Jump to content
You must now use your email address to sign in [click for more info] ×

Perspective grid an plane to create the illusion of depth


Recommended Posts

Affinity Designer already has a grid and axis manager, to select diferente types of grids, like isometric, dimetric, ….
Also it also has the isometric panel, that change planes to edit or fit objects in a given plane.

A new option of grid could be developed to do perspective drawings, like on paper with One Point Perspective or Two Point Perspective, that uses vanishing points to create the illusion of depth in an drawing. And probably using a similar organization of isometric drawings.
This use of the geometry in the grid and panel, would be very useful for to make this type of drawings.
Now, we need to create lines to achieve this, and it is not practical or productive.

Does anyone also finds this useful?

Cheers

Edited by va2m
add missing text and a better text organization
Link to comment
Share on other sites

The existing grids all have a limited number of well-defined planes to operate on.  A perspective grid would need to juggle an infinite number of planes leading away from the eye, which is a whole other ball game.

I'm not sure that this is practical without actual 3D support, which Designer does not have, and probably never will based on past comments from the developers.

Link to comment
Share on other sites

15 hours ago, fde101 said:

The existing grids all have a limited number of well-defined planes to operate on.  A perspective grid would need to juggle an infinite number of planes leading away from the eye, which is a whole other ball game.

I'm not sure that this is practical without actual 3D support, which Designer does not have, and probably never will based on past comments from the developers.

Lazy Nezumi Pro does the thing va2m is asking for and it does not rely on 3D support to work, or a grid for that matter. Heck, you can even load that add-on manually in the Affinity Suite to make all its perspective tools work, although it's slightly buggy since it doesn't fully support Affinity, and will not be supported as long as the programs lack high precision tablet input (which will be added at some point according the the devs). These perspective tools rely on mathematical equations to fake the perspectives in a non-3D environment. It's not really that different from the already existing perspective transform tools in Photo, which are also not real 3D.

Above is a quick video I recorded showing how this add-on fakes perspective when loading it into Affinity.

Link to comment
Share on other sites

Quote

I'm not sure that this is practical without actual 3D support, which Designer does not have, and probably never will based on past comments from the developers.

The kind of "vanishing point" converging perspective va2m is describing is entirely based on 2D construction. And it's nothing new. Macromedia FreeHand implemented a grid system for it way before it was copy-catted into Adobe Illustrator (after Adobe's acquisition of Macromedia).

But I'd suspect it would be cleaner (both from the programming perspective and for the user interface) to implement it separately from the parallel perspective grids feature still under development in Affinity.

JET

Link to comment
Share on other sites

  • Staff

I've thought about vanishing perspective tools - and as @JET_Affinity says, they will have to be completely separate to our standard grids and planes system.  From a 2D construction point of view, there is a lot more can be easily determined or inferred in parallel systems.  Converging perspective needs distortion - and to preserve any notion of the plane an object lives in is more involved (or even impossible without over-complicating the tools).

 

I will be working more on snapping, grids and custom guides soon.  It may well fall into that work, but I am making no promises.

SerifLabs team - Affinity Developer
  • Software engineer  -  Photographer  -  Guitarist  -  Philosopher
  • iMac 27" Retina 5K (Late 2015), 4.0GHz i7, AMD Radeon R9 M395
  • MacBook (Early 2015), 1.3GHz Core M, Intel HD 5300
  • iPad Pro 10.5", 256GB
Link to comment
Share on other sites

Meanwhile, as Frozen mentioned, to those who haven't yet, it really is worth looking at Lazy Nezumi Pro.

LPN's claim-to-fame is that it is not an "add-on" for any particular graphics program.  It sort of gives your computer some drawing smarts, instead of just a particular graphics program. The drawing or painting program you're using doesn't even know it's there. You just launch it and then click the application window that you want to work within. It works by affecting cursor movement at the system level, "smoothing" your pointing device movements to constrain to the on-screen guides that it effectively "overlays" within that window. It's the closest thing I've seen to software actually emulating the use of physical drawing aids like rotatable rulers and ellipse templates. The way it works gives it these huge advantages:

  • Although it has a quite sophisticated collection of setup options, it's worth the learning curve because you learn it once and it works the same way regardless of what drawing or painting program you're using at the moment.
     
