![]() ![]() ![]() The tetrahedra each share three of their four oxygen ions with neighboring tetrahedra to produce a hexagonal sheet. In most micas, one in four silicon ions is replaced by an aluminium ion, while half the silicon ions are replaced by aluminium ions in brittle micas. The tetrahedral sheets consist of silica tetrahedra, which are silicon ions surrounded by four oxygen ions. It is the relatively weak ionic bonding between TOT layers that gives mica its perfect basal cleavage. The TOT layers in turn consist of two tetrahedral sheets ( T) strongly bonded to the two faces of a single octahedral sheet ( O). The crystal structure of mica is described as TOT-c, meaning that it is composed of parallel TOT layers weakly bonded to each other by cations ( c). Deposits of mica tend to have a flaky or platy appearance. Micas are translucent to opaque with a distinct vitreous or pearly luster, and different mica minerals display colors ranging from white to green or red to black. All crystallize in the monoclinic system, with a tendency towards pseudohexagonal crystals, and are similar in structure but vary in chemical composition. I have changed the mechanism a now, at all crossing point, we should have the 1min rolling avg target nits instead (if reachable) as I wanted initially.Ĭombination of a) and b) solved issues and the one I have looked at.Īnd as I said before, the "Chapter FALL merge" of 500% could be lowered and could be the value of the "Smart max Speed" of 5 (0,5%) if any other brightness change would be visible.īut let's wait for feedback before changing those.The mica group is composed of 37 phyllosilicate minerals. This meant that the crossing point value was not representative of the LOCAL situation. The new FALL algo helps a lot in that matter.ī) the crossing point of the "rolling average" within each chapter was NOT using (in most cases) a 1min rolling average value but the FULL chapter average instead (but it should have.) I looked at the issue with visible brightness change reported by but also at some other movies where I had spotted such behaviour myself.Ī) in all case, it was for low target nits changing between 50nits and 150+nits. What it does it set the target nits in such a way that the BT2390 Knee is equal to the current frame FALL (at least with 100% dynamic tuning).ģ) Improvement of the "Smart Max Speed" algo. It's a hidden algo hidden in the FALL algo if you select a "No compression limit" value of 1. If anybody prefer to use the good old FALL algo, then he can choose 150nits instead of the 0 for the no compression limit.īut I like the new FALL algo with BT2390 100nits Knee calculation better (and this also helps test scene)Ģ) A new Target nits experimental algo just for fun that we could call BT2390Knee Algo: In between the additionnal FALL part is scaled. If the FALL was high, then it would add up to those 150, and you would end up most of the time with a pretty higher target nits like ~200nits even if the peak itself was also only 200nits.Īnyway, now we only have Target= Peak Nits for any frames with a peak lower than 100nits.įor frames with a peak higher than 100nit, we calculate what the Bt2390 target nits needs to be to protect the 0-100nits from compression.Īnd we use the Min between 150 and the calculated BT2390TargetMin (target to protected 0-100nits) instead.Īlso the FALL is not added if BT2390TargetMin=100nits but is fully added by the time BT2390TargetMin reaches 150 nits (which is for about 250 peak nits as can be seen in the table below). Target Nits(i)= Math.Min(2 * arrHL(i), dblMinTarget + Math.Max(1, 2 * dblTuning / 50) * arrSKYFALL(i))which meant that any frame with a peak higher than 150 nits, had a target nits higher than 150 as well. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |