One-bit representations of the audio signal used as part of the conversion process contain the same information as multibit ones and so can be processed directly. This approach offers significant system advantages and comes closer to the phase response of analog. The paper discusses this approach, in particular, looking at the problems of overflow, rounding, coefficient accuracy, and implementation. Solutions to these problems are also presented.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.