  • Although originally presented as a drawing aid for raster-based drawing, it's not just for raster imaging. It's just as applicable in windows of vector-drawing applications; it just depends on your having selected an appropriate drawing tool. Yeah, it works with path drawing tools you think of as "freehand" (pencil, brush, etc.), but it's just as useful for a line tool (or Affinity's Pen Tool in Line Mode). It's basically appropriate for any tool you drag along the desired direction because that's what it does; it affects the smoothness or straightness of your drag inputs.
     
  • Its perspective features include configurations for constructing both converging perspective and parallel perspective.
     
  • Contrary to common misconception, it is not just for using a stylus (which I very seldom use). On desktops or laptops, it doesn't know or care what you are using for a pointing device. (I can't speak to the matter of finger gestures on cell phones because I quit finger painting before kindergarten. ;-)

So LNP is not something that competes with Affinity, it just serves as a drawing enhancement for it or whatever graphics program(s) you are using. I think you can download it as a demo to try it out in whatever program(s) you have in mind.

JET

Link to comment
Share on other sites

I really don't know the technical implications, but like Ben, JET_Affinity and Frozen Death Knight said this is not that complex, like 3D engines, and also not new. So thanks the context and extra info!
But I do know, that on paper, is easy, it's a long process in drawings, but the process of making it possible is not rocket science. It's just vanishing points where the lines of a given axis (or plane) converge/meet.

4 hours ago, Ben said:


I will be working more on snapping, grids and custom guides soon.  It may well fall into that work, but I am making no promises.

That is great, even without the promise. Thanks for that, and I wish you the best!
And if I may, know that you talk about guides, I never quite understood why no oblique line guides or other types of guides come as standard. Not to mention to auto define and update the 3 thirds rule, golden ratio, rectangle, …

18 hours ago, Frozen Death Knight said:

Lazy Nezumi Pro does the thing va2m is asking for and it does not rely on 3D support to work, or a grid for that matter.

The Tool "Lazy Nezumi Pro" as Frozen Death Knight mentioned appears to be a very cool tool indeed …unfortunately is only for windows (no Mac and no Linux).
I have found some similar tool for Mac (it is called  "Hej Stylus" for those who are wondering) but no quite like the Lazy Nezumi Pro, and no vanishing points to make the illusion of perspective using vanishing points.
Anyway, I still think this is necessary as a standard at any software drawing, and others tools like that to assist the process of drawing.

Thanks once again!

Link to comment
Share on other sites

21 hours ago, va2m said:

It's just vanishing points where the lines of a given axis (or plane) converge/meet.

Convincing 2D converging perspective is not that simple. If it were, we wouldn't need any special features for it. We can already just draw arbitrary horizon lines and perspective rays like we do on paper.

Nor do I assume it's that simple to implement in software. Have you tried the Perspective Grid feature in Adobe Illustrator (which it copied from FreeHand)? Many users no doubt think it must be rigorous just because it's in software. But it's easy to catch it committing the common error in which foreground grid "squares" become less foreshortened vertically than horizontally and therefore look like elongated rectangles instead of squares (not at all how the eye would see it).

Here's the thing: The kind of casual "vanishing point" perspective typically taught in basic art classes is actually not very rigorous. Students get exposed to so-called "one-point", "two-point", and "three-point" perspective and are led to assume that three-point perspective is some kind of end-all of ultimate realism just because the subjects we draw reside in 3D space. It's not. A view of 3D objects in 3D space can have an infinite number of "vanishing points." (Think of tossing your toddler's collection of building blocks upward into the air and taking a photo.)

