Systems and methods for enabling reduced bit-depth processing and memory reduction in video-related data processing

1. A method for dequantization and inverse transformation of an image signal, said method comprising:

receiving a plurality of quantized coefficient levels (c α ) that define an image;

receiving a quantization parameter (QP);

determining a periodic mantissa portion depending on QP % P, wherein P defines the period of said mantissa portion;

determining an exponential portion;

computing reconstructed transform coefficients ({tilde over (w)} α ) from said plurality of quantized coefficient levels (c α ), said periodic mantissa portion and said exponential portion;

performing an inverse transformation on said reconstructed transform coefficients ({tilde over (w)} α ) thereby yielding scaled samples ({tilde over (x)}′ α );

computing reconstructed samples, (x α ″) by normalizing the scaled samples.

2. The method of claim 1 wherein said computing reconstructed transform coefficients comprises using a formula equivalent to {tilde over (w)} α =c α ·R(QP)·b E(QP) , where R(QP) is a scaling factor mantissa portion and b is a base for a scaling factor exponential portion, E(QP).

3. A method for dequantization and inverse transformation of an image signal, said method comprising:

receiving a plurality of quantized coefficient levels (c) that define an image;

receiving a quantization parameter (QP);

determining a periodic mantissa portion depending on QP % P, wherein P defines the period of said mantissa portion;

determining an exponential portion;

computing reconstructed transform coefficients ({tilde over (w)} α ) wherein {tilde over (w)} α =c α ·R(QP)<<E(QP), where R(QP) is a scaling factor mantissa portion and E(QP)

is a scaling factor exponential portion;

performing an inverse transformation on said reconstructed transform coefficients ({tilde over (w)} α ) thereby yielding transformed samples ({tilde over (x)}′);

computing reconstructed samples, ({tilde over (x)}″ α ) wherein

x″ α =[{tilde over (x)}′ α +(1<<(M−1)]>>M, where M is normalization parameter for converting from said transformed samples, ({tilde over (x)}′) to said reconstructed samples, (x″ α ).