Tweet
Auto Voice Filter
(Forensics Version Only)
The Auto Voice Filter is an adaptive system optimized for separating voice signals from random noise automatically. It uses two independent mathematical processes that work in harmony with one another in order to perform its job. One process identifies statistical noise while the other process identifies the human voice component of a signal. The Auto Voice Filter uses both of those pieces of information to separate the voice from the noise using sort of a “push-pull” methodology. The Auto Voice filter will continually adapt itself to varying signal to noise ratio conditions. It can improve the signal to noise ratio of spoken word recordings. It can also be used to reduce room reverberation (de-verb) sound on speech recording made in large venues. To operate this filter, merely adjust the “Attenuation” slider control for the optimal signal quality while previewing. The optimum quality will be found as the best balance between noise and/or reverberation reduction and speech intelligibility. Moving the attenuation slider upwards raises the aggressiveness of the filter and vice versa. Adjust the “Output Level” control such that the “Overload” LED does not light up on transient signals in order to minimize the creation of distortion. When you have discovered the best settings, highlight the entire sector of the file in need of filtering and then click “Run”. As with other Diamond Cut filters, you can save your favorite presets for future recall. This filter is limited to files using sample rates between 22.05 kHz to 48 kHz. If your file uses a different sampling rate, convert it to 44.1 kHz first before attempting to use the Auto Voice Filter. It works best with 16 bit resolution files. The Auto Voice Filter is monophonic only. Stereophonic files are summed to monophonic first and then processed. The resultant processed output file is presented in a dual-channel monophonic format.
The Auto Voice Filter
Note: This filter is optimized for Forensic Audio only; as such, it is not what one might call a “High Fidelity” system.
(Forensics Version Only)
The Auto Voice Filter is an adaptive system optimized for separating voice signals from random noise automatically. It uses two independent mathematical processes that work in harmony with one another in order to perform its job. One process identifies statistical noise while the other process identifies the human voice component of a signal. The Auto Voice Filter uses both of those pieces of information to separate the voice from the noise using sort of a “push-pull” methodology. The Auto Voice filter will continually adapt itself to varying signal to noise ratio conditions. It can improve the signal to noise ratio of spoken word recordings. It can also be used to reduce room reverberation (de-verb) sound on speech recording made in large venues. To operate this filter, merely adjust the “Attenuation” slider control for the optimal signal quality while previewing. The optimum quality will be found as the best balance between noise and/or reverberation reduction and speech intelligibility. Moving the attenuation slider upwards raises the aggressiveness of the filter and vice versa. Adjust the “Output Level” control such that the “Overload” LED does not light up on transient signals in order to minimize the creation of distortion. When you have discovered the best settings, highlight the entire sector of the file in need of filtering and then click “Run”. As with other Diamond Cut filters, you can save your favorite presets for future recall. This filter is limited to files using sample rates between 22.05 kHz to 48 kHz. If your file uses a different sampling rate, convert it to 44.1 kHz first before attempting to use the Auto Voice Filter. It works best with 16 bit resolution files. The Auto Voice Filter is monophonic only. Stereophonic files are summed to monophonic first and then processed. The resultant processed output file is presented in a dual-channel monophonic format.
The Auto Voice Filter
Note: This filter is optimized for Forensic Audio only; as such, it is not what one might call a “High Fidelity” system.