Yet just watch a few of the many amateur YouTube videos on the subject and listen to how many times you hear the demonstrators use phrases like "There you have it; a perfect perspective of [this or that]!" Also note that they typically just place their vanishing points at any random locations relative to each other, without any kind of actual geometric basis.

Rigorous converging perspective must be based on the geometry of optics; of conic vision. Grid squares don't merely shrink toward the distance, they change in shape. (I wonder how many graphic designers and even illustrators understand that if you take a photo with a "normal lens," crop it down to a distant part of the subject, and then enlarge it, the resulting image looks like it was taken with a more "telephoto" lens.)

There are more formalized methods of constructing converging perspective. For example, before computers architects used a more elaborate construction method which did a better job of constraining things to realistic proportions. Technical illustrators had special equipment like Klok Perspective Boards and Tables and their graduated ruler scales.

That's not to say that the willy-nilly kind of informal converging perspective is "illegitimate" as an illustrative style. It's not. Grossly exaggerated perspective is commonly used just to add overstated drama in everything from Marvel's super hero comic books to the Marketing Department's rendition of the corporate trade show booth.

My point is that "vanishing point perspective" is a broad subject encompassing methods and purposes that range from rigorous to whimsical. So a "good" software implementation of it is an ill-defined target.

Axonometric drawing, by comparison, is not like that at all. There's nothing ambiguous or arbitrary about it. It is either geometrically correct or it is not, in fact, axonometric.

Again, the Lazy Nezumi Pro approach to interactive drawing aids causes me to re-think the need for program-native features for converging perspective. For example, one of the demo video clips on the LNP site demonstrates its Fisheye perspective variant. How often have you ever seen that kind of 2D perspective convincingly executed by the traditional "vanishing point" methods? That novel implementation alone represents a lot of potential.

In my previous post, I mentioned that the applicability of LNP to vector-based drawing is largely dependent upon the number and quality of Bezier path drawing tools that are used by dragging along the onscreen guides. It may be more advantageous for Affinity development to simply focus on optimized Bezier path tools that are of superior quality in terms of functional economy and editing elegance; i.e., the number and placement of nodes and handles automatically generated when the tool is dragged.

JET

Link to comment
Share on other sites

Like I said before I really don't know the technical implications, but the point of his post is that this features are necessary.
That's why we see "Perspective Grid" in Adobe Illustrator or even "Vanishing Point" in Adobe Photoshop.

So, I would like it to see something like "Perspective Grid" (if you like) in Affinity Designer.
The ability to edit, place objects and transform (to name a few things) in planes is very useful, especial if it's done in a non destructive way.
One, needs the ability to later change things and recover the original form of the objects.

On 7/17/2019 at 1:44 PM, JET_Affinity said:

Again, the Lazy Nezumi Pro approach to interactive drawing aids causes me to re-think the need for program-native features for converging perspective.

I don't think tools like Lazy Nezumi Pro can fill the requirements that I pointed out above, but they can be very helpful and importante in drawing with pointing devices … must like rulers (and alike) with pencils on paper. Probably, software like that should come with every pen table or other point device for professional use, or even in the operation system … how know's.

On 7/17/2019 at 1:44 PM, JET_Affinity said:

Convincing 2D converging perspective is not that simple

Thanks for explain what you know, it would be nice to talk more about it (and some links to the sources would be nice)…but I don't know if it's possible or advisable to deviate from the post. Anyway, I just like to say that probably it wasn't clear when I gave the example of the process on paper, or you just simple remove it from the context to make a point.
But, if we are talking "not that simple ", and "rigorous" and "geometry of optics", we could talk even about 3d software, lens cameras, even the human eye it self…
Hey, some could even argue that everything about this media it not accurate or real, it's just illusions to trick our mind…
This is not suppose to be scientific rigorous, that is another field…not the target of Affinity Designer software, I supposed.

Away, better to simplify and just focus on the topic.

