Ay, assumptions + questions about distortion,,, ya don't have to answer them, but would be great to know if i'm completely misunderstanding something. Thanks! Assumption1: The term "distortion" in linear perspective drawing = warping/stretching of a projection/image due to wrong location (sp) and/or direction (cp) of the viewer, causing exaggerated proportions and making dimensions/iconic shapes unrecognizable/hard to read. note: distortion has nothing to do with the actual 3D world directly,,, only the interpretation of a snapshot of it,,, Example1: Some cubes floating Infront of you. You paint what you see on a glass picture plane in front of you while only looking at CP... now if you look away from cp at any of the cubes you painted, they will appear slightly distorted,,, different to how they would look if you looked directly at each of the cubes irl (moved your cp to them). Question1: You take a wide angle photo from the corner of a square field looking towards the opposite corner, and hang it on your wall.... On the photo the squares VP's fall on the horizion lines 45deg points 1m apart... If you hang the photo on a wall, standing 2 meters infront of it at eyelevel looking at the same CP,,, The photos will appear zoomed out (smaller than life), for the projection to look the same as real life, the angles/cov must match (45deg points of the rotated square must fall at the same points on the horizon a rotated square in reality),,, The station point is wrong distance from subject,, but cp is correct? is this zooming in/out also called distortion? ASSUMPTION 2: Taking a photo from 1ft above a chessboard, the center squares are same size as corner squares which are relatively much further away from the cameras station point/center point... yet with our eyes/curvilinear lens they would appear smaller,,, also the diagonal corner squares are greater distance than the horizontal/vertical edge squares,,, if you looked at this photo to scale, and looked at the center point, standing at station point, our own eyes would correct these side/diagonal squares to the correct shape/dimensions? Question2. If you wanted the drawing to look correct in person, you would need to hold it flat away from one eye at the appropriate distance, and our eyes curved vision would correct the circle to look normal when focusing on the cp? Same for a camera/youtube audience, they would also have to be certain distance from the subject & screen. (you could calibrate it with 45deg vp's being certain distance apart on screen + user holding screen a certain distance from 1 eye) this is the benefit with linear perspective (if correct sp/cp enitre scene will look same as real life) Question3. Why tilt the paper when trying to show the view from station point, won't our eyes do that for us if we are looking at cp? Does tilting the paper emulate the fisheye effect (so that the circle looked like a circle you would see in that 3d scene if you moved your CP/eyeline to it) I assume if someone wanted all these circles in a linear drawing to look "circular" wherever you looked you would want to project the drawing onto a curved glass window/sphere your standing behind, and match the 45deg vp + sp... transforming it into curvilinear? (if correct sp,,, then you can move your cp/eye line around and all the shapes/views in a narrow fov would look undistorted/real,, .... (at the cost of the full scene not looking accurate when at correct cp,, unlike in normal linear)? Thanks again for the videos, i find distortion so interesting :)
@trustyourperspective
9 ай бұрын
Even curvilinear perspective will be distorted if you flatten out the curved picture plane, which you must do if you are drawing on a flat paper. Horizontal lines become curved, unless your eye is at the SP and the paper you are drawing on is also curved. Then you have no distortion. I needed to tilt the paper because my camera doesn't have that wide of a field of view
@Meowch3
9 ай бұрын
I'm still just a baby perspective chick, but I'd like to give this a try based on what I know. There is barely any info on this topic, so if anyone out there is confident they've got it absolutely right, please feel free to swoop in and correct me. Assumption 1: Yes, you've got it right. The thing to remember is that distortion occurs because we are drawing on a flat piece of paper. There is only one accurate point on that paper that looks the same as it would in real life, and that is the dead center of the flat paper because it is the only place that is perpendicular to our viewpoint (assuming it’s exactly in front of us). Photographs are the same since they are flat, though they may also be subject to lens distortion as well as this flat surface distortion. In real life, since our eyeball is round, everything is perpendicular to our viewpoint, so there is no distortion. Everything appears accurately. Example 1: The original image you drew was from a specific viewpoint (station point) AND a specific direction of view (the perpendicular line going from your viewpoint to your flat surface, aka the direction our flat surface is facing us). If you keep your viewpoint/station point the same and just rotate your eyeball to look at the cubes you drew, your perspective does not change, so there is still no distortion. Looking around with your eyeballs does not change the perspective of anything because your position in space remains unchanged. On the other hand, if we moved our head to look at the cubes, our station point would also move, so there would be distortion then, but if within the cone of vision, it wouldn’t be noticeable. Question 1: Yes, even if the cp is the same, if you change your distance from the picture, you will no longer be standing at the station point, so distortion will become more noticeable outside the cone of vision. It might not really be accurate to call this zooming in and out, especially if the picture is going well past the 60 degree cone. The picture will never look exactly like a smaller or larger version of itself by changing your distance since distortion will cause the sides to look more stretched out past the cone of vision the farther away