A.I. DE-FEEDBACK PLUGIN V1

[@devinlsheets_alphasound]

This plugin is currently being used on tour with Mariah Carey at FOH by Paul J. Falcone (Platinum Grammy winner, music director for Missy Elliot). tinyurl.com/bdf2v4nw The audio at the beginning of the video is 80% direct sound from the SM58 which goes to the speaker, and 20% from a stereo ambient mic nearby to capture the general sound of the room. The hardware options through the website come with the plugin preinstalled but don't include the license, you need to buy the license separately. It is mono only, Windows VST3 only. We're working on stereo, and Mac compatibility, but it might be a while. Read the full FAQ on the website.

[@shumoyoyoyaya]

man wish you had a macbook m chip option

[@chrisjaustin88]

The faq button on the site isn't working?

[@devinlsheets_alphasound]

@@chrisjaustin88are you trying to open it on your phone?

[@boydw1]

This is a good idea, but really needs to be based on a dedicated DSP/NPU chip, and integrated into hardware mixers & rackmount processors. Running it on a PC? No thanks. Running it on windoze or iOS? F*ck no!

[@duanesheets5484]

Try the demo.

[@Max-bh1pl]

The examples in the video sound fantastic, and my curiosity is definitely piqued. I wouldn’t be surprised if this becomes a success, enabling you to keep advancing De-Feedback. I’m looking forward to using it!

[@LuigiChelli]

...you did whaat??! Man, this is unbelievable, congratz!

[@demarcuscousinsthe65th]

That is one of the most beautiful speakers I've ever seen.

[@李垠-s8c]

Looks like NEXO P12

[@deegio2]

Deep learning neural feedback remover (for the V2).. your market is mainly sound engeneers and A.I. is a term negatively perceived because it threathens people's job. But also, what a great job. Like sylenth2 i was waiting for those tools for a long time, thank you for your perceverance and tenacity, hope you can dev a v2 to make it compatible with a lot of systems (like ARM & live integrated systems)

[@threepe0]

People are stupid.

[@pacosoundaruba]

This product is very good for end users which are not Pro Audio engineers such as conference rooms, churches, ballrooms and small live bands. But even in the live Pro Audio this is working perfect as a extra tool. So all you guys who are negatively about this product.. get a life!

[@parkeranderson1172]

Would love to see more about the controls/behind the scenes on this. Eventually incorporating it into a Live Professor rig seems like a great option.

[@StereoAnthony]

Really great work! However, I think it would be prudent to explain in the video that this is only for 1 mic, and that there is a PRO tier for up to 4 mics. Also, that its currently Windows only. It was the first questions that came up for me, and I didn't find an answer until going to your website.

[@duanesheets5484]

If you subgroup you can use more mics to some extent. We subgroup as many as four mics on the single channel with no problems.

[@mattmattga]

@@duanesheets5484are you involved in the dev? you’re in every reply.

[@duanesheets5484]

@@mattmattga Devin is my oldest. I'm the owner of Alpha Sound, the partner company.

[@MicroCircus22]

PLEAAAASE give us a Offline Activation :O Tested it on my Fourier Transform Engine and it works perfectly..... Now i NEED it!

[@devinlsheets_alphasound]

We are actively working on this as fast as we can!

[@TridentEventGroup]

Would love a more indepth of how the routing of sound works from the computer to the sound card to the mixer and out. Thank you!

[@duanesheets5484]

Read the FAQ

[@mattmattga]

@@duanesheets5484they’re trying to sell us on the product in initial marketing materials. The answer to every question is not “read the FAQ”. If somebody is trying to sell me something, i’m sure they appreciate this kind of feedback and will address it in future videos.

[@duanesheets5484]

@@mattmattga in the FAQ it's recommended to route through a sub group and use the insert.

[@bikeenjoyer977]

@@mattmattga not many products have a faq on how to use them directly attached to their marketing materials so I'm sure you can appreciate them doing more than usual to answer questions right at the source :)

[@ZachariahConnor]

Sounds incredible Devon!!

[@devinlsheets_alphasound]

Thanks!! It’s been an amazing year of development

[@saharasound8337]

Amazing great model. How it is connected? Mic to mixer then into audio interface and then speakers?

[@M0M...]

Ok this is huge. I would love to see more reviews on the product

[@channel19428]

Will this help in punishing conditions like a person with (omni)headset mic standing in front of a wedge ? Would love to see a demo.

