Change Angsizing Usage (THIS AFFECTS MORE THAN 1000 CALCULATIONS!!!!!)

[Flashlight237]

If that's the main issue, it could just be taken slowly. Encouraging people to remake old blogs, and not accepting new blogs with a bad basis.

Yeah, I wouldn't mind this route.

 

[Kulf_Boba]

Because providing experimental data that is relevant to the topic at hand and touching on Agnaa's concerns with the workings on camera functionality is "not contributing anything to the discussion"? This is a site-wide revision where any relevant data both current and previous, including my own, should be considered in EQUAL weight if it needs to be done as smoothly as possible. By deliberately claiming someone who has contributed to the discussion didn't contribute to the discussion, you're actively ignoring what was presented, thus stagnating possible avenues where an important revision can go. If you're not going to be civil about it and accept my contributions equally to others' contributions, then I don't know what to do with you. We're all staff members trying to figure out where this should go and how to handle it, and I do not appreciate you leaving me out like that. I provided evidence that I personally had to collect in order to contribute; please do me a kind favor and take it with due diligence.

It is you who should approach this more seriously and diligently. First, it sounds like you haven't even taken the time to read this OP, yet you brag about answering it.

I will go out of my way to say that this topic is one that I've personally answered to with a 4:3 camera: https://vsbattles.com/threads/angsizing-formula-2-0.171235/post-6700369

Then, watch how you present your arguments:

ChatGPT? Really? Okay, I would very much prefer to get back into the topic, but that is just... Wow.

It gives the impression that you are only seeking to contradict instead of taking the discussion seriously.

Furthermore, your experimentation does not seem interesting to me for this discussion, since the important thing here would be to focus on the cinematographic field.

 

[Kulf_Boba]

I would like to discuss what angle of view could be used for the horizontal FOV, taking into account the following information about the most commonly used lenses in cinema:

35 mm and 50 mm Lenses (Standard Lenses)

• 35 mm: This lens is considered slightly wide-angle, offering a broad field of view without distorting perspective too much. It is commonly used for indoor scenes or situations where a larger portion of the environment needs to be shown without making objects appear distorted.
• 50 mm: Considered the standard lens closest to what the human eye sees in terms of perspective. It does not produce distortion and is ideal for medium shots and general takes.

24 mm and 28 mm Lenses (Wide-Angle)

• 24 mm: Used for wide shots that capture more of the environment. Ideal for landscapes, action scenes, or small interiors where more space needs to be encompassed.
• 28 mm: This is a wide-angle lens that does not distort as much as more extreme lenses (like the 18 mm), making it suitable for shots where a broad view is needed without distortion.

Lens Focal Length	Diagonal	Vertical	Horizontal
24mm	84.1°	53.1°	73.7°
28mm	75.4°	46.4°	65.5°
35mm	63.4°	37.8°	54.4°
50mm	46.8°	27.0°	39.6°

Value table source:

https://www.nikonians.org/reviews/fov-tables

Angle of view (photography) - Wikipedia

en.wikipedia.org

 

[Mr. Bambu]

It is you who should approach this more seriously and diligently. First, it sounds like you haven't even taken the time to read this OP, yet you brag about answering it.




Then, watch how you present your arguments:


It gives the impression that you are only seeking to contradict instead of taking the discussion seriously.

Furthermore, your experimentation does not seem interesting to me for this discussion, since the important thing here would be to focus on the cinematographic field.

While I agree that the method and phrasing on Flashlight's behalf shouldn't be as it is, I would also suggest, for future formal discussions you'd like to have, to avoid citing Chat GPT. It simply isn't a reliable enough source to be taken seriously. As far as I'm aware, this was substantiating evidence, not your full evidence, that was hailed from GPT: as such, it's not a big deal, but it really is entirely irrelevant what it concocts to say.