you are. The stuff inside the 60 degree cone of vision is also distorting, but not enough to be noticeable. Assumption 2: Yes. Looking at this from the station point, our eyes would correct the side/diagonal squares to the correct shape/dimensions, just as you say. In order for that to happen, you'd have to draw them all the same size. If you draw them "correctly" on a flat surface using curvilinear perspective the first time around, it will never look completely right from anywhere, including the station point. Some people insist that curvilinear perspective is accurate and is how we actually see, but they say that because they don't understand this phenomenon. The truth is, both curvilinear and linear perspective give us distorted versions of reality (except of course when we are standing at the station point for linear perspective). Our retina is round, so there is just no way to capture reality perfectly on our flat piece of paper. Question 2: To your first question, yes and no. Yes in that holding the paper away from our eye at the correct point in space will cause the image to appear correctly, but no in that you do not have to focus on the cp. Just by being in the right spot, it will appear correctly, no matter which direction you look with your eyeball. The important thing is that our flat surface is facing us so that the center is perpendicular to our viewpoint. Where we look with our eyeball doesn't matter. For a camera/youtube audience, yes, but in order for the audience to feel like they are right there with you, they'd probably have to put their eye right up to the screen, or more likely, it will be impossible to get close enough. Question 3: It is not enough to look at cp. We must also be the right distance to the paper for it to appear correctly. And technically it is not the act of looking at the cp that makes it appear correctly, but the flat paper being the right orientation to our viewpoint. Tilting does not emulate the fisheye effect. It is just to demonstrate how the stretched circle would correct itself if seen from the right position. The circles in a linear perspective drawing will appear exactly as they would in real life if seen from the station point, so it is not necessary to draw them on curved glass for them to appear "circular." From my understanding, the act of looking all around with your eyeball does not change the perspective of anything. Although we can only point our eyeball in one direction, we can still SEE in many different directions all at once, which is how we're able to watch big movie screens and drive cars and all that. Linear perspective is taking that one direction we're pointing our eyeball to and saying, "OK, we will place our flat surface perpendicular to this line of sight and calculate everything from there." So we can look around with our eyeballs in many different directions while our viewpoint stays the same, and thus what we see stays the same too. One viewpoint = one perspective.
@58elrond
9 ай бұрын
@@Meowch3 Thanks for this by the way, alot to think about!
Пікірлер: 6
Thanks for your time
Perfect!!
Ay, assumptions + questions about distortion,,, ya don't have to answer them, but would be great to know if i'm completely misunderstanding something. Thanks! Assumption1: The term "distortion" in linear perspective drawing = warping/stretching of a projection/image due to wrong location (sp) and/or direction (cp) of the viewer, causing exaggerated proportions and making dimensions/iconic shapes unrecognizable/hard to read. note: distortion has nothing to do with the actual 3D world directly,,, only the interpretation of a snapshot of it,,, Example1: Some cubes floating Infront of you. You paint what you see on a glass picture plane in front of you while only looking at CP... now if you look away from cp at any of the cubes you painted, they will appear slightly distorted,,, different to how they would look if you looked directly at each of the cubes irl (moved your cp to them). Question1: You take a wide angle photo from the corner of a square field looking towards the opposite corner, and hang it on your wall.... On the photo the squares VP's fall on the horizion lines 45deg points 1m apart... If you hang the photo on a wall, standing 2 meters infront of it at eyelevel looking at the same CP,,, The photos will appear zoomed out (smaller than life), for the projection to look the same as real life, the angles/cov must match (45deg points of the rotated square must fall at the same points on the horizon a rotated square in reality),,, The station point is wrong distance from subject,, but cp is correct? is this zooming in/out also called distortion? ASSUMPTION 2: Taking a photo from 1ft above a chessboard, the center squares are same size as corner squares which are relatively much further away from the cameras station point/center point... yet with our eyes/curvilinear lens they would appear smaller,,, also the diagonal corner squares are greater distance than the horizontal/vertical edge squares,,, if you looked at this photo to scale, and looked at the center point, standing at station point, our own eyes would correct these side/diagonal squares to the correct shape/dimensions? Question2. If you wanted the drawing to look correct in person, you would need to hold it flat away from one eye at the appropriate distance, and our eyes curved vision would correct the circle to look normal when focusing on the cp? Same for a camera/youtube audience, they would also have to be certain distance from the subject & screen. (you could calibrate it with 45deg vp's being certain distance apart on screen + user holding screen a certain distance from 1 eye) this is the benefit with linear perspective (if correct sp/cp enitre scene will look same as real life) Question3. Why tilt the paper when trying to show the view from station point, won't our eyes do that for us if we are looking at cp? Does tilting the paper emulate the fisheye effect (so that the circle looked like a circle you would see in that 3d scene if you moved your CP/eyeline to it) I assume if someone wanted all these circles in a linear drawing to look "circular" wherever you looked you would want to project the drawing onto a curved glass window/sphere your standing behind, and match the 45deg vp + sp... transforming it into curvilinear? (if correct sp,,, then you can move your cp/eye line around and all the shapes/views in a narrow fov would look undistorted/real,, .... (at the cost of the full scene not looking accurate when at correct cp,, unlike in normal linear)? Thanks again for the videos, i find distortion so interesting :)