[@zazyczech]

How about musical applications with body mics (around 20 pcs) - can you train it for stereo on vocal SUB? And for Apple Silicon pls 🙂 Gread job!!!

[@jensdroessler3575]

Do a SoundGrid version. That would reduce latency a lot.

[@magoostus]

ya, sell the idea to Waves, before they figure it out on their own

[@UncleWalter1]

@@magoostus Nah. That enshittification factory can keep their grubby mits off this one...

[@markkapner]

Awesome work! I’m glad everyone here has so much control over their mixing environments. But for those of us that live in the real world, mixing corporate shows in ballrooms, or worse, this fits right in. 5045’s are great, own them, but they only get you so far. I would imagine Dante PCIe cards would work?

[@ShiminyDadardon]

What CPUs were tested to work well? I'm assuming an x86 64 processor. Also does the model use integrated graphics for processing like other types of machine learning algorithms, and if so, could a dedicated GPU be utilized? Thank you!

[@cykatroopa]

You can add a slight reverb with a pitch down envelope. Needs a premium mixer though.

[@jamespwalters]

Definitely would be interesting to see more live demos. I mix about 30 musical theater performances annually for elementary, middle and high school students and would love to have a more automated way to avoid feedback. We typically use 20-30 wireless mics per show and average 3-5 open mics per scene ... though large scenes might have 12-15 open mics.

[@jameswatson5576]

Hey James, unrelated question but what kind of wireless mic systems have you had success with? We need a reliable and durable pack and headset mic setup for speakers here in our overseas ministries. Thanks!

[@jamespwalters]

@@jameswatson5576 We use a combination of Shure and Sennheiser for the receiver and body pack. We have both old and newer equipment that we currently use. For example our current show is using the following models: 8 Shure ULXD, 16 Sennheiser G4/G3 and 6 Shure BLX. For the headset mic again we use a range of brands: DPA, Point Source, Galaxy Audio and Thor Hammer. We prefer the double ear headset mics over the single ear.

[@justinszeto9401]

​@jameswatson5576 a number of jts wireless headset was used in our Christmas play in church last year. Sounded great

[@alakovideo]

Could this be applied to an aux channel handling instruments + vocals? Saves individual or subgroup channel processing, leaves the FOH the same but fixes the main feedback issue which is onstage monitoring.

[@devinlsheets_alphasound]

Yes but right now it only works for vocals

[@ViciousButter]

This is similar to Clear by Supertone. It wasn’t made to fight feedback, but it uses AI to cut out background and reverb noise, which could have a similar effect since it breaks the amplification loop.

[@devinlsheets_alphasound]

We tried just about everything in our development phase. Not Clear specifically, but lots of AI denoisers, none of them work to fight feedback. Feedback is a fundamentally different kind of sound than "noise" or even room reverb. An AI has to be specially trained to fight feedback specifically for it to work.

[@lazymass]

Awesome use of AI, I like projects like this

[@christianschneider4033]

Awesome ... is there a Mac Version (Universal Binary / ARM) of it ?

[@devinlsheets_alphasound]

working on it, but it might be a while

[@hiimlawrence]

@@devinlsheets_alphasoundso this isn’t compatible on M1 Mac silicon chips?

[@Dolle_Man]

I want that speaker, show in the beginning, witch model/brand is that?

[@christianschneider4033]

Nexo P12

[@Rehbet]

Hope this can run on all the new NPUs they're putting in everything now

[@ilrosso666]

Interesting hope you manage to put everything in a compact unit with 6/8 ch

[@harleycurnow5242]

Can’t wait for waves to bring this out for $29

[@devinlsheets_alphasound]

We’ve been waiting for someone to do that for years lol

[@duanesheets5484]

If they could have they would have. They practically have a plug in to run your espresso machine.

[@scottbowman4844]

100% needs a Mac version. Looks interesting, but hope a Mac version is being worked on. For Intel and Apple Silicon.

[@Uforicmusic]

Agree

[@ES60Hz]

It will be nice to have it as a VST 3 for macOS + compatible with other audio interfaces.

[@DRAGDIESEL]

Ran this through the shop today, the guys were impressed...good job @alphasound ! has some great promise! Any sense if we buy version 1, would the releases coming to mac, soundgrid or any other variant require re-buying? We would love to start a spec with this for the volunteer HOW market for our builds.

