology to do everything.

[@kwon-illee6362]

🎤

[@masterastakhovsound]

🤝👍👍👍👍