Angsizing: Showing Our Work?

[Flashlight237]

Yeah, I'm starting another tired angsizing debate that'll just go nowhere. So here's the deal. Our current angular sizing methods account for only the height of a panel. This leads to stuff like this nonsense where we get something like 22 kilometers for something well within Earth in a cloudy sky: https://vsbattles.com/threads/issues-with-angsizing-and-manwha-panel.177186/

There was an attempt to account for field-of-view on the topic, but that amounted to that debate going nowhere: https://vsbattles.com/threads/angsizing-formula-2-0.171235/

While browsing through the Outskirts Battledome trying to work out a calc for Jafar, I looked for how angular sizing is handled over there and they way they did it accounts for both height AND width using the following:

2*atan(tan(70/2)*(panel width in pixels /panel height in pixels) ; measurements in degrees

Then this:

2*atan(Object size in pixels/(panel width /tan(x/2))) where x is the panel angle in degrees

Followed by using the trusty angsizing calculator: https://www.1728.org/angsize.htm

In the calc I did for Jafar, I determined the distance of the tower launched at a specific frame to be 1720.7 meters away from the camera: https://www.fanverse.org/threads/aladdin-calc-jafar-sends-a-tower-flying.1368147/

Using our formula, it's more like this:

29.45260805*900/[11*2*tan(70deg/2)]=1720.746142 meters

Which, granted I got the same results out of the math here, but at the same time, the movie is 1497x900 pixels wide. The Manhwa panel from the thread in the first link has a panel size of 800x4192. If I were to assume this particular sail circle from the manhwa panel is 10 meters...

Using our methods, we get something like this:

10*4192/[24*2*tan(70deg/2)]=1247.24925922 meters

Whereas using the OBD's method gave me this:

2*atan(tan(70/2)*(800/4192)=15.2223936°

2*atan(24/(800/tan(15.2223936/2)))=0.459374677°

Angsizing Calculator answer: 1247.249260 meters.

So uhh... What should we be doing here? Should we keep using our very simplified angular sizing formula without an explanation as to how it worked? Or should we use the OBD's angular sizing method as an explanation as to how we came to the conclusion that our formula worked practically identically?

 

[Breadbear83]

In a panel like the one shown in the "acceleration" section of this calc I made: https://vsbattles.fandom.com/wiki/U...spear,_Lifting_Strength#Lifting_Strenght_Calc

The panel for angsizing has a weird shape, smaller on the left and bigger on the right side, and the left side is higher than the right one

I was gonna ask if for panel height I'm supposed to rather choose the right part of the panel height (as it's bigger) or using the lower part of the right panel and the left part of the right panel to have a full panel height

But now on top of that, with you requesting using panel height and width, in cases like my calc should i use the later method, plus panel width?

Sorry if im derailling but I believe that this should be clarified for weird panel shapes in angsizing

 

[KLOL506]

@DontTalkDT

 

[Floxy178]

both height AND width

I don't think it does. You'll get panel width/panel width there which won't change anything, formula is identical to ours.