[@devinlsheets_alphasound]

Anything we come out with for version 1 will be free to current users. This includes stereo, Mac OS, offline activation, etc. None of this is guaranteed to happen of course, we’re trying though!

[@softmoneyy]

question is when does it get acquired by waves? lol

[@devinlsheets_alphasound]

Waiting for the right offer

[@powderwang5233]

amazing but how about a new plugin that canceling the drum or guitar noise on the stage from vocal mkc，that willbe great

[@ViciousButter]

Drum cage + guitar amp modelers will do the same thing.

[@powderwang5233]

@ drum cage sometimes unaviliable in some venues😂 or on the post production stage noise get it tough

[@powderwang5233]

@ drum cage sometimes unaviliable in some venues😂 or on the post production stage noise get it tough

[@ChrisMutchJones]

Question, do you need an instance of each plugin on each vocal channel or can you insert it on the master bus?

[@duanesheets5484]

read the FAQ

[@danilorosich]

Omni headset in front of a stage monitor!

[@devinlsheets_alphasound]

Challenge accepted

[@DeusXMedia]

Actually impressed. 😶

[@ChimeTsnare-h7r]

Do you have a live autotune plugin too ? Combined with this would be awesome

[@GahGoes]

that's amazing! unfortunelly, expensive for Brazil people... the license price is about twice the minimum wage here =/

[@curly_]

I love it guys great graet idea, i barely run into feedback problems even in club situations , but realy would love to test it

[@Audio_Simon]

Was the example room used in this video part of the training data? I'm curious how well it can adapt to different environments, and I guess that is something you will also discover as it rolls out. Anyways congrats on what seems like a brilliant result from your hard work! IMO a de-reverb plugin would be great for small / home studios.

[@vilddyr]

Can we hear it with a band playing in the background please :)

[@fkmobile1]

Any plans to support the new M4 Mac Mini?

[@devinlsheets_alphasound]

Know anybody who can do that?

[@NathanCaudle-k8f]

Is this a plugin that you could add to Waves? Is that in the works at all?

[@TheRoss140]

Have you tried this with Dante ??

[@aleksandarstojceski3139]

Will it work on guitar feedback?

[@WaterTimeLapse]

Got Dante?

[@devinlsheets_alphasound]

The FAQ on the product site has all the answers

[@juniorsilvabroadcast]

You using windows? Did you license it for this?

[@devinlsheets_alphasound]

Right now the plugin only works on Windows, but we’re working on a Mac compatible version. You can buy a license to use the plugin on our website.

[@elcohado2835]

Wir brauchen ein Tutorial dazu. Was für ein Programm brauche ich noch zusätzlich? Die .vst3 datei lässt sich ja nicht einfach so öffnen.

[@jamesjudsontantepudi3871]

Hey I downloaded the plugin but when I trying to open or run it, it asks me to open with which application. What should I do.

[@disklamer]

Plus, we use the authentic medieval Captcha technology to annoy the bejesus out of our website visitors!

[@DJGEEGONZALES]

How do you connect the unit to the mixer?

[@duanesheets5484]

Read the FAQ

[@oliverheywardmusic]

Wow! Anyone tested this with Fourier Transform or Waves Livebox yet?

[@duanesheets5484]

Works great with Livebox.

[@duanesheets5484]

And with Fourier!!

[@TomSmith-hq1ok]

How will this work with a sound desk? Am I going to have to run an interface alongside?

[@devinlsheets_alphasound]

Yes, you can easily run analog lines to and from the interface

[@mixplacetube]

@@devinlsheets_alphasoundDude, do you expect that pro audio on tour will run mics through your interface for that?

[@devinlsheets_alphasound]

@@mixplacetubeif they want to get rid of feedback with super low latency and reliability, they will

[@timhoffmann5022]

If this could be implemented in QSC Cores (Linux Based) it would be huge for installs

[@duanesheets5484]

We use it in installs now and it works great.

[@arthurskudra50]

Q-SYS has notch feedback controller blocks, however most programmers don't bother to read the help file to properly implement them.

[@charlesl2762]

@@arthurskudra50 Q-Sys notch feedback controller is actually very destructive for audio, tone in particullary, right ?

