Why in the procedure oc_enc_block_transform_quantize() is used function oc_idct8x8()? After all, Inverse DCT only need to decoding stage, rather than coding?
> Why in the procedure oc_enc_block_transform_quantize() is used function oc_idct8x8()? > After all, Inverse DCT only need to decoding stage, rather than coding?Future frames are predicted from past frames. It's necessary for the encoder to know what the decoder will have decoded so that it knows what the decoder will be using for its prediction.