In the game I'm running, the PCs are about to embark on a wilderness trek during which they will be harried by an opponent who has access to the major image spell. This opponent seeks to distract, mislead, and delay them, forcing them to spend more time and resources reaching their objective than they would otherwise.

I am trying to determine what is possible within the constraints of the major image spell description. One thing that occurred to me is that, while the spell's effects are confined to a 20' cube, they might seem to be spread over a much greater area by tricks of forced perspective. For example, an illusory, 20' tall castle suspended in mid-air between the party and a distant hill might appear to be a full-size castle settled on the hill. Or a 20' wide illusion of a wall of flames might seem to be consuming a vast span of the horizon.

Or alternatively, an illusory pit might seem to be much deeper than 20' by "drawing" its walls and bottom in such a way as to make it seem deeper.

Can forced perspective be used with major image in this way?

Are there potential drawbacks to allowing this? Would any modifiers be appropriate for investigation checks to "disbelieve" them?

  • 1
    \$\begingroup\$ Hi ktt4d, welcome to rpg.se! Take our tour for the usual badge and visit the help center for site specific guidance. I think this is a great question. Thanks for participating. Good luck and happy gaming! \$\endgroup\$
    – linksassin
    Commented Aug 25, 2020 at 23:59
  • 2
    \$\begingroup\$ As the DM, there's no need for you to stick to spells exactly as they're written, especially outside of combat. If you want your NPC to use a hybrid of Major Image and Hallucinatory Terrain, go for it. Maybe adjust the DC downwards by a point to account for the fact that your NPC's using a higher level spell than would normally be possible \$\endgroup\$ Commented Aug 27, 2020 at 15:29
  • \$\begingroup\$ @LogicianWithAHat the main thing I’m trying to avoid is the players feeling like they should’ve had clues that something was amiss, or that they can’t trust my descriptions of the world. I want the PCs to feel frustrated and confused, but for the players themselves to have fun and have a vivid picture of what’s happening. So, I’m trying to hold myself to consistent “rules” for illusions, and the best way I can think of to do that is to stick to spell descriptions as written. \$\endgroup\$
    – ktt4d
    Commented Aug 28, 2020 at 1:44
  • 1
    \$\begingroup\$ @ktt4d You're using illusion magic. They can't (and shouldn't?) trust your descriptions of the world while that's in play. As long as they know that their opponent is manipulating their impression of their surroundings (and they certainly will after the first time), that's their clue. You can also make the illusions slightly sloppy - massive wildfires don't normally suddenly appear, maybe a castle on a hill isn't there the first time they look, or textures and details 'pop in' as they get closer, or it has a different number of towers each time they look at it. \$\endgroup\$ Commented Aug 28, 2020 at 10:44

2 Answers 2


Disclaimer: I have no experience with this in a gaming context, but I do work in a visualization research lab that gives me some insight into what this entails.

Binocular (i.e., two-eyed) perspective and depth perception only works in the first place because the world is highly consistent in its geometry. The human visual system, from the pupil, through the retina, and up multiple sections of the brain absolutely requires these consistencies to interpret what it is seeing. When I move ten feet forward, the pattern of light striking my eyes from different objects changes in highly consistent ways that let me infer-- with very high confidence-- what the positions of those objects are.

I mention this because one of the weaknesses of forced perspective illusions is that they in some sense depend on the "accident" of a person standing at just the right spot to create the illusion, and the farther you get from that spot, the less effective the illusion becomes until it just falls apart. (Note that I have seen this with my own eyes using virtual reality equipment. It doesn't take very many steps away for perspectives to get messed up, even when forced perspective illusions are not the subject being modeled. I have seen people in shared environments argue over the size of virtual objects purely because they were drawn precisely for the perspective of Subject A, while Subject B was standing five or ten feet away.)

And as soon as the brain detects these anomalies, it will almost involuntarily focus attention on them until the anomaly is understood. Usually this will involve moving the head, craning the neck, walking around something, etc. Human brains don't like inconsistent visual data.

So if we think about this in terms of virtual reality, that brings to mind the following very probable limitations:

  1. If the image cannot move to respond to the movement of a viewer, it will almost certainly fail (or at least grant some significant mechanical bonus to the viewer on their Investigation check.) But this is not automatically insurmountable, because the image can move... as long as the caster is within 120 feet.

  2. For this type of application specifically, affecting more than one creature at one time will be difficult. It is very hard to convey in words how rapidly perspective anomalies add up, but they do, even over a few paces. I would interpret this as mounting bonuses with additional numbers of viewers.

  3. The more freedom someone has to walk around the illusion, or the more it depends on great distance, the faster it will break down. This makes things like "The floating castle above a hill" or "a wall a flame across a broad plain" are really hard to pull off-- when the subjects move, the illusion will inevitably break down.

  4. As a corollary, in enclosed spaces, this technique might be highly effective. The illusion of a bottomless pit passes my smell test, for instance (subject to 2 above.) The illusion of a vast cave containing a great dragon at the end of a passage, likewise. Possibly even the castle on a hill or wall of flame if seen from inside a building through a single window.