[@arthurskudra50]

​@@charlesl2762 In saying that, it seems like you're trying to "straw man" that AI De-Feedback doesn't cause some compromise in the signal chain? Sorry, I've been in the industry long enough to know that there isn't any free lunch, whether or not AI is involved. It seems like AI is the new buzzword in our industry, and I've seen some clever implementations, but OK, perhaps the AI based De-Feedback plugin seems like a promising application. Please provide supporting technical documentation on your website if that is the case, not some cursory youtube video or scant webpage FAQ that suggests this is "magic". But to answer your insinuation... IF the QSYS notch feedback controller is properly setup (not the factory default settings BTW), it is very effective and less invasive than the slicing and dicing one typically does with the EQ strip on a mixing console provided that you have a properly aligned, balanced, and optimized loudspeaker system using the right measurement tools. Maybe the QSYS NFC is not the best one out there, but "good enough". I will say that I have yet to encounter any processor that comes even close to matching the Rupert Neve 5045 Primary Source Enhancer (or Yamaha Portico 5045 processor or plugin) for background noise and feedback suppression for signal transparency. It's used by many A-Level acts throughout the world. Perhaps your plugin might be something that might come close to it? I'd love to hear an AB comparison at the next ISE or Infocomm trade show. I'm always on the prowl for new and interesting solutions, and I do my best to cut past the marketing BS and keep an open mind. Kind regards, Arthur Skudra (industry veteran, sound system optimization expert, Rational Acoustics instructor)

[@duanesheets5484]

@@arthurskudra50 We've used the Neve PSE for many years on both live events and installations. While it helps considerably it has artifacts and you have to deal with thresholds. This has no thresholds and significantly fewer artifacts while giving you at least 10db more gain than the PSE. Try out the demo and compare for yourself.

[@hubionmac]

Nice, that seems to be a far less expensive alternative to CEDAR Audio DNS… Would be interesting to see if your hardware can handle Dante’s Virtual Soundcard in addition.

[@devinlsheets_alphasound]

The issue with Dante DVS is latency. We are working on a native Dante option for those recommended hardware devices but that will take us some time to develop. There are other products already in existence that run VST3 plug-ins with native Dante connectivity but they’re expensive.

[@jerryseigle7434]

Wish we had something like this that could tune the room without using a measurement mic to collect data to tune the room.

[@christiaanbezuidenhout6984]

I'm sorry, but $499 is insane. Only a few people will afford that. That's more than some DAW's cost. You'll make way more money, by dropping the price and selling tons of licenses, rather than a high price, selling only a few licenses. You should even consider what Qlab is doing: Rent To Own. You'll have way more people jumping onboard then. But otherwise, I'm happy to use Waves' new Feedback hunter to EQ wedges, even if it uses an EQ method which affects the tone, it gets the job done.. I'd love to invest in this, but that price would need to be a thenth of what it is, and also be Mac compatible. Sorry, just being honest.

[@devinlsheets_alphasound]

Points for being honest! Here’s the way we think about the price structure: let’s say you’re a church that struggles every week with feedback, how much would you be willing to pay every week to not have feedback? This plug-in costs $10 a week even if you only use it for a year. That’s $2 a week if you use it for five years. Even small churches spend multiple times this amount of money every week on donuts and coffee in the lobby, and that’s probably way less important than not having feedback during the entire sermon.

[@mixworks-de]

Agree. Plus, there are other AI neural network based VST/AU plugins already out there for less that might work just as well

[@duanesheets5484]

What are they? I’m really interested in knowing.

[@duanesheets5484]

Please let me know those plug ins you were talking about. I’d like to try them.

[@sergeantseven4240]

@devinlsheets_alphasound If churches are your primary market target, then yeah, but i feel like you are missing out on a much larger market pricing yourself for churches with bigger budgets. Lots of churches have spent good money for a sound engineer to design their system correctly, and they dont have an issue with feedback. Churches that didn't do that aren't going to throw down $1k plus for a single mic feedback preventer when they could get a rack mount one for cheap that does arguably well enough for a small church.

[@MG53v8]

How handy would this be❤

[@KAPOW6000]

how many inputs dose it have, cause if i was to buy this i would need at lease like 6

[@duanesheets5484]

Read the FAQ

[@tigeraudio]

What kind of speaker is that?