 

[Kulf_Boba]

While I agree that the method and phrasing on Flashlight's behalf shouldn't be as it is, I would also suggest, for future formal discussions you'd like to have, to avoid citing Chat GPT. It simply isn't a reliable enough source to be taken seriously. As far as I'm aware, this was substantiating evidence, not your full evidence, that was hailed from GPT: as such, it's not a big deal, but it really is entirely irrelevant what it concocts to say.

The information I provided from ChatGPT is easily verifiable through simple searches, and I had previously corroborated it. I do not intend to use ChatGPT as a primary source. I cited the information it provided because I felt its explanation of the most commonly used lenses in cinema was suitable for summarizing what can be found online.

 

[Agnaa]

The information I provided from ChatGPT is easily verifiable through simple searches

Then use information from those simple searches instead.

and I had previously corroborated it.

Going through your posts from before that, I can't find that.

But if you did, then there was no reason to have ChatGPT reiterate it.

I do not intend to use ChatGPT as a primary source. I cited the information it provided because I felt its explanation of the most commonly used lenses in cinema was suitable for summarizing what can be found online.

We don't trust LLMs to summarize things without confabulations. Summarize it yourself, direct us towards the full explanations, or direct us to a summary done by another person.

 

[Kulf_Boba]

Then use information from those simple searches instead.

If necessary, I will do it.

I have already provided these sources before:

1. https://www.tutti.space/blog/quick-guide-to-cinematic-and-cine-lenses
2. https://artlist.io/blog/cinematic-f...UACHdhkxRO41RD6nWKVz1sX-tQ9xoALxoCJkcQAvD_BwE
3. https://www.studiobinder.com/blog/different-types-camera-lenses-explained/

Going through your posts from before that, I can't find that.

I said that I had corroborated them, not that I had initially published them.

But if you did, then there was no reason to have ChatGPT reiterate it.

I already explained the reason why I used the information provided by ChatGPT.

We don't trust LLMs to summarize things without confabulations. Summarize it yourself, direct us towards the full explanations, or direct us to a summary done by another person.

Well, I do trust in this case, because it is adequate information and I have already corroborated that it is correct.

 

[Agnaa]

You trust it, cool, but we don't; so I'd still suggest listening to myself and the three other users that have advised against using LLMs as part of your posts in the future.

 

[Flashlight237]

It is you who should approach this more seriously and diligently. First, it sounds like you haven't even taken the time to read this OP, yet you brag about answering it.

I wasn't bragging, I literally did answer it. The entire argument concerns different FOVs, and the link I provided is experimentation using an everyday camera (in that case, a Google Pixel 8, which is more close-at-home to everyday situations). That's another example of a different FOV, one that clearly is outside the film industry as most people's cameras are not the kind used to film the likes of Jaws, The Matrix, Independence Day, and even recent films like The Avengers: Endgame and Wakanda Forever. In fact, what isn't serious and diligent is your decision to cherry-pick a select few sentences out of several entire posts, which I have taken the entire discussion seriously and diligently up to this point despite my wording. My response to the ChatGPT thing is the point where my switch flipped to condescending mode since, like others have pointed out, is not something that should even be considered a reliable source, and I don't like how it is worded as a reliable source.

It gives the impression that you are only seeking to contradict instead of taking the discussion seriously.

Furthermore, your experimentation does not seem interesting to me for this discussion, since the important thing here would be to focus on the cinematographic field.

My dude, both supporting AND contradictory points are serious points in a debate. You can't bolster an argument by only choosing the supporting parts. And furthermore, most people aren't IN the cinematographic field; in fact, we literally had members in this thread even say that they aren't professional photographers, never mind filmographers. You can't expect them to have the same level of tech LucasFilms used.

Also, you literally brought up video games in the OP, something that ignores cinematography entirely as, as I've pointed out in one of my posts, video games rely on the capabilities of the game engine, something programmers would need to handle in order to get it to work. You don't film a game, you program a camera into the game. Look at GTA IV; you can see a clear example of the cameras being programmed by getting the character drunk and watching the camera wave around in response. This isn't including the more questionable stuff you see in video games like the Fuzzies from Yoshi's Island basically making the camera not see in Mode 7.

