I noticed that in video_calc_gammatable() you do
gamma[i][j] = (uae_u32)v;
where v is of type float.
Since the float-to-int conversion truncates, should you do this instead?
gamma[i][j] = (uae_u32)(v + 0.5f);
Edit to add: Similar issue in limit256() and limit256rb()? The limit256rb() case may be slightly different since there v can be negative.
Last edited by mark_k; 10 February 2016 at 22:05.
|