Link to comment
Share on other sites

  • Staff

Here is the problem.  Our current parallel axonometric system allows for non-destructive transformations as the inverse is easy to determine.  The transformation is also just a simple shear-scale-rotation applied to the original object.  The underlying geometry itself remains unchanged.

We also use affine transformations for our document model, allowing us to deduce an inverse.

 

Converging perspective requires deformation of the original geometry.  Anything nearing the vanishing point will tend towards a size of 0.  It is impossible to then determine an inverse transform for anything that has been transformed in this way.  We do not currently preserve any notion of "original geometry" as such.  Only, a transform for which we can return to a 2D representation.

 

If we do converging perspective, it will have to use destructive deformations to stay in keeping with our current Document Model.  This is not a 3D application - it's a strict 2D application.  Any notion of 3D (or 2.5D) is purely illusion.  I've reiterated this point many, many times.  In order to provide a full perspective model a full concept of 3D position is required.  We are not going there.

 

SerifLabs team - Affinity Developer
  • Software engineer  -  Photographer  -  Guitarist  -  Philosopher
  • iMac 27" Retina 5K (Late 2015), 4.0GHz i7, AMD Radeon R9 M395
  • MacBook (Early 2015), 1.3GHz Core M, Intel HD 5300
  • iPad Pro 10.5", 256GB
Link to comment
Share on other sites

26 minutes ago, va2m said:

 if we are talking "not that simple ", and "rigorous" and "geometry of optics"...better to simplify and just focus on the topic.

va2m,

I am on topic. As I said:

That's not to say that the willy-nilly kind of informal converging perspective is "illegitimate" as an illustrative style. It's not. Grossly exaggerated perspective is commonly used just to add overstated drama in everything from Marvel's super hero comic books to the Marketing Department's rendition of the corporate trade show booth.

My point is that "vanishing point perspective" is a broad subject encompassing methods and purposes that range from rigorous to whimsical. So a "good" software implementation of it is an ill-defined target.

Again: The familiar kind of 2D "vanishing point" perspective drawing as taught in general art classes and as usually practiced, it is not a consistent, specific geometrically-derived construction discipline. It is done differently by different people for stylistic reasons. And there's nothing "wrong" with that. I'm saying because it is more style-driven than geometry-driven it is rather ambiguous; there is not a single clear-cut, well-defined, consistent construction method for it that embraces all that arbitrary stylistic freedom.

3D modeling software does indeed emulate what a lens sees. But that's an entirely different thing from just setting up one or two or three 2D converging grids in a 2D drawing program based on a few arbitrarily placed "vanishing points."

Quote

…we see "Perspective Grid" in Adobe Illustrator…I would like it to see something like "Perspective Grid"…in Affinity Designer.

 

Okay, so now you're more specifically saying you want something like Illustrator's (i.e., FreeHand's) feature. Frankly, I don't. It is stylistically limiting. You can't, for example rotate its horizon. It's always assumed to be horizontal. Again, look at comic books. The common bird's-eye-view perspective of a super hero soaring above NY sky scrapers often has non-horizontal "horizons" just as do views from the perspective of inside an airplane cockpit.

And why just three vanishing points? "Vanishing points" are not based on any three universal "locations in space", they are really just based on lines that an illustrator extends from parallel lines in whatever objects he is drawing. That works fine for cliché' rows of boxy objects like tall buildings conveniently oriented parallel to each other on streets that all conveniently intersect at right-angles. Depending on what you're drawing, you could need any number of "vanishing points."

Said another way, there is only one "real" vanishing point; the one corresponding to your line-of-sight, located at the center of your field of vision.

So how do you emulate that as guides in 2D drawing software? You'd need to provide as many arbitrary focal points as the illustrator wants to place on the page, and somehow make the all their envelope distortions interdependent relative to a center-of-vision.

Quote

I don't think tools like Lazy Nezumi Pro can fill the requirements that I pointed out above...

