ReDither
As used in the QUADRIGA BatchProcessor
It is advisable to conclude any DOBBIN process with the ReDither FPU, as all internal processing is performed using 32 Bit Floating Point. The algorithm is the same as used in the Cube-Tec AudioCube platform.
Dithering is a method for reducing quantization errors in digital recordings. Often, dithering is applied when recording. In the case of WaveLab though, dithering is applied when reducing the number of bits in a recording, for example when changing from 20 to 16 bits and when applying processing.
The theory behind this is that during low level passages, only a few bits
are used to represent the signal, which leads to quantization errors
and hence distortion. To the ear, this is perceived as a "graininess"
during low level passages in a recording.
When "truncating bits",
as a result of moving from for example 20 to 16 bit resolution, such
quantization noise is added to an otherwise immaculate recording. By
adding a special kind of noise at an extremely low level, the
quantization errors are minimized. Indeed, the added noise can be
perceived as a very low-level quiescent hiss added to the recording.
However, this is hardly noticeable and much preferred to the distortion
that otherwise occurs.
When to use Dithering
When preparing 20-bit files for mastering to CD (converting to16 bits).
When preparing 16-bit recordings for multimedia (converting to 8 bits).
The Dither Options
Unfortunately, there are no hard and fast rules for the following options, which to select depends on the type of material. We recommend that you experiment and let your ears be the final judge. To set the parameters, click the white arrows repeatedly until the desired value appears.
Type 1: it is the most "all-round" type of method
Type 2: This method emphasizes higher frequencies than Type 1.
Noise shape (0,1 or 2)
Using this parameter one can change the noise that is inserted during dithering. There are also no generally valid rules here but you will that the noise will be moved further away from the middle area (which is the most audible area for the human ear) the higher the value is.
A special NoClippingControl is used that prevents your audio material from clipping.