[@duanesheets5484]

Nexo P15

[@tigeraudio]

@@duanesheets5484 Oh interesting. THe newest Nexo PA I’ve used are S12, but I have been a Nexo user since 2000. PS15, PS10, Also the R2 versions, Alpha E, Alpha, T, S8, S12, and D. I was still using S8 regularly until early 2023, but I haven’t seen it since. D has the best rigging of any line array ever and I don’t know why the others haven’t copied it. The JBL A is pretty fast to rig, but still not something a single person can fly that fast, even with a single point.

[@rhythmixenterprises1720]

how to contact you for the plugin?

[@duanesheets5484]

Read the FAQ

[@vadimmartynyuk]

What if there are 18+ mics ?

[@devinlsheets_alphasound]

You could make a subgroup and put all the mics into it and send it through a plugin instance. But, if too many of those mics have people talking/singing at the same time, like a choir, the plugin might not work perfectly. We're working on the choir situation but it'll take time to get right. You could also group the mics into multiple subgroups and send each subgroup through their own instances, so that there is less chance of more that two or three mics with active vocal content running through any individual instance at the same time.

[@Artcore103]

This is cool, but can I integrate this into my church's setup with their behringer mixing board, and have it active for multiple channels... like how many channels? 4, 6, 8? I need this solution to be usable on every mic channel (but toggle on/off depending on the situation, because clearly it does affect the audio quality... no need to oversell it, the quality is remarkably good FOR the level of feedback removal it can do, it's worth the sacrifice).

[@duanesheets5484]

On the scope the effect is extremely small. Read the FAQ.

[@Artcore103]

@duanesheets5484 I will, but I could hear it in the video. That was a worst case scenario as far as feedback though .. maybe without feedback present the system can be active and not affect the sound as much? Anyway that's quite secondary, I want it regardless... Can this be implemented with an N-channel mixing board, with XLR in and out? Or digitally routed for killing feedback in all the audio passed through it at once (downmixed channels)?

[@duanesheets5484]

What you think you’re hearing is the ambient mic. Listen to the de reverb section for the sound of the plug in alone.

[@aleksandarkikivideos]

Good Job!

[@niccooss]

I have downloaded the plugin put it in the correct folder of my mac run a new instance of ableton but it doesn't show up on the list Did i have do something wrong ?

[@devinlsheets_alphasound]

If you read the FAQ section on the website, it specifies this is for windows only. We may eventually make this Mac compatible

[@curly_]

@@devinlsheets_alphasoundplease

[@alexanderniejaki9756]

