Crimes Against Curvature

[DMB_1]

The only thing brown I see is an entire Continent

That seems way too large to be North America though.

 

[ByAsura]

Just to be absolutely clear, this calc, based on the curvature of a spherical neutron star fragment, is ok, but this isn't?

 

[Dargoo_Faust]

The only thing brown I see is an entire Continent

Alright, let's do a demonstration:


Diameter of Earth is 7,917.5 miles.

7,917.5 * 511.3 / 766

Pixelscaled Distance from LA to Atlanta is 5,284.87 miles.

Actual Distance from LA to Atlanta is 2,339 miles.

Technically speaking I didn't account for the curvature of the Earth, so the pixelscaled difference should be somewhat larger.

 

[Andytrenom]

That's...a surprisingly smaller difference than I would have expected

 

[Dargoo_Faust]

The difference should technically be larger since it's a flat line projected onto a sphere, but yeah.

 

[DMUA]

Now I think I'm seeing it

 

[Ugarik]

The difference should technically be larger since it's a flat line projected onto a sphere, but yeah.

It's easy to calc though

Projection (chord) 5284.87 mi

Earth radius 3958.75 mi

Angle: 2asin(5284.87/2/3958.75) = 1.46 rads

True distance: 3958.75*1.46 = 5779.78 mi

 

[Ugarik]

Roughly 2.5 time larger that it actually is. That means KE feats would be about 97.66 times greater than they actually are

 

[Ugarik]

Btw the satellite which made this photo was 832 km above the Earth surface

This is apparently the minimal altitude you need to reach in order to see the entire planet

 

[Dargoo_Faust]

Thanks @Ugarik.

But yeah, there seems to be some visual forkery going on. Is there a name for this phenomenon, and any way we could take it into account for calcs where the POV isn't absurdly far away from earth?

 

[Ugarik]

Not really that's the same good old horizon view limitation. Here's the visualisation


It also works with any spherical object

 

[Dargoo_Faust]

I guess we could at least solve for the minimum distance needed to be away from the planet for the horizon that is viewed to be equal to the equator; after that I'm not entirely sure how to solve this issue unless we know exactly how high up the POV is.

Again thanks very much for the help, we'll need to update a lot of calcs to fix this.

 

[Mephistus]

I think you can approximate the true size of an object on a planets surface by angsizing the distance to the planet's horizon and working it out from there.

https://vsbattles.fandom.com/wiki/User_blog:Mephistus/Knuckle_Punch_feat

compare to what I think is wrong here:

https://vsbattles.fandom.com/wiki/User_blog:FanofRPGs/Knuckles_punches_The_Moo

 

[Ugarik]

Why? We can just use continents

 

[Dargoo_Faust]

Why? We can just use continents

Apologies, I mean when it's a fictional planetoid, which we wouldn't have defined sizes for any given landmass.

 

[Mephistus]

I kinda did that above; you need to know the diameter of the planet/moon/sphere tho.

 

[Antvasima]

Just a note that you can send me a message after you have reached a conclusion, if you would need my help with clearing something for application.

 

[Callsign_Castle]

I guess we could at least solve for the minimum distance needed to be away from the planet for the horizon that is viewed to be equal to the equator; after that I'm not entirely sure how to solve this issue unless we know exactly how high up the POV is.

Again thanks very much for the help, we'll need to update a lot of calcs to fix this.

I'm confused.

 

[Callsign_Castle]

Any suggestions for the rule's wording?

 

[DMUA]

actually, looking at it

Ugarik's image in general is a bit off, considering this image is much more proportional


So, one of these images has something going on

 

[AKM sama]

Both seem to be different images taken from a different distance/angle.

 

[Callsign_Castle]

I'm not on a computer to do scaling but I think using an image from live International Space Station recordings would be the most accurate.

I'm not sure if images or generated images would be made exactly to scale.

 

[Ugarik]

@DMUA

This was not my image. This was an actuall photo of the Earth called "Blue Marble 2012" taken by NASA. And the photo you have provided is called "Blue Marble 2002".

The reason why "Blue Marble 2002" looks much more proportional is because is was taken from 29,000 km above the planet's surface while "Blue Marble 2012" from only 2,100 km. I don't even understand your point

 

[Antvasima]

I was told to comment here again, but am not sure how I can be of help.

 

[DMUA]

The point is that using the planet's diameter might not necessarily be entirely void depending on circumstance.

 

[Antvasima]

Well, this is a calc group issue, so I will leave it to you to reach a decision.

DontTalkDT is good at modifying calculation instruction pages, if you need help with that later.

 

[Dargoo_Faust]

The point is that using the planet's diameter might not necessarily be entirely void depending on circumstance.

You'd need to be a certain distance away from the planet for the perspective to not be skewed by sphere geometry, which is kind of hard to prove in calcs where the distance away from Earth isn't given.

Even harder when it's a fictional planet which we can't reference country sizes for.

Ugarik has a point, which means a ton of our calculations are just sort of wrong.

 

[Callsign_Castle]

I think this should be tackled at two different stages.

Stage 1 would be correcting any calcs that use heavily exaggerated artistic choices or treating the horizon as planet curvature.

Stage 2 would be correcting calcs that use Earth shots from space.

Does this sound reasonable?

 

[Dargoo_Faust]

Sure. Although prior to both we should compile all calcs that do either, which is an undertaking in itself.

 

[Elizhaa]

Politely calling the calc group members, along with DontTalkDT and Antoniofer, here seems best.

I contacted DontTalkDT and Antoniofer yesterday and bumped my messages today. I hope they can be available.

 

[Elizhaa]

I think Callsign Castle's steps for the solution sound reasonable

 

[Callsign_Castle]

Sure. Although prior to both we should compile all calcs that do either, which is an undertaking in itself.

No time like the present, I'll get started. It would probably be quicker to start with the more popular verses.

 

[DontTalkDT]

Haven't read through the entire thread yet, but yes our earth curvature based calcs need some improvement.

I have the intuition that if one assumes angsizing assumptions then one could figure out formulas that account for the limitations of perspective without needing more information. At least for images where one sees a big portion of the planet.

I think that comes down to solving a fixpoint equation, which may or may not has a unique solution, but I don't have the time to get into the math right now.

 

[Callsign_Castle]

Calculations list (incomplete, some might need to be checked to see if they are actively used as a main calculation or source on profiles)

I don't think there will be that many.

https://vsbattles.fandom.com/wiki/User_blog:M3X/MoS_-_Flying_to_the_satellite

https://vsbattles.fandom.com/wiki/User_blog:TataHakai/Kusuo_Saiki_flies_to_America

https://vsbattles.fandom.com/wiki/User_blog:Darkanine/X-Men:_Storms_Storm

https://vsbattles.fandom.com/wiki/User_blog:Gwynbleiddd/Saitama_Splits_the_Atmosphere

https://vsbattles.fandom.com/wiki/User_blog:Golden_Void/Possible_Saitama_Casual_Feat (not used but still)

https://vsbattles.fandom.com/wiki/User_blog:Assaltwaffle/Pokemon:_Shadow_Mewtwo_Destroys_the_Earth

https://vsbattles.fandom.com/wiki/User_blog:ThePerpetual/More_Moon_Punching_By_Todd_Ingram

https://www.t5forums.com/forum/the-...ball-super-calc-frieza-destroys-planet-vegeta

https://vsbattles.fandom.com/wiki/User_blog:Kepekley23/Toeiverse:_Freeza_destroys_Planet_Vegeta

https://vsbattles.fandom.com/wiki/User_blog:Kepekley23/Naruto_and_Sasuke_seal_Kaguya_(recalc)

https://vsbattles.fandom.com/wiki/User_blog:M3X/Naruto:_Moving_the_moo

https://vsbattles.fandom.com/wiki/User_blog:Kepekley23/Naruto_Revision_Calcs#Feat_6

https://vsbattles.fandom.com/wiki/U...the_Discussion_Rules_(5-C_to_5-B_Toneri_Calc)

https://vsbattles.fandom.com/wiki/User_blog:DodoNova2/Chibaku_Tensei_Moon_Sized_-_Naruto

https://vsbattles.fandom.com/wiki/User_blog:Spinosaurus75DinosaurFan/Thanos_pulls_down_a_moon_(MCU)