  5. Do not try this against ettins, Tiamat, or other multi-headed monsters. They will eat you.

The real answer to what's possible depends on a bunch of trigonometry for each application. But the five points above, I think, are a useful set of guidelines for what's possible, and what will keep your players or your GM from crying foul.

(Side note: The acoustic interactions, if any, will suffer the same basic problems but, for humans, much less severe. I would not assign a penalty to an illusion that incorporates an acoustic effect, because it would not be more difficult for the caster and probably would not help the subject pierce the illusion.)

  • 3
    \$\begingroup\$ An excellent example of multiple viewers screwing up forced perspective is in the Mission Impossible:Ghost Protocol movie. The scene has the agents behind a screen that is showing an empty hallway. There is also a device that is tracking a single guard's eye movement. As his line of sight changes, so does the image shown. Later, when a second guard enters the scene, the camera is forced to switch between the two causing the image to stutter. When three and four show up, it's just a mess. Attached video shows the setup. \$\endgroup\$
    – MivaScott
    Commented Aug 26, 2020 at 16:36
  • \$\begingroup\$ @MivaScott yeah, that's a surprisingly faithful portrayal of the basic idea. Most modern systems use a tracking dot on the forehead or on glasses of the primary target. In the cold light of the morning, though, I am considering whether I should muddy the waters by mentioning that real systems like this present a subtly different image to each eye in order to complete the verisimilitude of 3D rendering. \$\endgroup\$
    – Novak
    Commented Aug 26, 2020 at 17:30
  • 1
    \$\begingroup\$ Ettins seem to be biologically close to human conjoined twins (perhaps even related?) - with each head having its own consciousness, and the two heads are really close together. I believe they would be quite susceptible. Tiamat on the other hand... has a singular mind with a large amount of eyeballs, and despises trickery. You will be eaten. \$\endgroup\$ Commented Aug 27, 2020 at 4:52
  • 1
    \$\begingroup\$ Excellent answer. Worth noting that things like flames, clouds, mist, etc. might work a lot better as there are less clear boundaries \$\endgroup\$ Commented Aug 27, 2020 at 8:27
  • \$\begingroup\$ I've seen chalk drawings that can do this convincingly for a wide array of distances and angles, but they are all the bottomless pit (or similar) case, and they do fail for well over half the visual field. \$\endgroup\$ Commented Aug 29, 2020 at 23:19

Major Image is open ended

The spell description for major image doesn't go into a great deal of detail on how it can or can't be used. I believe this is deliberate, allowing for players to get creative and DMs to be flexible in its use. So the answer is definitely up to the DM, but lets look at what we are given to help decide if this is a potential use.

You create the image of an object, a creature, or some other visible phenomenon that is no larger than a 20-foot cube.

A forced perspective image isn't an object or a creature so we are looking at "some other visible phenomenon". This phrasing covers basically everything, you can make an image of anything you want. The question is really if the forced perspective will be believable.

It seems completely real, including sounds, smells, and temperature appropriate to the thing depicted.

At least initially we have some support that this should work. The phrase "seems completely real" is fairly unambiguous. But the wording generally supports a single entity rather than an optical illusion so it is unclear if this will work.

A word on Forced Perspective

Forced Perspective is a form of optical illusion to make thing appear smaller/larger, closer/further away than they are. The limitation however is that Forced Perspective illusions only work from a single direction. Moving your perspective or viewpoint usually breaks the illusion.

The spell allows you alter it to appear realistic. Therefore you can shift the perspective to mirror the movements of the creatures you are trying to deceive, provided they are close enough to have a single perspective on your image. If the viewers spread out and view the image from different directions it won't appear realistic for all of them. Novak's answer goes into this in more detail if you are interested in the science.

Allowing it at the table

In the past I have allow my players to use major image to conceal themselves within a dungeon. Hiding in a dead-end hallway they cast the spell in front of them to make the hallway appear empty, thus avoiding attention from the patrolling guards.

If you are to allow the spell to work this way it becomes a major tactical advantage for those who have access to it. I personally choose to allow it as I like the kind of creative play that it encourages. However you may need to be wary of overuse. If you find your players begin to use this single spell to avoid entire encounters you may like to revisit your ruling.

Bonuses to disbelief

As written a creature will disbelieve the illusion if physically interacted with or if it:

uses its action to examine the image [to] determine that it is an illusion with a successful Intelligence (Investigation) check against your spell save DC.

As I already mentioned, forced perspective illusions only work for a single direction. Therefore it is fairly reasonable to grant advantage on this check as this kind of illusion is less realistic than the typical uses of this spell.

Whether or not you choose to grant advantage is really a choice in how effective you want this tactic to be. From your question is appears that you want this to be a major plot point and something you wouldn't want the players to easily uncover. Therefore you might prefer not to grant advantage in this situation.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .