Core music-making skills dont have a shelf life if you could write, record and produce a great track in the past, those abilities are just as valid now as they ever were. Much like riding the proverbial bike.
Those skills might still need a little updating though. Music technology both in terms of how we make it, and how we listen is changing all the time, and its always worth keeping up with the latest developments.
When we talk about intelligent or smart plugins, were generally using these as catch-all terms for a variety of technologies involving some combination of forensic audio analysis, artificial intelligence and machine learning.
These are all terms that crop up in plugin promotional materials with increasing regularity. Their proliferation is down to more than fashion though these are technologies that make use of the increasing processing power of modern computers. They are processes that simply wouldnt have been feasible on home computers in the past.
Broadly speaking, these kinds of plugins are ones that can respond to context. Take an EQ as an example. Although a traditional EQ will let users boost or cut a variety of frequencies and may come stocked with presets aimed at specific types of material, the effect itself pays no heed to the material coming into the input it simply applies a process as instructed.
An intelligent EQ, on the other hand, will listen to the incoming audio and adapt its processes based on what it thinks necessary, such as boosting the top end on a vocal or adding punch to a kick.
These technologies all come down to training algorithms to find patterns. Machine learning involves, essentially, educating a plugin to recognise things so that it can make decisions on how to respond.
At their best, AI-powered plugins tend to support the creative decisions of the music maker, rather than attempt to overrule them
Take a noise reduction tool like those in iZotopes RX, for example. Its ability to remove unwanted noise is based on the developers training the algorithm to differentiate between desirable and undesirable elements of an audio file. Armed with this knowledge it can break the file down into very small elements and make decisions on what to keep and what to remove.
To some, these kinds of technologies hold a lot of negative connotations. Its easy to write off smart plugins as automated music-making tools ones that take the human skill out of music production. Probably the most controversial is LANDR, the AI-based online mastering service that claims to be able to provide a service traditionally seen as a dark art requiring years of training and a studio full of vintage gear.
At their best, AI-powered plugins tend to support the creative decisions of the music maker, rather than attempt to overrule them. In the case of mixing tools, the tech is best thought of as a modern take on the usual stock presets found in many plugins. As you might turn to a kick insert preset in a compressor plugin as a starting point, a smart plugins AI-powered compressor settings for your kick will likely provide a rough, convenient approximation of whats needed that the user can adjust to taste.
iZotope are relatively old hands when it comes to making use of AI and machine learning. You can find variations on these ideas present in Ozones Mastering Assistant or the smart features of Neoverb. RX is probably the companys most powerful deployment of modern tech, though.
This audio restoration tool uses machine learning to offer smart suggestions on how to clean up and improve audio. It works remarkably well too. Its thanks to this tech that RX keeps getting more and more effective, making it a must for studio, TV and other audio editing applications.
Focusrites recently-launched range of plugins are powered by tech from AI-specialists Sonible, whose Smart series of mixing tools could warrant inclusion in this list too. The reason weve opted for Focusrites take, however, is that these tools are an excellent example of how despite sounding complex intelligent features can actually help to make mixing tools far simpler and more user-friendly.
The four plugins in the range, focusing on EQ, compression, reverb and spectral ducking, each use their AI tools to help apply context-specific presets that can gently steer even total beginners toward the right settings for the job.
With Playbeat, Audiomodern apply an algorithm-driven approach to drum sequencing. On the surface, the plugin functions like a lot of other software drum machines. Users can import samples and sequence them across an eight-track step sequencer. There are parameters for varying pitch, level, flams, pan position, etc.
Where Playbeat is unique, however, is in its algorithms that can be used to randomise or remix a pattern based on current groove. This is where the intelligent element of design lies rather than simply randomise hits and parameters, it makes creative decisions inspired by the existing pattern.
RipX is divided into two versions DeepRemix and DeepAudio. The former of these promises the ability to split a fully mixed and mastered audio file down into individual instrument stems. Does it always work perfectly? No. But its still remarkably impressive, particularly when it comes to working with simple band material (classic rock and pop tracks, for example).
Its functionality relies on machine learning to intelligently identify different elements within an audio file. The pricier DeepRemix does the same, but adds some impressive editing tools too. We particularly like the ability to isolate and edit noise elements.
Atlas uses AI to take the pain out of wading through your unwieldy drum sample library. It works by analysing your sample folders and creating an elegant map coloured nodes that create clusters of samples grouped by sound-type and tone.
Even if you import a disorganised bundle of one-shots, melodic sounds and loops, Atlas will filter the non-percussive sounds to leave neatly-organised maps of beat hits. This functionality is paired with a slick, pad-focused drum machine and step sequencer. Theres a good level of onboard sample editing too, and those smart features make it easy to swap sounds on-the-fly.
TAIP is an interesting marriage between modern tech and vintage effects. The concept itself is nothing new a tape emulation that aims to impart the desirable inconsistencies of analogue tape machines. However, the approach taken is a little different.
According to Baby Audio, rather than use straight DSP, TAIP makes use of AI/neural networks to accurately decipher the sonic qualities that make a tape machine sound and behave in the way it does. This is done by feeding an algorithm with examples of dry and tape-processed audio in order for it to learn the precise differences between the analogue and digital versions of the audio.
View original post here:
6 AI-powered intelligent plugins that could change the way you make music - MusicRadar