Is Dreamtonics Vocoflex a must-have vocal plugin or just an AI toy?$199 (excluding tax), dreamtonics.com
Available in a bundle with Synthesizer V Studio
It’s getting tough to tell whether an online voice is real or fake, isn’t it? Our era of vocal deepfakes and generative AI vocals is eerie and precarious — but that hasn’t stopped some developers from trying to improve the prowess of music producers with new AI vocal tools. Though we’ve seen software like Voice-Swap capable of transforming and replacing vocals in any song, Japanese brand Dreamtonics has gone a step further with its Vocoflex plugin, with seriously impressive results. But is it actually useful?

READ MORE: DJ Fresh doesn’t want AI to fall in the wrong hands — so he started Voice-Swap

Vocoflex uses AI technology to transform your vocals in real-time and morph between several voices simultaneously. The output may not sound as pristine as something like SoundID VoiceAI or Voice-Swap but its real-time workflow makes it fast and flexible.
Available as a plugin and standalone software, Vocoflex can also be used alongside another Dreamtonics instrument, Synthesizer V, to greatly expand the range of possible voices.
Aside from experimental vocal morphing, Vocoflex can change the voice’s gender, blend multiple targets to create an entirely new voice, create new backing vocals that differ from the lead, or create vocal demos with a more appropriate vocal sound. You can even try it on instruments instead of vocals, which can lead to surprising sound design results. If using it live, then you can switch on a low-latency Realtime mode, with the fastest option being around 35ms. Most will want to leave it off, for a higher-fidelity output.

How do you use Vocoflex?
You can choose your target voice in three ways. The first, and easiest, is to load up one of the 40 presets that include male and female pop singers, operatic soprano, husky R’n’B, mellow baritone and more. Although they’re high-quality models, there’s a learning curve to get to know the sound of each. You can generally guess if it’s male or female based on the name (e.g. Tenor), but we’d find it more useful if the list was split into low, medium and high registers so you can pick one that best fits your vocal.
The second way to create a target voice is with the Voice Generator, which presents an X/Y pad with gender on the X axis and Tone on the Y axis. You can click Random to generate a new voice timbre, which is linked to the colour of the pad. The new colour is also listed as a hex colour code at the bottom, which you can save and recall later. Once you’re happy, click Apply to generate the target. Easy, right?
Vocoflex interface
The final, and most impressive option, is to load in your own vocal sample. You only need around 10 seconds of audio for the process to work, which is far less than some competing products that require 15 minutes of analysis. Dreamtonics suggests that quality is more important than quantity; you’ll get the best results with clean and dry, solo vocal stems.
Once you’ve chosen a target, the presets and imported voices are shown as curves on the user interface, with several dots representing small portions of the analysed audio, and the Voice Generator showing up as a single node. If you choose more than one node, then the curves or dots are distributed around the screen, with similar voices placed nearer to each other to make morphing smoother. You can then play your vocal, either via a track in the DAW or through a microphone, and Vocoflex will transform it in real-time. It’s truly impressive how smoothly it all works.

To get the best vocal sound, you’ll want a target and main vocal that are as close in register as possible. If you click on the tuning fork icon, you can transpose the output by 18 semitones, and you get a helpful graphic that suggests the most suitable range.
As the voice is synthesised, it offers a novel, alternative way to transpose vocals. To put this to the test, we take our own vocal track and feed a portion into Vocoflex. We then use the generated target on the same vocal part (i.e. on itself) and pitch it down five semitones. If you were to do this using more traditional pitch-shifting effects, and even with something like Melodyne, then you get shifts in formants that alter the voice away from the original. However, amazingly, Vocoflex sounds just like the original singer, just in a lower register, albeit with subtle artefacts.
Vocoflex’s Waypoint system lets you select a number of points for the cursor that can then be jumped to either by hitting a key or via MIDI input. This means you can morph vocals live onstage, where different sections of a song are sung in different voices. Unfortunately, you can’t slide between points, but cursor positions can be automated.
Vocoflex voice generator
When tested against other vocal-changing products such as Sonarworks’ SoundID VoiceAI and Audimee, Vocoflex falls short in terms of audio quality and how natural the voice sounds – much like the difference between real-time and offline stem separation algorithms. At times, it sounds indistinct and is more likely to have subtle audio artefacts if any portion of the original vocal isn’t perfect. But remember — Vocoflex can be used more like an instrument than the VoiceAI and Audimee, so you’re getting a slightly unique product here.
For example, when using SoundID, you need to undertake many more steps to process and re-import the audio — and you’ll probably come up against some trial and error, as you don’t necessarily know what’s going to come out when you hit ‘process’. Vocoflex, on the other hand, is instant. Simply move the cursor and hear the changes in real-time so you can fine tune the output to your liking. It’s also lightning-quick to build your own voice presets.
With that in mind, we would probably only use Vocoflex for a lead vocal if the original recording and performance are of immaculate quality. However, it’s a perfect tool for creating fast mock-ups, or adding harmonies and backing vocals.
Dreamtonics has found a novel, if frustrating, way to help make Vocoflex ethical, given the potential for misuse or copyright infringement. When you register, you have to upload a copy of your ID. This is then used to generate an inaudible watermark in audio you create with the software, which can be traced back to you. It remains to be seen how effective this is, but it’s a step in the right direction for protecting artist rights.
Vocoflex waypoints
You might feel uncomfortable with uploading a copy of your passport or driving licence to access music software, which is totally fair. With Dreamtonics, it’s all handled via a third-party in London (ComplyCube) which specialises in identity verification and uses bank-grade security to encrypt your data. We’ve already seen this as a dealbreaker for potential buyers online, while others aren’t as concerned.
Vocoflex uses incredible technology to make vocal transformation fast and fun. If you’re working with imperfect audio, then you’re more likely to get the odd artefact or unnatural sound, but experimentation with different targets can improve the results. It would improve flexibility if you could go in and alter the generated targets further.
For $199, it’s definitely on the pricey side. However, compared to a tool such as Audimee, which is $25 per month for the Pro tier, you could save money in the long run. Although, as mentioned, the results with Vocoflex might not be as convincing. Voice-Swap, meanwhile, is $10 for the Pro tier. Both are slightly different prospects to Vocoflex, but will be more affordable if you’re just looking to change the timbre of a vocal part.
Overall, the streamlined approach and fast workflow of Vocoflex make this a worthy contender to more cumbersome online voice changing options, particularly for live performers.
Key features

Real-time vocal transformation and morphing, fine-tuned for singing
Plugin (VST3, AU, AAX) and standalone
40 target vocal presets
Import your own vocal a capellas for analysis
Morph between presets by dragging a cursor around the interface
Change genders or blend and create new vocalists
Set Waypoints to perform vocal switches via key presses or MIDI
Vocal Generator can generate a new vocal timbre at random
Edit the gender and tone of the Vocal Generator on an X/Y pad
Low latency Realtime mode
Watermarking feature to help reduce unethical use

The post Is Dreamtonics Vocoflex a must-have vocal plugin or just an AI toy? appeared first on MusicTech.