>
>
>On Thu, 6 Mar 1997, Nick S Bensema wrote:
>
>> >- Do you think that a different way of storing X coordinates might lead to
>> >a more efficient code during screen drawing? For example:
>
>> Though it could be as simple as this:
>
><tech stuff deleted>
>
>Nick, what you were talking about was unnecessary, the values that would
>be stored in the byte are already adjusted after the calcpos routine was
>called. That routine takes an X coordinate and translates it into the
>correct rough and fine positioning values. Those can then be combined
>into a 1 byte value as the original poster stated. Please remember that
>the code I uploaded is VERY sloppy. I was just winging it, and plan a
>total rewrite now that the concept is down.
Your code was sloppy? Mine was WRONG... at least the LSR parts.
But one could save a lot of memory with minimal quality loss by simply
using the FC_XPOS (fine+coarse XPOS, for clarity) as the official
register for storing something's X position; it really wouldn't need
to be calculated further unless one wants to do the EOR #8 ahead of
time to save a few cycles in the drawing subroutine.
To convert a literal X position into FC_XPOS, and amke it accurate along
the entire screen.... that would take a divide-by-15 routine, which really
demands either a tedious INY-SBC-BCC loop, or a long network of CMP-BCC's.
But it would save time if one wanted to re-use a sprite while in the middle
of scanning a screen, certainly.
--
To unsubscribe, send the word UNSUBSCRIBE in the body of a message to
stella-request@xxxxxxxxxxx