 

[Elizhaa]

@DontTalkDT, no problem. I know the thread can finish soon but I honestly didn't mind waiting for your proper solution as I will be willing to make another Calculation Group Discussion thread for this topic proper conclusion.

 

[Kepekley23]

Alright, let's do a demonstration:


Diameter of Earth is 7,917.5 miles.

7,917.5 * 511.3 / 766

Pixelscaled Distance from LA to Atlanta is 5,284.87 miles.

Actual Distance from LA to Atlanta is 2,339 miles.

Technically speaking I didn't account for the curvature of the Earth, so the pixelscaled difference should be somewhat larger.

This is a camera-related problem though. The picture here is taken with a lens that strongly distorts the edges of the planet, thus creating an illusion of an enlarged America. It would have no bearing on, say, the entire planet being shown from afar in a manga panel.

 

[Kepekley23]

This suggestion is perfectly fine for calcs that go like "an energy wave explodes and some hints of curvature are shown". For calcs that show pretty much the whole planet and especially calcs that show said planet from afar, this is absolutely not fine.

 

[Dargoo_Faust]

This is a camera-related problem though. The picture here is taken with a lens that strongly distorts the edges of the planet, thus creating an illusion of an enlarged America. It would have no bearing on, say, the entire planet being shown from afar in a manga panel.

Where exactly is it stated this is a camera-related problem and not just a simple geometric problem?

Ugarik does well to explain that even if it seems the globe is fully visible, that can certainly not be the case.

Not really that's the same good old horizon view limitation. Here's the visualisation

It also works with any spherical object



Despite the second image being a full circle, the proximity of the viewpoint (the camera, in this case) causes the perimeter of the circle to not actually encompass a hemisphere but a spherical cap.

Unless, in "calcs that show pretty much the whole planet", the viewpoint is at a sufficient distance, the objects/explosion being viewed is actually much larger than it would actually be when you use stuff like the Earth's diameter for comparison.

 

[Kepekley23]

> Where exactly is it stated this is a camera-related problem and not just a simple geometric problem?

Those Earth pictures are taken with a fish-eye lens, which is well-known to "round" and distort objects at the edges. Not only that, those pictures aren't in real time, they are mashups/composites of several different shots over several months of footage. Each one taken over different distances, thus further complicating your point.

The perspective and distance between the Point of View and the planet are part of the problem, but that's solely in the cases where we use shos that are very close to the ground and are by all means are ******-up curvature (such as my aforementioned example of an energy blast detonating, the camera panning out over a few hundred kilometers, and it showing significant curvature), which, as I said on the very post you responded to, I agree shouldn't be used.

 

[Dargoo_Faust]

Those Earth pictures are taken with a fish-eye lens, which is well-known to "round" and distort objects at the edges. Not only that, those pictures aren't in real time, they are mashups/composites of several different shots over several months of footage.

Makes sense.

The perspective and distance between the Point of View and the planet are part of the problem, but that's solely in the cases where we use shos that are very close to the ground and by all means are ******-up curvature, which, as I said on the very post you responded to, I agree shouldn't be used.

Doesn't the example that Ugarik provided (the second, not the first), demonstrate that even when a full circular "globe" is visible there can still be issues?

If that's the problem, how exactly do we determine if we have the "whole" globe and aren't effected by what Ugarik called the "horizon view limitation"?

Obviously this doesn't need to be 'close to the ground' as going off of this you could be ~2000 kilometers off the ground and still not get the whole picture.

I get this not being an issue for deep space shots, although I'm worried about shots where the globe still takes up much of the view.