The balance between direct sound and reverberation is essential for the acoustic quality of audio recordings. DeVerberate is a reverb reduction plug-in that can attenuate or boost the reverberation that is already present in recorded material. The new version DeVerberate 3 includes an entirely new algorithm based on deep learning for fully automatic reverb reduction of recorded dialogue. By training a neural network on thousands of high-quality voice recordings and a wide variety of acoustical surroundings, the artificial intelligence can automatically separate dialogue from reverb.
The reverb reduction algorithm from version 2 is included for music and other audio sources. Both algorithms split the incoming audio into an estimation of the direct sound and the reverberated sound which can then be recombined freely. Additionally, a unique Early Reflections Filter that works in conjunction with the reverb suppression algorithms is included. The Early Reflections Filter uses sophisticated statistical methods to estimate the so called impulse response of the early reflections, which is then used to cancel out the effect of the early reflections.
- Attenuate reverberation in existing recordings
- Change the perceived distance in a mix by changing the balance between the direct sound and the reverberation
- Improve dialog and location sound
- Forensic specialists can improve speech intelligibility
- Boost the natural reverberation in existing recordings
The impulse response estimation from the Early Reflections Filter correctly detects this flutter echo.
- Fully automatic dialogue reverb reduction based on deep learning
- Manual mode for music and other signal types with adjustments of the original reverberation’s level and decay times
- Separate adjustments of the direct sound and the reverberation levels
- Early reflections filter
- Frequency spectrum representation of the following signals:
- Input signal
- Output signal
- Estimated reverberation
- Frequency emphasis filter for the the reduction level
- High and low shelving filters with variable slopes (-3 to -96 dB/ octave)
- High and low peak filters with adjustable bandwidth (0.1 to 3.0 octaves)
- Graphical representation of frequency response
- Editing of filter settings through handles in the freqency response curve
- EUCON and full Pro Tools HW controller support