ZDoom

[quote="Gutawer"]All your models should have a sprite backing them up anyway (I've had issues where actors with a model and no sprites just can't be placed down in the editor). The cheat can only show sprite actors since its not drawing the map as a 3d space, but as a 2d space.[/quote]
But all that still projecting to 2D plane of monitor...

[quote="Hypersonic"]It would be neat to have a separate 3D automap which uses the same 3D engine as when playing, just move the camera away from the player's eyes to way up in the air looking down like in Diablo, like Doom Builder visual mode.[/quote]
Try play in descent to point "I destroy reactor, where exit?". This have good look only in dead space.

[quote="Graf Zahl"]The automap doesn't use a depth buffer so there will be some problems with such an approach because things may overlap in a non-trivial fashion (that is, unless you clear the depth buffer after each model render.)[/quote].
If I correct, map sprite draw as billboard x/y axis. So what if draw model at side view? Find actor slope, in case if it use some dark trigonometric magick to be always perpendicular to floor with any slope, through cathetus of right angled triangle, with one angle equal to finding angle, than find normal to this cathetus in direction of north-south and draw it from this point of view.

*just*... :?

It would be neat to have a separate 3D automap which uses the same 3D engine as when playing, just move the camera away from the player's eyes to way up in the air looking down like in Diablo, like Doom Builder visual mode.

Trying out the chase command with chase_dist 9999 it seems that it would work out fairly well, if only the camera could 'puncture' through the ceiling. Though vertical sprites are difficult to see, but just like the current map mode they could be drawn horizontally instead of vertically. Models would solve this problem. Orthographic projection would solve the problem of walls blocking the view of floors of adjacent sectors.

The automap doesn't use a depth buffer so there will be some problems with such an approach because things may overlap in a non-trivial fashion (that is, unless you clear the depth buffer after each model render.)

Would it be possible to render the model (and/or voxel) on the automap using the expected angle with orthographic projection?

I wonder if it's possible to "render" the model offscreen (say from the front angle) and output that as a texture to be drawn as a sprite... :?:

(similar to the idea me, Rachael and dpJudas talked about on Discord once regarding a "dressing room" feature)

All your models should have a sprite backing them up anyway (I've had issues where actors with a model and no sprites just can't be placed down in the editor). The cheat can only show sprite actors since its not drawing the map as a 3d space, but as a 2d space.

am_cheat 2 dont show actors that use model instead of sprite on automap.
Is it ok and iddt cheat can show only sprite actors?