The "problem" is that the Quattro cameras has an image offset of 2000. That might be correct, or it might not. If it not is correct, we risk to clip the data at over exposure.
In practice, this means that the differential Huffman decoding starts at 2000, instead of 0 or a low value. If it had been 0 instead, nothing would have happened (except maybe avoiding clipping) as we deduct black level later in the chain.
No no no - it is not the Y2K bug.
The "problem" is that the Quattro cameras has an image offset of 2000. That might be correct, or it might not. If it not is correct, we risk to clip the data at over exposure.
In practice, this means that the differential Huffman decoding starts at 2000, instead of 0 or a low value. If it had been 0 instead, nothing would have happened (except maybe avoiding clipping) as we deduct black level later in the chain.