Software Review
AI-based Tool For Vocals
A few days ago I published a review of an interesting plugin from Karanyi Sounds, the Cloudmax Breezer. It was a software that promiesd to fix vocal recordings so they could sound a little more professional. At the same time, a software appeared from another place in Europe. From German company sonible came prime:vocal, which is a continuation of the theme of “simple software that fixes most things almost on its own” … with just a little help from AI (yes, there we have this buzzword again). Today we focus on sonible prime:vocal.
The great thing about sonible’s various software is that they can often, with very little effort, make a huge difference to the end result. Previously, we have looked at reverb, EQ plugins, compressors and limiters from the diligent developers at sonible. Prime:vocal, as you all probably already understood, handles various vocal recordings on a more general level.
sonible prime:vocal
The developers promise that this plug-in will allow you to create world-class vocal recordings, regardless of your recording environment.
They say AI has been used to accomplish this. The AI technology analyzes the recorded sound and removes things that should not be heard on a vocal recording – a bit like what was done on John Lennon’s old demo recordings – and only beautiful vocals remain (yes, of course, that assumes that there is one from the beginning). Sonible prime:vocal also offers a number of other controls with which you can improve your recorded vocal performance.
The thing is that most people who have a project or home studio and want to record a song don’t have access to a perfectly adapted and tuned vocal booth. The result from a vocal recording is often slightly colored by the acoustic environment in which you find yourself. It can be disturbances from the outside in the form of hum and other noise, but even more often you can hear strange resonances from the room. It could be anything from a reverb effect that originates from standing waves in the room, but it could also be that the room itself amplifies certain frequencies and creates resonances in an undesirable way. All of these are of course quirks that you would rather not have, but once they have ended up on a vocal track, it can be difficult to get rid of them. Until now, says sonible…
Several audio tools
Sonible prime:vocal has various tools that are intended to, if not eliminate, then at least reduce the impact of these effects on your recorded vocals. After the built-in AI algorithms have analyzed your audio recording, the app comes up with a suggestion for how it could sound. But the developers at sonible are not satisfied with this. They also have no less than five different special tools at their disposal to further fine-tune the sound.
Noise Reduction, Room Reduction and Vocal Clean-up are designed to address different aspects of this acoustic problem. Noise Reduction is designed to remove additional unwanted acoustic phenomena such as background noise, hum and crosstalk, Room Reduction is designed to tone down the room’s influence on your recording (reverbs, resonances, standing waves, etc.), and Vocal Clean-up is designed to handle sharp and disruptive consonants such as S, P (as in puff sounds) and the like.
But it doesn’t stop there. A dial for the so-called spectral balance parameter has also been inserted. They have probably taken a lot of algorithms from the company’s own excellent smart EQ4. With this knob you can control the sound in a brighter or darker direction.
Last in the row of knobs we have dynamics which allows you to control a so-called “level riding” (where you ensure that the sound level is kept at a steady and decent level in three different steps) and compression which can also be set in three different basic levels – in addition to the knob turns of course.
So how does it work?
I had very high hopes for sonible prime:vocal – sonible has developed so many excellent pieces of software over the years – but unfortunately I was disappointed by some details of prime:vocal.
On paper, this sounds like an extremely capable and suitable plugin for both home studio geeks who are forced to record their vocals in a room that doesn’t sound quite optimal, and for post-production where a sound engineer may need to fix dialogue that was recorded outside in a messy industrial space, in an office, or similar. There can really be many potential sources of interference that need to be eliminated (I suppose you recogize this).
When it comes to simpler fixes prime:vocal works really well. A vocal recording that has a bit too much natural reverb can be made to sound almost “dead muffled”, and both compression and the spectral balance function are okay. But if we have a really difficult recording with, for example, a motor or a fan that can be heard in the background, or a recording with far too lively acoustics, it immediately becomes much more difficult for prime:vocal. If you try to remove the reverb effect with room reduction from a very lively recording, you either get some funny tail that lingers as a kind of artifact/side effect, or the vocal/speech can instead feel both muffled and cut short. It just doesn’t sound very good.
There are many plugins from other developers that do a better job here. I’m thinking of the Waves Clarity Vx series, where the Pro plugins in particular do a really good job, and Cedar VoiceEX, Hush audio’s Hush and accentize dxRevive, all of which sound a bit better to my ears.
It actually feels as if sonible’s machine learning hasn’t been allowed to listen to really difficult recordings. But you might not want to record dialogue (or vocals) in a noisy workshop to begin with, you might object. Yes, but sometimes you actually have no choice.
Another detail is that prime:vocal for some reason cannot handle recordings longer than 60 minutes. This can be a problem for both dialogue and live recordings.
Then I would like prime:vocal to be better integrated into different DAW programs. The choice to release it only as an ARA plug-in and standalone feels very strange. The developers admit that both the AU and VST3 versions with ARA support are still in beta, and there is no native real-time version yet, nor is there AAX for all Pro Tools users.
This means, among other things, that on newer Apple computers, you are forced to start Logic Pro in Rosetta mode, otherwise the program cannot handle plugins with ARA support. This is of course a big problem and something that definitely needs to be addressed by Apple. But now Rosetta mode is what needs to be used, and this means that other functions that need support from the new Apple-developed processors (M1, M2, and so on), such as the excellent Stem separation function, do not work at the same time. Plague or cholera…
When it comes to Steinberg Cubase and Nuendo, you have to open the prime:vocal Extension on a new audio track and then manage it in an Event Editor. A bit cumbersome and rather clumsy, in other words.
In Presonus’s DAW program Studio One, it works the same way (Event Editors), but both here, in Steinberg’s program and in the standalone version of prime:vocal, the analysis of the audio track takes its fair share of time. Other plugins with full ARA support have spoiled us by allowing the plugin to get to work immediately, but that is not the case with prime:vocal. In its stand-alone version, the time for importing and exporting an audio file between prime:vocal and the DAW you use is also added.
Altogether
I have used and also written about several different sonible plugins over the years. All of the ones I have tested in the past have really impressed, so I naturally had really high expectations for prime:vocal as well.
The developers themselves claim that with prime:vocal they present the world’s most advanced processor software for vocal recordings.
The program also includes some special settings for spoken sound or dialogue. It is therefore not surprising that people want to be able to use this plugin both to fix slightly below par vocal recordings and dialogue for film or TV that was recorded under poor acoustic conditions.
If you have vocal recordings where you have a decent sound to begin with, prime:vocal can definitely be of some help, but if it is about really bad acoustic recordings, it doesn’t work quite as good.
The lack of common native versions in AU, VST3 and AAX formats also makes the plugin a more questionable choice. It feels like the development team was forced to rush the release of this software for some reason. It’s a bit telling that the ARA version of both the AU and VST3 versions is still in beta. Sonible prome:vocal shouldn’t have been released in December last year. I’m afraid it’s simply not ready yet.
LINK
System requirements
- PC at least Windows 10 (64 bit)
- Mac MacOS 11+
- RAM at least16 GB
- GPU OpenGL Version 3.2+
- CPU Intel Core (i5), Apple Silicon
Software formats
- Standalone app, ARA VST3 (Beta), ARA AU (Beta)
Supported sampling frequencies
- 44.1 kHz to 192 kHz
Authorization
- Machine based or iLok (dongle or cloud)
Support for Apple Silicon
- Native
Developers sonible gmbh
Info www.sonible.com/primevocal
Price 179,00 €