Also, outside of a live action film or a professional photographer's work, you can't expect everyday situations to make use of a professional-grade camera. At best, you can assume they were filmed with a Camcorder, maybe, and even then, an iPhone or an Android would be more at home with everyday situations camera-wise than a professional film-studio camera. I don't expect Garfield comic frames to have been taken using a professional-grade camera because of this logic. Even then, our decision to use 70° for the vertical FOV is entirely based on the FOV of a human eye looking dead-on at an object, not the work done for the MCU. You can't invalidate someone because they don't own or have not rented a film studio.

 

[Antvasima]

So, regarding the issue of finding the most reliable calculation method for this...

@DontTalkDT @Executor_N0 @Spinosaurus75DinosaurFan @Mr. Bambu @Therefir @Ugarik @DMUA @Damage3245 @TheRustyOne @DemonGodMitchAubin @Jasonsith @Wokistan @Armorchompy @Migue79 @Psychomaster35 @CloverDragon03 @Dark-Carioca @AbaddonTheDisappointment @Aguywhodoesthings @Agnaa @Dalesean027 @DemiiPowa @Flashlight237 @SeijiSetto @SunDaGamer

 

[Mr. Bambu]

As I've said, I am not a reliable source in determining the outcome of this thread, I merely believe that it is important to do so. I would rather it not be sidetracked by the discussion of AI involvement or tone policing, as the issue at hand is more important than either of those.

 

[Antvasima]

@Chariot190

You are also free to help out here if you wish.

 

[Kulf_Boba]

I wasn't bragging, I literally did answer it. The entire argument concerns different FOVs, and the link I provided is experimentation using an everyday camera (in that case, a Google Pixel 8, which is more close-at-home to everyday situations). That's another example of a different FOV, one that clearly is outside the film industry as most people's cameras are not the kind used to film the likes of Jaws, The Matrix, Independence Day, and even recent films like The Avengers: Endgame and Wakanda Forever. In fact, what isn't serious and diligent is your decision to cherry-pick a select few sentences out of several entire posts, which I have taken the entire discussion seriously and diligently up to this point despite my wording. My response to the ChatGPT thing is the point where my switch flipped to condescending mode since, like others have pointed out, is not something that should even be considered a reliable source, and I don't like how it is worded as a reliable source.

That was a response to a previous proposal. When I presented the new proposal in that thread, you didn't make any concrete comments regarding it. Then you introduced your experiment in this new context as if it were a response to my proposal, but you didn't explain how it impacts or affects the ideas presented.

1. The sizing calculations should use the panel length (not the panel height) in scans of movies and series, preferably at 70°, as it falls within the range of the most common viewing angles in cinema.
2. In video games, the standard for the horizontal field of view (FOV) is around 70°, so this value should also be used for the panel length unless an exact value is known.
3. For comic and manga panels, a smaller angle should be used, preferably 40° for the panel height, while panels similar to those in cinema should use 70° for the panel length.

Your experiment is not relevant to the topic of movies and series, as it would be more appropriate to investigate what lenses are commonly used in the film industry. It is also not relevant to video games. For comic and manga panels, it might be relevant; however, you didn’t take the time to properly argue your point. By the way, an extra note is that in your camera, the photos have a horizontal FOV of 70° because the vertical FOV is 55°.

My dude, both supporting AND contradictory points are serious points in a debate. You can't bolster an argument by only choosing the supporting parts. And furthermore, most people aren't IN the cinematographic field; in fact, we literally had members in this thread even say that they aren't professional photographers, never mind filmographers. You can't expect them to have the same level of tech LucasFilms used.

I don't consider your points to be contradictions; I simply think they are not relevant. If we are discussing the FOVs used in series and movies, I believe it would be more appropriate to provide information within the cinematographic context.

Also, you literally brought up video games in the OP, something that ignores cinematography entirely as, as I've pointed out in one of my posts, video games rely on the capabilities of the game engine, something programmers would need to handle in order to get it to work. You don't film a game, you program a camera into the game. Look at GTA IV; you can see a clear example of the cameras being programmed by getting the character drunk and watching the camera wave around in response. This isn't including the more questionable stuff you see in video games like the Fuzzies from Yoshi's Island basically making the camera not see in Mode 7.

I didn't mention anything about video games in relation to cinematography. What I mentioned was that the horizontal FOV in video games usually hovers around 70°, and that you should use that value as a standard or, alternatively, check the settings to see what type of FOV is being used. For video games, the discussion seems straightforward; at most, one could talk about the horizontal FOV, which should be used as a standard when an exact value is unknown.

Also, outside of a live action film or a professional photographer's work, you can't expect everyday situations to make use of a professional-grade camera. At best, you can assume they were filmed with a Camcorder, maybe, and even then, an iPhone or an Android would be more at home with everyday situations camera-wise than a professional film-studio camera.

This would be relevant if we were analyzing feats filmed with a mobile phone camera.

I don't expect Garfield comic frames to have been taken using a professional-grade camera because of this logic. Even then, our decision to use 70° for the vertical FOV is entirely based on the FOV of a human eye looking dead-on at an object, not the work done for the MCU. You can't invalidate someone because they don't own or have not rented a film studio.

Wasn't the value of 70° supposed to be based on the common FOV of modern cameras?

The idea is not to assume that comic panels were captured with professional-quality cameras, but rather to apply an appropriate FOV for the type of panel, using the FOVs employed in cinema and photography as a reference to maximize their utility. For example, a horizontal FOV of 70° is achieved with lenses considered wide-angle, which are used in cinema and photography to create panoramic shots, capture narrow interiors, or achieve a sense of depth and closeness in scenes. These lenses are perfect for landscapes, architecture, and situations where more elements need to be included in the frame; therefore, it doesn't make much sense to use them for the height of panels like these:

 

[TheRustyOne]

My opinion is pretty much the same as Bambu and Agnaa, I do not have the knowledge required to accurately suggest much here. Obviously, any angular size calc where the distance can be proven incorrect should be disregarded. However, our rules already cover stuff like this to begin with.

These are the summarized points of the OP, correct? I think I understand this, but I don't want to assume right now.

Maybe the OP should show their suggestions in action? Show why that would be more accurate than our current methods.

I imagine it'd be easier to understand your points if one could see them in a calculation.

 

[Kulf_Boba]

Maybe the OP should show their suggestions in action? Show why that would be more accurate than our current methods.

I imagine it'd be easier to understand your points if one could see them in a calculation.

I have already explained the reasons why my proposal is more accurate, but likewise, I will present several calculations below to allow for a more detailed comparison:

Movies and Series​

Black Adam: 88px | 1.95m
Panel Height: 804px
Panel Width: 1920px

Distance to Black Adam (70° Vertical FoV): 1.95 * 804/(88 * 2 * tan(70deg/2)) = 12.7m
Distance to Black Adam (70° Horizontal FoV): 1.95 * 1920/(88 * 2 * tan(70deg/2)) = 30.3m

Earth: 1411px | 12,742 km
Panel Height: 1080px

Distance to the center of the Earth (70° Vertical FoV): 12,742 * 1080/(1411 * 2 * tan(70deg/2)) = 6964.3 km
Distance to the Surface: 6964.3 - 6371 = 593.3 km (Subtracting Earth's radius)

Distance to the center of the Earth (70° Horizontal FoV): 12,742 * 1920/(1411 * 2 * tan(70deg/2)) = 12,380 km
Distance to the Surface: 12,380 - 6371 = 6009 km

Games​

I don't know Spider-Man's horizontal FoV, so I will assume a default of 70°.

Sandman: 32.57px | 24.93m
Panel Height: 720px
Panel Width: 1280px

Distance to Sandman (70° Vertical FoV): 24.93 * 720/(32.57 * 2 * tan(70deg/2)) = 393m

Distance to Sandman (70° Horizontal FoV): 24.93 * 1280/(32.57 * 2 * tan(70deg/2)) = 699.61m

The second method matches the distance indicated by the game where Sandman is after this panel, which is 640 meters away.