@trustyourperspective
9 ай бұрын
Even curvilinear perspective will be distorted if you flatten out the curved picture plane, which you must do if you are drawing on a flat paper. Horizontal lines become curved, unless your eye is at the SP and the paper you are drawing on is also curved. Then you have no distortion. I needed to tilt the paper because my camera doesn't have that wide of a field of view
@Meowch3
9 ай бұрын
I'm still just a baby perspective chick, but I'd like to give this a try based on what I know. There is barely any info on this topic, so if anyone out there is confident they've got it absolutely right, please feel free to swoop in and correct me. Assumption 1: Yes, you've got it right. The thing to remember is that distortion occurs because we are drawing on a flat piece of paper. There is only one accurate point on that paper that looks the same as it would in real life, and that is the dead center of the flat paper because it is the only place that is perpendicular to our viewpoint (assuming it’s exactly in front of us). Photographs are the same since they are flat, though they may also be subject to lens distortion as well as this flat surface distortion. In real life, since our eyeball is round, everything is perpendicular to our viewpoint, so there is no distortion. Everything appears accurately. Example 1: The original image you drew was from a specific viewpoint (station point) AND a specific direction of view (the perpendicular line going from your viewpoint to your flat surface, aka the direction our flat surface is facing us). If you keep your viewpoint/station point the same and just rotate your eyeball to look at the cubes you drew, your perspective does not change, so there is still no distortion. Looking around with your eyeballs does not change the perspective of anything because your position in space remains unchanged. On the other hand, if we moved our head to look at the cubes, our station point would also move, so there would be distortion then, but if within the cone of vision, it wouldn’t be noticeable. Question 1: Yes, even if the cp is the same, if you change your distance from the picture, you will no longer be standing at the station point, so distortion will become more noticeable outside the cone of vision. It might not really be accurate to call this zooming in and out, especially if the picture is going well past the 60 degree cone. The picture will never look exactly like a smaller or larger version of itself by changing your distance since distortion will cause the sides to look more stretched out past the cone of vision the farther away you are. The stuff inside the 60 degree cone of vision is also distorting, but not enough to be noticeable. Assumption 2: Yes. Looking at this from the station point, our eyes would correct the side/diagonal squares to the correct shape/dimensions, just as you say. In order for that to happen, you'd have to draw them all the same size. If you draw them "correctly" on a flat surface using curvilinear perspective the first time around, it will never look completely right from anywhere, including the station point. Some people insist that curvilinear perspective is accurate and is how we actually see, but they say that because they don't understand this phenomenon. The truth is, both curvilinear and linear perspective give us distorted versions of reality (except of course when we are standing at the station point for linear perspective). Our retina is round, so there is just no way to capture reality perfectly on our flat piece of paper. Question 2: To your first question, yes and no. Yes in that holding the paper away from our eye at the correct point in space will cause the image to appear correctly, but no in that you do not have to focus on the cp. Just by being in the right spot, it will appear correctly, no matter which direction you look with your eyeball. The important thing is that our flat surface is facing us so that the center is perpendicular to our viewpoint. Where we look with our eyeball doesn't matter. For a camera/youtube audience, yes, but in order for the audience to feel like they are right there with you, they'd probably have to put their eye right up to the screen, or more likely, it will be impossible to get close enough. Question 3: It is not enough to look at cp. We must also be the right distance to the paper for it to appear correctly. And technically it is not the act of looking at the cp that makes it appear correctly, but the flat paper being the right orientation to our viewpoint. Tilting does not emulate the fisheye effect. It is just to demonstrate how the stretched circle would correct itself if seen from the right position. The circles in a linear perspective drawing will appear exactly as they would in real life if seen from the station point, so it is not necessary to draw them on curved glass for them to appear "circular." From my understanding, the act of looking all around with your eyeball does not change the perspective of anything. Although we can only point our eyeball in one direction, we can still SEE in many different directions all at once, which is how we're able to watch big movie screens and drive cars and all that. Linear perspective is taking that one direction we're pointing our eyeball to and saying, "OK, we will place our flat surface perpendicular to this line of sight and calculate everything from there." So we can look around with our eyeballs in many different directions while our viewpoint stays the same, and thus what we see stays the same too. One viewpoint = one perspective.
@58elrond
9 ай бұрын
@@Meowch3 Thanks for this by the way, alot to think about!