Ok, this product demonstration has some wild and highly incorrect assumptions that I cannot leave uncorrected. First of all: I know that the folks at Alpha Sound are highly trained professionals. I watch this channel regularly, you know what you are talking about. I am also not attacking you personally in any way. You seem like a great guy and this is also why i am so disappointed by you and why I take the time to produce this response. Also, this is not a discussion about whether A.I. can or should be deployed in a live environment, I am strictly talking about the errors in this video. But let's face them in chronological order. 0:35 - You are showing a channel EQ of a DM3. But those settings make absolutely no sense when dealing with feedback. If an engineer makes cuts that wide to avoid feedback, his job should be replaced by AI. No one in their right mind would do this as we all know this causes a lot more issues (from unintelligibility, decreased volume to phase issues). If I need cuts that wide, I did something else horribly wrong, like speaker deployment or mic choices. 0:37 - "often requires hours of work". What the hell? If it takes me HOURS to ring out a system, again, AI should take my job. Any decent engineer will achieve this in minutes, if not seconds. Most of the corporate jobs I do (which is what this is made for I assume) don't even give me any time to ring out a system. Sometimes I have no more than ten seconds to ring out a monitor before the CEO or whoever is talking gets really mad at me. 1:22 - No, your plugin has not "zero latency". If your colleagues told you that is impossible, they were right. Any signal processing takes time. It might take very little time, but it still takes time. If I input a signal into any DSP, the program has to analyze it. How else will the AI know if there is feedback present? It then has to make an educated guess on how the signal would sound without feedback, process it that way, and output it. This takes time. I don't care how long it takes, even if it is less then a ms, it still takes time. Even sending an analog signal through a wire takes time. It happens at almost lightspeed, but even that is not instantaneous. If the processing was really "zero latency", it would mean that your AI can either 1) look into the future to process the signal in time before it even arrives or 2) is so fast, that the analyzing, processing and outputting of the signal takes less then one sample of time. With a standard sample rate of 48kHz, this would mean that all of the processing of the AI can take no longer than 20us per sample. I don't want to say this will never be possible but without further proof, I unfortunately have to call BS on this. Your FAQ is especially interesting when it comes to this point. I quote: "IS THE PLUGIN LINEAR PHASE? No, the plugin is not technically linear phase, although we tried our best to keep the phase shift to a minimum, and it shouldn't be noticeable under normal usage. Also, the small bit of phase shift is not always static, it randomly changes as the AI removes the feedback." - Please tell me how you achieve a phase shift without latency. 1:52 - You are showing the same bs as before. This is not how ringing out a system works. If I have feedback at 400Hz its not wrong to assume I might get feedback at 800Hz as well. But it would be wild to expect it at the same intensity though. If I have to cut -20dB on a fundamental and also -20dB on the octaves, I just show that I didn't understand how acoustics or signals work. Or my system deployment has some catastrophic flaws. 1:55 - This takes the cake for me. I have nothing to add to this, execpt again: If your EQ looks like this, there is either something going wrong elsewhere or you should be replaced by an AI. 2:06 - You're acting like a threshold is something bad. We both know this is quite a stupid thing to say. A threshold is not a limitation, it gives you extra control over when you want to apply a certain effect. Just imagine a compressor without a threshold. Pretty useless, right? If I cannot control when your plugin is engaging and when not, it has no use for me. 2:35 - Again. Either show how it works, or I continue calling BS. 2:43 - "But I don't give up that easily". Yeah! Don't let physics tell you what is possible and what not. 3:34 - This makes me sad. You as a systems engineer should know why your diagram is utter bs. No one in their right mind would deploy a system like this. You know how sound works in a reverberant room. You know that we use small collumn arrays or small point sources to tackle exactly this problem. I am not saying that this doesn't come with its own set of problems. But I am saying that your examples are highly misleading. So what is my conclusion? This product is not meant for any serious engineer. This is meant for venues that don't want to bother hiring real people and cannot wait for AI to mix their church services or whatever. I have a huge problem with that but am too biased to talk about this. You are a fantastic engineer and a great teacher but you are selling a product like its snakeoil. Why? You are creating a solution for a problem that is not nearly as big as you describe, which takes a lot away from your credibility. And this makes me sad. Because I know you can do better. So please.. Rethink how you market this.

[@duanesheets5484]

Your comment won't age well.

[@alexanderniejaki9756]

I wrote over 1.000 words as to why I feel like I do and this is all you have to say? Why won't my comment age well? ​@@duanesheets5484

[@duanesheets5484]

Give it time.

[@pindopurba]

I tested it in the studio, and it performed impressively. It successfully handled feedback from a Shure omni gooseneck mic, even with maximum gain from the Behringer X32, along with additional gain from the compressor and digital processing. It maintained performance effectively up to a certain volume threshold on the studio monitors. Next, I'll be testing it in my church, where the acoustics are more challenging due to hard surfaces like plain walls, granite, marble, glass, and hardwood. Regarding CPU usage, the plugin places a noticeable load on the system that runs the usual buffer size (256-512). On an AMD Ryzen 5 series processor, a single plugin consumed around 10% of the CPU. When tested on an Intel i9, this dropped to approximately 5%. Memory usage, however, remained relatively stable and didn't show any significant impact, as for latency it's not noticeable.

[@李垠-s8c]

Great thanks for your test. Another question, does it have impact on the sound quality compared to adjusting the EQ?

[@pindopurba]

From what I hear, there’s not much difference, and plugin seems to have a more subtle effect compared to EQ. That said, this is based on past experience with EQs rather than a head-to-head.

[@duanesheets5484]

@@pindopurbathere is no change in equalization except up near 20k.

[@devinlsheets_alphasound]

You should test the real round-trip latency of your setup by comparing it against a parallel mic input channel that isn't running through the plugin. To do this, create a second channel on your mixer with identical mic input, then reverse the phase on that channel and combine it with the channel that is inserted through the plugin. Bring up the delay on that second channel until you hear the total sound of the two channels cancel each other to silence. This will require sample-level precision, not just millisecond-level precision. Whatever that amount ends up being, that's your total round-trip latency for using the plugin in your situation.