 

Which requirements? The ability to freely add and remove drawn shapes to and from the "planes"? No, LNP does not do that, because LNP is just providing you guides on your screen, not creating actual envelopes in your drawing program. That's what FreeHand's / Illustrator's Perspective Grid feature is: a set of envelopes. Affinity doesn't yet have an envelopes feature, but mesh and warp envelopes are on the planned feature list. Depending on how innovatively implemented that is, as far as we know, it could serve for "vanishing point" perspective among the other things for which envelopes are employed.

JET

 

Link to comment
Share on other sites

1 hour ago, Ben said:

The underlying geometry itself remains unchanged.

Can we expect a way to "undo" the "Fit to plane" operation applied to an object? And not only by going Ctrl + Z, but for any selected object. Or maybe it's there but I am missing it.

Thanks in advance.

Best regards!

Link to comment
Share on other sites

  • Staff

Yes - at some point.  There will be some improvements to axonometric tools.

SerifLabs team - Affinity Developer
  • Software engineer  -  Photographer  -  Guitarist  -  Philosopher
  • iMac 27" Retina 5K (Late 2015), 4.0GHz i7, AMD Radeon R9 M395
  • MacBook (Early 2015), 1.3GHz Core M, Intel HD 5300
  • iPad Pro 10.5", 256GB
Link to comment
Share on other sites

A few days ago I used the isometric perspective tool to make a mockup of a board game. The result is not bad, but with convergent perspective would be much better.
I have created the mockup in a separate document in which I copy and paste from other files the 2D elements of the game (map, cards, counters...), then pass them to isometric perspective using the function fit to plane.

For me, the convergent perspective would be very useful even if it was destructive.

Does anyone else find it interesting, is there any possibility of having this function in the future?

Link to comment
Share on other sites

I am trying to help, with all my limitations. I am no expert, or anything close, by no means …I am just a user.
If for some reason I am not helping, please let me know.

On 7/23/2019 at 11:57 AM, Ben said:

Our current parallel axonometric system allows for non-destructive transformations as the inverse is easy to determine.  The transformation is also just a simple shear-scale-rotation applied to the original object.  The underlying geometry itself remains unchanged.

We also use affine transformations for our document model, allowing us to deduce an inverse.

Ok, this is interesting and probably a base for a solution.
So, if I understood it correctly it relies on transformations for parallel axonometric system.
And at the moment, Affinity Designer has the Rotation and Skew/shear transformations.
Scale transformation apparently is also present, for what I understood, but doesn't have a visible control for the user (at least to my knowledge, see image below), which I think it should have.

1088653533_Capturadeecr2019-07-26s16_17_28.png.66a836b3104f2f202e38dc0031245f00.png

Well, trying to use the same principle that Affinity already uses, without changing the underlying geometry itself, the transformations could also be used to achieve the illusion of "Perspective".
Some new transformations would be required, but also useful for other effects and adjustments.
So, some transformations useful for achieving this illusion could be Scale, Rotation, Skew/shear, Stretch/compression, "Perspective" (Incline or tilt).
Note, that when I say "perspective" transformation I say some like the image below, where you can incline or tilt (probably there is a better name). So, this combine with Skew/shear could be possible to get different placements along the "z-axis", and combine with the Scale and Stretch/compression different placements along the "y-axis". 

169387486_Capturadeecr2019-07-26s16_58.05(2).png.678a51d9d2c51c9b30dbe143c104ece1.png

And the combinations between this transformations controls, would also allow many more non-destructive effects and adjustments.

Or, a Distort or/and a Warp transformations, but more controls are necessary for this.

By the way, now that we are talking about transformations (non-destructive and with controls), that can achieve other illusions and effects, why not Reflection (Flip / mirror). So useful!

On 7/23/2019 at 1:28 PM, Mithferion said:

Can we expect a way to "undo" the "Fit to plane" operation applied to an objec

Let just say, that It's possible (to my knowledge), to partially and manually "undo", at least the Rotation and Shear/Skew, putting back to 0, but because there is no scale transformation control visible, one can't put that back to 0. So it isn't an a full manual "undo". The possibility to "undo" would be very useful, manually or automatically.

On 7/23/2019 at 11:57 AM, Ben said:

Converging perspective requires deformation of the original geometry.  Anything nearing the vanishing point will tend towards a size of 0.  It is impossible to then determine an inverse transform for anything that has been transformed in this way.  We do not currently preserve any notion of "original geometry" as such.  Only, a transform for which we can return to a 2D representation.

But if Affinity Designer preserves the width, height (not the original geometry, but the current one) and the scale transformation, it is possible to change only the scale value, and the current width and height are preserve or not? So, the scale transformation is used to calculate the new dimensions (width and height). This could be used for the illusion of "perspective" as well, as for example, putting back the original size of a place image (very useful).

On 7/23/2019 at 11:57 AM, Ben said:

If we do converging perspective, it will have to use destructive deformations to stay in keeping with our current Document Model.  This is not a 3D application - it's a strict 2D application.  Any notion of 3D (or 2.5D) is purely illusion.  I've reiterated this point many, many times.  In order to provide a full perspective model a full concept of 3D position is required.  We are not going there.

Let me put this very clear, I really don't want Affinity Designer to become a 3D software. For 3D I use other softwares, better suitable for that purpose.

This is just an ideia, still in it early stages I suppose … it worth what is worth!
I try to explained the best that I could, but I don't know if is making sense or not.

But if, it's really not possible using the current Document Model, or other reason, then the use destructive deformations this is better than nothing.
But one of the thinks that I really like about Affinity Designer is the possibility to work with so many non-destructive options.

But anyway, probably this is not the way to go.
I just hope it helps in particular way!

Link to comment
Share on other sites

9 minutes ago, va2m said:

Let just say, that It's possible (to my knowledge), to partially and manually "undo", at least the Rotation and Shear/Skew, putting back to 0, but because there is no scale transformation control visible, one can't put that back to 0. So it isn't an a full manual "undo". The possibility to "undo" would be very useful, manually or automatically.

In tried it once with some text and clearly it didn't keep the proportions right. That's when I realised that it'd work great with this added feature to "undo" the transformation.

Best regards!

Link to comment
Share on other sites

  • Staff

@va2m I don't think you quite understand.  Anything that needs to distort an object such that it is scaled non-uniformly (ie, such as perspective) cannot be reversed using an inverse transform that can simply be derived from just inverting the transform components.  If you have something that has been squashed towards the infinite vanishing point, it is impossible to restore the size through a simple inverse scaling - since the object can potentially have been distorted to zero size.

 

Without preserving the data prior to the perspective transform it is also not possible to determine the starting geometry.  The perspective plane cannot be derived from the geometry alone.  That is what I mean when I say you'd have to preserve some 3D type knowledge of the initial object.

 

Affine transforms can be expressed as translation, scale, rotation and shear components.  It is easy to determine an inverse transformation of these.  If you need to warp the geometry, the inverse is not easily derived unless you have some way of knowing how that geometry was created.  Also, to apply a distortion/warp we have to apply it direct in a destructive manner (akin to how Photoshop applies all its transforms).

 

Our parallel axonometric system requires no destructive changes to the base geometry.  This is the major difference.

 

Vanishing perspective tools will come eventually, but they will work in an entirely different way to our parallel axonometric tools.

SerifLabs team - Affinity Developer
  • Software engineer  -  Photographer  -  Guitarist  -  Philosopher
  • iMac 27" Retina 5K (Late 2015), 4.0GHz i7, AMD Radeon R9 M395
  • MacBook (Early 2015), 1.3GHz Core M, Intel HD 5300
  • iPad Pro 10.5", 256GB
Link to comment
Share on other sites

  • 5 months later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...

Important Information

Terms of Use | Privacy Policy | Guidelines | We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.