Comics​

Electro's Head: 24px | 0.225m
Panel Height: 369px

Distance to Electro (70°): 0.225 * 369/(24 * 2 * tan(70deg/2)) = 2.47m

Distance to Electro (40°): 0.225 * 369/(24 * 2 * tan(40deg/2)) = 4.75m

Sam Wilson: 430px | 1.88m
Panel Height: 900px
Panel Width: 1827px

Distance to Sam (70° Vertical FoV): 1.88 * 900/(430 * 2 * tan(70deg/2)) = 2.8m

Distance to Sam (70° Horizontal FoV): 1.88 * 1827/(430 * 2 * tan(70deg/2)) = 5.7m

Manga​

Maki Zen'in: 114px | 1.7m
Panel Height: 903px

Distance to Maki (70°): 1.7 * 903/(114 * 2 * tan(70deg/2)) = 9.6m

Distance to Maki (40°): 1.7 * 903/(114 * 2 * tan(40deg/2)) = 18.49m

Tatsumaki: 61.32px | 1.35m
Panel Height: 509px

Distance to Tatsumaki (70°): 1.35 * 509/(61.32 * 2 * tan(70deg/2)) = 8m

Distance to Tatsumaki (40°): 1.35 * 509/(61.32 * 2 * tan(40deg/2)) = 15.39m

Panel Height: 397px
Panel Width: 995px
Monster: 10px | 1.7m

Distance to monsters (70° Vertical FoV): 1.7 * 397/(10 * 2 * tan(70deg/2)) = 48.19m
Distance to monsters (70° Horizontal FoV): 1.7 * 995/(10 * 2 * tan(70deg/2)) = 120.7m

Explosion: 133px
Panel Height: 250px
Panel Width: 722px
Distance to the Explosion: 20 km

Explosion Diameter (70° Vertical FoV): 2 * tan(70deg/2) * 20 * 133/250 = 14.9 km
Explosion Diameter (70° Horizontal FoV): 2 * tan(70deg/2) * 20 * 133/722 = 5.15 km

 

[Mr. Bambu]

Some of these don't actually seem more accurate under your proposal, to be honest with you.

Take the first manga one for example. The character scaled is 7 tiles away from the POV- probably add 1 or 1.5 to account for ground leading to the POV's feet, which is out of sight. Each tile, though, is self-evidently shorter than the human (Maki) being scaled (per your suggestion, 1.7 meters). This means that under this approximation, and even generously (and obviously wrongly) assuming the tiles are equal to this character's height, the character is at most 14.45 meters away- damn near the middle point of the two results you gave. In reality, the distance is certainly even less.

Now, I chose to inspect this one because there was an easily grabbed measuring stick that did not require me to bust out any programs to check pixels directly. I would wager, then, that the others at least have a high possibility of being similarly flawed under inspection. Is it possible that 40 degrees isn't as sufficient a solution as you think?

 

[Kulf_Boba]

Some of these don't actually seem more accurate under your proposal, to be honest with you.

Take the first manga one for example. The character scaled is 7 tiles away from the POV- probably add 1 or 1.5 to account for ground leading to the POV's feet, which is out of sight. Each tile, though, is self-evidently shorter than the human (Maki) being scaled (per your suggestion, 1.7 meters). This means that under this approximation, and even generously (and obviously wrongly) assuming the tiles are equal to this character's height, the character is at most 14.45 meters away- damn near the middle point of the two results you gave. In reality, the distance is certainly even less.

Now, I chose to inspect this one because there was an easily grabbed measuring stick that did not require me to bust out any programs to check pixels directly. I would wager, then, that the others at least have a high possibility of being similarly flawed under inspection. Is it possible that 40 degrees isn't as sufficient a solution as you think?

I got the pixel scaling from another calculation, and I verified that the girl's height was incorrect, it was actually 142px:

The grid squares measure 0.658m, and the characters we see near the panel should be around 8 squares away.

Distance: 0.658 * 8 = 5.26m

For the formula to give a similar distance, the field of view should be close to 85°, which seems quite exaggerated. The grid squares were probably not drawn accurately, or if a calculation is made, it would be better to rely on that rather than use the angsizing.

Distance to Maki: 1.7 * 903/(142 * 2 * tan(85deg/2)) = 5.89m

Distance to the girl near the panel: 0.24 * 903/(350 * 2 * tan(85deg/2)) = 0.33m

Distance between characters: 5.89 - 0.33m = 5.56m

 

[Kulf_Boba]

In the other examples that use 40°, the value seems more precise to me:

Comics​

Electro's Head: 24px | 0.225m (measures 1.8m)
Daredavil Head: 87px | 0.225m (measures 1.8m)
Panel Height: 369px

Distance to Electro (70°): 0.225 * 369/(24 * 2 * tan(70deg/2)) = 2.47m
Distance to Daredavil (70°): 0.225 * 369/(87 * 2 * tan(70deg/2)) = 0.681m
Distance between characters: 2.47-0.681 = 1.78m

Distance to Electro (40°): 0.225 * 369/(24 * 2 * tan(40deg/2)) = 4.75m
Distance to Daredavil (40°): 0.225 * 369/(87 * 2 * tan(40deg/2)) = 1.31m
Distance between characters: 4.75-1.31 = 3.44m

Manga​

Tatsumaki: 61.32px | 1.35m
PPP Head: 47px | 0.31m (measures 2.2m)

Panel Height: 509px

Distance to Tatsumaki (70°): 1.35 * 509/(61.32 * 2 * tan(70deg/2)) = 8m
Distance to PPP (70°): 0.31 * 509/(47 * 2 * tan(70deg/2)) = 2.39m
Distance between characters: 8-2.39 = 5.6m

Distance to Tatsumaki (40°): 1.35 * 509/(61.32 * 2 * tan(40deg/2)) = 15.39m
Distance to PPP (40°): 0.31 * 509/(47 * 2 * tan(70deg/2)) = 4.61m
Distance between characters: 15.39-4.61 = 10.7m

 

[TheRustyOne]

Using Marvel's Spider-Man (PC), I'm able to see if using horizontal or vertical pixel size is more accurate. I do not know the FOV as the game doesn't tell you, you can change the FOV but the slider doesn't actually say what the FOV is. It's just a slider that goes from -25 to 25, with 0 being the default.

Scan 1

Spider-Man's Height = 177.8 cm or 5'10 (268 px)

Window Height = 436.538 cm or 4.36538 m (658 px)

Scan 2

As you can see there, the actual distance is 236 meters.

Window Height = 4.36538 m (28 px)

Screen Height/Length = 1080px and 1920px respectively.

Pixel Height Angular Size = 4.36538*1080/[28*2*tan(70deg/2)] = 120.235 m

Pixel Length Angular Size = 4.36538*1920/[28*2*tan(70deg/2)] = 213.751 m

In this case, using the horizontal length gave me a more accurate result. So I believe in cases like this, we should use the horizontal length. Obviously it's not exact, but we already acknowledge that angular size is a bit inaccurate. Although, there could be issues on my scaling as well, which could change the results.

Obviously in this case we can get the true distance, but I did this just to see what was more correct.

So I agree we should use the horizontal screen length for stuff like this.

Note: I cannot give any opinion on the different degrees, I'll leave that to more knowledgeable members. Although the OP is only suggesting 40 degrees for comic and manga panels, and that's only for panels that don't have paneling similar to a 16:9 aspect ratio. I agree with using the horizontal length for that stuff as well.

It's also possible I'm just being super dumb here, so obviously I'll be paying attention to what everyone has to say. I'm trying to understand here.

 

[Golden_Void]

Games​

I don't know Spider-Man's horizontal FoV, so I will assume a default of 70°.

