Techivation Releases AI-Loudener: Smart Loudness Enhancer

[Techivation]

Hello everyone!

We're thrilled to introduce our latest and the first from our AI series of plugins, set to transform your audio mixing and mastering experience: Techivation AI-Loudener.

Designed with precision and simplicity in mind, this AI-powered plugin is the key to achieving the perfect loudness in your mixes with unparalleled ease.
 
When developing the DSP algorithms for our plugins, we often encounter features or parameters that perform exceptionally well in certain situations, but not in others. Usually, features like these don’t make it into the final plugin, as it needs to function effectively across all material. This is why we turned our attention to machine learning. If the plugin can precisely identify what it is processing, then we can utilize even the most specialized features of our DSP algorithms. The AI series of plugins is designed to analyse the input using a system of convolutional neural networks and configure the enhanced version of an M-series algorithm to uniquely suit your track.


Watch the Teaser



Watch the Overview Video




Pricing

AI-Loudener is now available at techivation.com for an introductory offer of 45.00 USD; the original price will be 90.00 USD after April 30, 2024.

Crossgrade & Upgrade Offers: Owners of other premium Techivation plug-ins can take advantage of a special loyalty offer at 35.00 USD. Those who own all other Techivation plug-ins can acquire it for only 25.00 USD. Techivation M-Loudener users can also crossgrade to AI-Loudener at 25.00 USD.

*Techivation Full Access Users get this plug-in for free.


Free Trial

There’s a fully functional 14-day free trial available on our website, offering users to test the plug-in before committing a purchase.


For full product details and specifications, visit:
https://techivation.com/ai-loudener/

 
We’re always happy to hear your feedback!


Thank you!

Amin
Founder & CEO.

[bmanic]

Why on earth is the install size so ridiculously large? It needs 780mb of space to install.. wtf?

[Techivation]

Why on earth is the install size so ridiculously large? It needs 780mb of space to install.. wtf?

It's because of AI.

[Vortifex]

That would represent as much code as an entire operating system, which I doubt this plugin has. What's all the space actually for?

[bmanic]

Ok, so I gave the demo a quick go and the results are a bit of a mixed bag. Ironically so far I've got the best results analyzing completely different tracks and applying them to my original intended source.

Would be a much more useful plugin in my opinion if the user had some amount of control over what the AI should be aiming for. For instance, the "silky" processing or whatever causes the top end to not be too harsh was something that sounded much better on the audio I tried it on but the AI refused to recognize that.. so I went through a bunch of other audio clips until I got one where it wanted to apply silky.. then just ran it through the original audio source as intended.

The idea behind the plugin seems really interesting but I'm not immediately satisfied with the results. I'll test it on more sources of course before jumping to too many conclusions.

[bmanic]

That would represent as much code as an entire operating system, which I doubt this plugin has. What's all the space actually for?

The installer forces the install of all plugin types, which is unfortunate. That probably makes it quite a bit larger than it has to be. It installs AU, VST2, VST3 and AAX.. each about 200mb in size, which is quite large for a plugin but not unprecedented size.

[Vortifex]

The installer forces the install of all plugin types, which is unfortunate. That probably makes it quite a bit larger than it has to be. It installs AU, VST2, VST3 and AAX.. each about 200mb in size, which is quite large for a plugin but not unprecedented size.

I see, fair enough.

[Techivation]

That would represent as much code as an entire operating system, which I doubt this plugin has. What's all the space actually for?

The majority of the file size comes from the embedded machine-learning model. It's not uncommon for a model to take several gigabytes, so our models are quite optimised. This is the drawback of using machine learning directly on a device, without having to contact some server.

[motomotomoto]

Would be a much more useful plugin in my opinion if the user had some amount of control over what the AI should be aiming for. For instance, the "silky" processing or whatever causes the top end to not be too harsh was something that sounded much better on the audio I tried it on but the AI refused to recognize that.. so I went through a bunch of other audio clips until I got one where it wanted to apply silky.. then just ran it through the original audio source as intended.

You might just want to try the original M-Loudener plugin which is quite nice imo and has all sorts of control over the end result. From what the dev said earlier I am assuming this has the same funcitons, but the settings are determined by the AI.

[Techivation]

Would be a much more useful plugin in my opinion if the user had some amount of control over what the AI should be aiming for. For instance, the "silky" processing or whatever causes the top end to not be too harsh was something that sounded much better on the audio I tried it on but the AI refused to recognize that.. so I went through a bunch of other audio clips until I got one where it wanted to apply silky.. then just ran it through the original audio source as intended.

You might just want to try the original M-Loudener plugin which is quite nice imo and has all sorts of control over the end result. From what the dev said earlier I am assuming this has the same funcitons, but the settings are determined by the AI.

It does not have the same functionalities as the M-Loudener, because the AI-Loudener is loaded by several extra DSP algorithms under the hood with different settings, which change based on the AI's decision.

[motomotomoto]

Would be a much more useful plugin in my opinion if the user had some amount of control over what the AI should be aiming for. For instance, the "silky" processing or whatever causes the top end to not be too harsh was something that sounded much better on the audio I tried it on but the AI refused to recognize that.. so I went through a bunch of other audio clips until I got one where it wanted to apply silky.. then just ran it through the original audio source as intended.

You might just want to try the original M-Loudener plugin which is quite nice imo and has all sorts of control over the end result. From what the dev said earlier I am assuming this has the same funcitons, but the settings are determined by the AI.

It does not have the same functionalities as the M-Loudener, because the AI-Loudener is loaded by several extra DSP algorithms under the hood with different settings, which change based on the AI's decision.

That's interesting, seems like it would have made sense to combine them into one plugin, but excited to try it none the less

[pekbro]

Another loudener, imagine that.