[@pindopurba]

@@devinlsheets_alphasound Thanks for the pro tip! I’ve been testing the plugin live since my primary focus is on using it in a real-world setting, like at my church, where feedback is often a concern. When I tried it yesterday, I didn’t notice any latency issues, and I’m hoping for a similar experience when I test it in the church environment.

[@fkmobile1]

How many channels can one unit handle?

[@devinlsheets_alphasound]

The FAQ on the website has all the information

[@AurelShen]

Am sorry to not be that excited, but it sound a little bit arsh, like there are little glitches in the high frequencies. OFC it's amazing compares to an EQ on a headset mic and it will works for conferences but still

[@duanesheets5484]

The pinned comment says he’s using 20% ambient mic to show it’s not faked but in a real room. That’s what you’re hearing. Listen later to the de reverb section to hear the clean plug in without the ambient mic.

[@adazbig2552]

Thank you

[@russellhltn1396]

I take it this is for vocals only and doesn't work on other things - like an acoustic guitar?

[@devinlsheets_alphasound]

We’ve only verified that it works for vocals at this point. It was trained on vocal content. It might work on other stuff but we haven’t yet trained it for that.

[@russellhltn1396]

@@devinlsheets_alphasound It would be interesting to see how it handles unusual voices. (Cyndi Lauper, Janis Joplin, etc.)

[@Fiend404]

what are those speakers????

[@jensdroessler3575]

Nexo P12 I guess.

[@Kart0nas]

Very impressive

[@awesome6423]

Is this different to nvidia rtx voice or krisp ai?

[@duanesheets5484]

Those are noise canceling. Completely different application.

[@devinlsheets_alphasound]

Those have lots of latency, can't be used for live sound unfortunately. They also don't fight feedback very well, they're designed to reduce background noise which is a really different task.

[@awesome6423]

@@devinlsheets_alphasound Fair enough.

[@lucasetten]

Is this just some sort of fancy AI trained AEC?

[@devinlsheets_alphasound]

It’s fancy, whatever it is

[@joebarber3651]

Has anyone tried to use this on waves LV1

[@pinanaplayer]

🔥🔥🔥🔥🔥

[@Ranakade]

I honestly think this is a great way to use AI for live music.

[@alt.entertainments5487]

Wouldn't it make more sense to build some sort of Interface that uses the AI so You could plug into any audio input? or am I not understanding the product?

[@PooNinja]

Very interesting 🤨

[@AshLordCurry]

This is crazy

[@artifactingreality]

i already have an AI plugin to remove reverb so this is just that but more expensive I guess?

[@duanesheets5484]

Yeah you’re right.

[@artifactingreality]

​@@duanesheets5484 I use Waves Clarity Vx, I was guessing they must work in very similar ways if not the same way, however I just tried to use Clarity Vx to stop feedback and it didn't work but my demo just has one mic. Maybe there is some trickery in his demo with 2 mics for some reason.

[@duanesheets5484]

2 mics? What are you talking about?

[@artifactingreality]

​@@duanesheets5484read the pinned comment FAQ from the channel.

[@georgefisher5648]

But you still have ring at hi frequency😊

[@devinlsheets_alphasound]

That’s the sound of the room (a bunch of empty sheetrock walls) the plugin is preventing feedback even at high frequencies

[@duanesheets5484]

The pinned comment says he’s using 20% ambient mic to show it’s not faked but in a real room. That’s what you’re hearing. Listen later to the de reverb section to hear the clean plug in without the ambient mic.

[@TomSmith-hq1ok]

AT LONG LAST

[@ThePahan007]

[@RedButtonTV1]

nice

[@besterspieler2285]

Needing a dedicated modern x86 CPU to run this thing is absolutely insane and completely unviable.

[@duanesheets5484]

Then don’t buy it. It’s not for you.

[@Reactor10k]

So here’s how this will play out. You’ll sell some copies at $500 each. $99.99% of people will forget about yours, and another company will do something similar for $50. Everyone will buy that one.

[@duanesheets5484]

So why haven’t they done it already?

[@Reactor10k]

@@duanesheets5484That’s just a really stupid question on many, many levels.

[@duanesheets5484]

So why haven’t they?

[@duanesheets5484]

So with 8 billion people in the world we’d be overjoyed to only sell to .01% of them.

[@JakobN-zg1st]

You want know something that’s truly stupid on many levels? You thinking you know how this can play out. You’re not omniscient. Get offline. Please return when you have designed, fully implemented, and began selling your $50 version.

[@JohnnysaidWhat]

bro please do some demos in reverberated rooms before and after. would love to hear the before and after

[@TinoSchulz1990]

digital audio processing with ZERO latency does not exist. Every conversation (be it hardware-wise or software-wise) takes time to process. Especially when coupled with an "AI". What part of your plugin is even part of the "AI" ? Does it train itself to reduce even more feedback? Or is it just the process of listening to the incoming signal and processing it in various ways? Also don't forget that some latency can perfectly cut down feedback. This is a trick that's mostly widely known. I like the concept. But using "zero latency" and "AI" in one sentence is not possible.

[@duanesheets5484]

Lots of Waves plug in have zero latency.

[@TinoSchulz1990]

@@duanesheets5484 They report zero latency to the DAW. Right. But that doesn't make them zero latency ;)

[@duanesheets5484]

The FAQ specifies zero added latency.

[@nomadcreative1382]

0:38 “hours of work”? 🙄🙄🙄

[@TheMaddoxfam]

It’s AWESOME to see but… in a world of free and cheap VSTs including subscription options, this is ridiculously expensive @ $499. Make it $49-$99 but non-transferable, like Dante virtual sound card or waves super rack (which would be my integration method). I bet you’ll see sales skyrocket.

[@devinlsheets_alphasound]

It’s already selling pretty well, the way we thought about the price structure was from the standpoint of a church that fights feedback every week. How much would they be willing to pay every week to not have feedback during the sermon? At $500, if they use it for a year, that’s $10 a week. Five years… that’s $2 a week. Churches spend way more every week on coffee and donuts in the lobby and those are way less important.

[@duanesheets5484]

Free and cheap gets you free and cheap. You get what you pay for.

[@TheMaddoxfam]

@ ah yes, Waves C6 always lets me down

[@duanesheets5484]

This is many orders of magnitude more complex than a multi band compressor.

[@TheMaddoxfam]

@ I definitely realize this - My point was that the price tag is far too high. Good for alpha that they’re selling well, but at $500 I won’t be touching it. I’ll stick to sound gain staging, speaker, and microphone placement with some basic EQ for my applications

[@MrBravo143]

Is this possible to connect directly to a mixing console for live applications? We don’t need an interface in between. For example, mic to interface to mixing console.

[@tonibutigan]

What PA speaker is that?

[@duanesheets5484]

Nexo P15

[@prodantech]

Wish this was Linux based and not Windows.

[@bounded2426]

Claiming you're the first to do this isn't right. I've seen AI plugins that isolate and clean up vocals realtime already. It's a useful plugin, but definitely not anything unique.

[@duanesheets5484]

Which AI plug ins?

[@bounded2426]

@@duanesheets5484 Not sure what the names were, I use Ultimate Vocal Remover since I don't need to do it real time, it's free on github and does a great job isolating everything but vocals. I used a VST beta that did it realtime though, it was something along the lines of gogoi voice isolator or gogoi vocal remover, but I think they changed the name. Just look around there's plenty of options.

[@duanesheets5484]

Does it get rid of feedback? That’s primarily what this plug in is designed for. The room ambience is a nice byproduct.

[@duanesheets5484]

@@bounded2426 have you tried the demo?

[@juliodiazgranados2131]

😂😂😂😂 Pink noise.

[@duanesheets5484]

What about it?

[@cesarjeanchristophe339]

technology will make us lazy I agree that this is a great tool and that it makes our lives easier but for me I prefer the challenge because I like this profession for me this technology must be given to this who thinks they are a sound engineer

[@watchmandave]

when you say it "listens" define "listen" please...enquiring hearing aids of deaf musicians would like to know...peace! thanks :)

[@watchmandave]

setting up in a 103-year-old methodist church building, high metal ornate ceiling, windows, and a choral shell back wall on the stage...ricochet rabbit is an easy target... lol

[@duanesheets5484]

Obviously the AI doesn’t have ears. But it does analyze the audio content it’s given so in a way listens to it.

[@o0westwood0o]

Can you make this a waves plugin?

[@devinlsheets_alphasound]

Waves would have to contact us about that

[@TinoSchulz1990]

Waves already as a anti-feedback plugin.

[@devinlsheets_alphasound]

@@TinoSchulz1990 true. There's lots of anti-feedback plugins. We tried a ton of them before deciding to make a better one