Sandman: 32.57px | 24.93m
Panel Height: 720px
Panel Width: 1280px

Distance to Sandman (70° Vertical FoV): 24.93 * 720/(32.57 * 2 * tan(70deg/2)) = 393m

Distance to Sandman (70° Horizontal FoV): 24.93 * 1280/(32.57 * 2 * tan(70deg/2)) = 699.61m

The second method matches the distance indicated by the game where Sandman is after this panel, which is 640 meters away.

Idk how the game scaled down size. The distance between the towers on the Brooklyn bridge is 486.3 meters. Sandman is flush with the second tower, which means he should be half of that distance from the middle of the bridge. The bridge's parts that are submerged in water or directly over it measure 594.36m, which is the overall width of the East River channel. However, Spidey was 650+ meters away from Sandman and was at the halfway point, or less.

Games scaling down real-life counterparts makes games in the OP's argument a special case. The second angsize method seems more accurate than the game itself if we're trying to find real-life distance and not game distance (given a 100+ meter discrepancy).

 

[Kulf_Boba]

Idk how the game scaled down size. The distance between the towers on the Brooklyn bridge is 486.3 meters. Sandman is flush with the second tower, which means he should be half of that distance from the middle of the bridge. The bridge's parts that are submerged in water or directly over it measure 594.36m, which is the overall width of the East River channel. However, Spidey was 650+ meters away from Sandman and was at the halfway point, or less.

This seems to be because the map is not actually at a 1:1 scale with real life.

Games scaling down real-life counterparts makes games in the OP's argument a special case. The second angsize method seems more accurate than the game itself if we're trying to find real-life distance and not game distance (given a 100+ meter discrepancy).

If Sandman’s head measures 24.9 meters (according to another user’s scaling) and the game’s horizontal FoV is approximately 70°, the distance calculated with the second method should be fairly accurate. This seems to align with the distance shown later in the game, although I’m not entirely sure how precise it really is.

 

[Kulf_Boba]

By angsizing the buildings near Sandman, which according to Google Earth are 4.8 meters wide in this section I get this distance:

4.87 * 1280/(6.32 * 2 * tan(70deg/2)) = 704.31m

 

[Kulf_Boba]

Using Marvel's Spider-Man (PC), I'm able to see if using horizontal or vertical pixel size is more accurate. I do not know the FOV as the game doesn't tell you, you can change the FOV but the slider doesn't actually say what the FOV is. It's just a slider that goes from -25 to 25, with 0 being the default.

Scan 1

Spider-Man's Height = 177.8 cm or 5'10 (268 px)

Window Height = 436.538 cm or 4.36538 m (658 px)

Scan 2

As you can see there, the actual distance is 236 meters.

Window Height = 4.36538 m (28 px)

Screen Height/Length = 1080px and 1920px respectively.

Pixel Height Angular Size = 4.36538*1080/[28*2*tan(70deg/2)] = 120.235 m

Pixel Length Angular Size = 4.36538*1920/[28*2*tan(70deg/2)] = 213.751 m

In this case, using the horizontal length gave me a more accurate result. So I believe in cases like this, we should use the horizontal length. Obviously it's not exact, but we already acknowledge that angular size is a bit inaccurate. Although, there could be issues on my scaling as well, which could change the results.

Obviously in this case we can get the true distance, but I did this just to see what was more correct.

So I agree we should use the horizontal screen length for stuff like this.

Note: I cannot give any opinion on the different degrees, I'll leave that to more knowledgeable members. Although the OP is only suggesting 40 degrees for comic and manga panels, and that's only for panels that don't have paneling similar to a 16:9 aspect ratio. I agree with using the horizontal length for that stuff as well.

It's also possible I'm just being super dumb here, so obviously I'll be paying attention to what everyone has to say. I'm trying to understand here.

Thank you for your input; this further supports the use of horizontal FoV in games, which is typically around 70°. However, if you can calculate the size of an object and position it at a specific distance so that it is centered on the screen during a screenshot, I could determine the game's horizontal field of view more accurately.