Accentize is a young startup from Darmstadt, Germany that develops cutting edge audio processing software using machine learning methods. After several years of academic research, we realized that there is a huge gap between modern theoretical discoveries and productive implementations available.
Our products are designed to fill this gap by applying intelligent algorithms to general audio processing tasks. Most of our products currently available are focused on speech enhancement and restoration applications for post-production and content creation.
DeRoom is a machine learning based module for reducing or removing room reverberation and resonance in real time. The artificial neural network has been trained in many different room scenarios to be able to separate direct sound from reflective components. Reverb and echo tails are automatically detected and then precisely suppressed or, if desired, removed entirely. DeRoom is ideal for enhancing the quality and intelligibility of speech recorded in highly reverberant environments.
DeRoom Pro is a machine learning based reverb remover designed for professional use. An artificial neural network has been trained for many different room scenarios to be able to separate direct sound from reflective components. The real-time processing circuitry provides instant feedback on parameter changes and enables seamless integration into your workflow without the need to switch applications.
DialogueEnhance is an intelligent tool for automatic speech processing. You can use it to easily improve the quality of your speech recordings without having to adjust many parameters. The signal is processed by four different modules: Noise Reduction, Auto EQ, Dynamic Recovery, Volume Up. A useful tool for podcast and video editors!
PreFET is a machine learning based transistor preamp emulation that was developed using the same concept as PreTube. The included artificial neural network learned to accurately simulate the transistor stage of the preamplifier of a tabletop cassette recorder of the 70s. Depending on the settings, PreFET can color your sounds from subtle analogue saturation to heavy overdrive.
PreTube is a machine learning based tube preamp emulation that adds subtle analog richness to your instruments or vocals. The included artificial neural networks have learned to accurately simulate the behavior of real high-performance hardware circuits. Choose from three different amplifiers with different characteristics to find the most suitable sound for your project. PreTube is also a valuable addition to your mastering chain, adding analog depth and warmth.
SpectralBalance is a smart equalizer plugin for recording dialogues. By continually analyzing the audio signal, it automatically adapts its EQ curve to correct spectral imbalance in order to achieve pure, neutral speech. An efficient workflow makes it a very useful and time-saving tool for professionals who have to deal with many different dialogues with individual spectral characteristics. Automatic adaptation is also useful for live performances with multiple potentially moving speakers.
VoiceGate is a machine learning based module that can be used to reduce noise in speech and voice recordings in real time. The artificial neural network has been trained on over 100 hours of audio data to study the characteristics of human speech. The plugin can automatically adapt to many different types of noise and allows you to easily control the level of noise reduction you want.
• Chameleon v1.0.7
• DeRoom v1.1.9
• DeRoom Pro v1.1.1
• DialogueEnhance v1.1.9
• PreFET v1.0.1
• PreTube v1.1.1
• SpectralBalance v1.0.7
• VoiceGate v1.6.7
