Obique vs isometric

[Sokurah]

Look at all of you - all grown up and wanting to be ONE HUNDRED PERCENT PRECISE ... like grownups ...
... after decades of "being okay with something that adequately describing".

... I wanna be young again

(j/k, I have turned into a grammar and spelling nazi in my later years too )

[hikoki]

Please serve me a 3dish of 3urdy 3stles

[p13z]

I really think [mention]4thRock[/mention] nailed this one.
- "Isometric" graphics -
with Isometric in quotes - applying to all games that would be colloquially referred to as Isometric within the context of 80s video games.
Surely enough to disarm the pendant trigger and not upset anyone or overcomplicate things.

[Rorthron]

Personally I think we are better leaving things as they are. But if we really want to open this can of worms, we should come up with a proper solution. Honestly I hate the option with inverted commas. It's trying to saying isometric, while pretending it's not. It's the sort of trick the Daily Mail uses in headlines like FACEBOOK "GIVES YOU CANCER" (hint: it doesn't). Either use the isometric description or don't, but hiding behind quotation marks doesn't really solve anything in my opinion.

Definitely not enough to disarm the pedant trigger.

[Joefish]

It seems to me there's a fairly clear distinction between games with two diagonal axes in their fixed-3D projection, like Knight Lore, and those with just one diagonal axis, like Rasputin. You can argue the toss over whether the 2:1 slope instead of a 30° angle is strictly 'isometric' or not (I'm inclined to agree it's close enough). But however long you can drag out that argument, it's pretty clear there's a fundamental difference between that and the 'oblique projection' of other games.

I'd argue the case that 'isometric' is fine to apply to Ant Attack, Knight Lore, etc. But should not be applied to Rasputin, Orbix etc. It's not just inaccurate in that case; it's fundamentally wrong! So label the 'Oblique Projection Graphics' games as such and the others can keep the label 'Isometric Graphics'.

Then the argument over whether that particular 'Isometric Graphics' label needs quoting, italicising, or shimmering colour effects applied to every third vowel or whatever can just twitter on all by itself without constantly being interrupted by anything approaching a rational argument.

(Also, I would posit that the general ignorance of 80s games journalists on this particular technical issue is not a valid reason for perpetuating the error today!)

[Einar Saukas]

It seems to me there's a fairly clear distinction between games with two diagonal axes in their fixed-3D projection, like Knight Lore, and those with just one diagonal axis, like Rasputin. You can argue the toss over whether the 2:1 slope instead of a 30° angle is strictly 'isometric' or not (I'm inclined to agree it's close enough). But however long you can drag out that argument, it's pretty clear there's a fundamental difference between that and the 'oblique projection' of other games.

I'd argue the case that 'isometric' is fine to apply to Ant Attack, Knight Lore, etc. But should not be applied to Rasputin, Orbix etc. It's not just inaccurate in that case; it's fundamentally wrong! So label the 'Oblique Projection Graphics' games as such and the others can keep the label 'Isometric Graphics'.

I was resisting the temptation to re-discuss everything but now that you mentioned it, I must confess that I agree with you...

IMHO there are basically 3 very distinct 3D graphic styles here:

• "Isometric-like 3D graphics" (i.e 2 symmetrical diagonal axis) like Ant Attack, Knight Lore, W*H*B, etc. Technically it corresponds to "isometric", "dimetric", and "oblique military" projections.

• "Cabinet 3D graphics" (i.e. only 1 diagonal axis) like Ball Breaker, Colony, etc. Technically it corresponds to the "oblique cabinet" projection only.

• "Trimetric 3D graphics" (i.e 2 asymmetrical diagonal axis) like Crystal Castles and Nether Earth. Technically it corresponds to the "trimetric" projection only.

[Joefish]

Kind of agree - Nether Earth is a bit of a nuisance there!

But I'd still use 'Oblique' as the middle description of those three. 'Cavalier' and 'Cabinet' are two very specific things, neither of which is strictly accurate. So there's no point trying to be that specific.

In 'Cavalier' projection the diagonal lines are drawn to their full length, which looks strange. Even without perspective, you kind of expect objects to appear foreshortened.

The Wikipedia definition of 'Cabinet' perspective contradicts itself though. First it says the depth lines are halved in length, then later shows an image using a 2/3 factor. As far as I can tell from searching for other sources, 'Cavalier' means precisely a 1:1 scale and 'Cabinet' means precisely a 1/2 scale. Rasputin isn't using either of these. You also have to measure along the diagonal to get the scale, so any count of X pixels along and Y up isn't going to give a simple result. But since it's neither 1:1 nor 1:2 it is neither of these. It is simply an 'Oblique Projection' with a scale factor all its own.

[Einar Saukas]

Perhaps "Cabinet-like 3D graphics"?

The term "oblique" doesn't mean only 1 diagonal axis. In particular, W*H*B graphics could be classified as military oblique!

[Rorthron]

Based on my authoritative Wikipedia-level knowledge...

[mention]Joefish[/mention] is right: Cabinet, Cavalier and Military are specific types of oblique projection, which don't necessarily correspond to what we see in games. In fact, as the differences depend on knowing the dimensions of the objects they are supposed to depict, it is impossible to draw a distinction between Cavalier and Cabinet in games, for example. It doesn't make sense to distinguish between the different types of oblique projection.

Also, [mention]Einar Saukas[/mention], I don't think you're right to construe the different forms of projection in terms of "diagonal" lines based on the orientation of the display. I am not sure the orientation of the axes relative to the canvas matters. The important issue is their orientation relative to each other. For example, W*H*B and Nether Earth are, I think, both oblique projection. In both two axes intersect at 90 degrees.

But perhaps we should defer to people who actually know what they're talking about.

[Einar Saukas]

Also, @Einar Saukas, I don't think you're right to construe the different forms of projection in terms of "diagonal" lines based on the orientation of the display. I am not sure the orientation of the axes relative to the canvas matters.

I disagree. These Wikipedia definitions are based on the orientation of the axes. If you rotate the "oblique military" image by 135 degrees counter-clockwise, it will look a lot like "cabinet" and "cavalier" images, but I don't think it could be called "oblique military" anymore!

[Rorthron]

I disagree. These Wikipedia definitions are based on the orientation of the axes.

Not in the way you describe. The Wikipedia definition of oblique projection does not refer to angles relative to a reference outside the projection. If you can find an example of what you are saying, I'm happy to be corrected, but I've re-read the passage a few times and cannot find the language you mention.

Wikipedia does use similar language to yours in respect of Military oblique projection, but that is a special case. It is fallacious to infer the properties of a superset from those of a subset.

I've checked sources other than Wikipedia and they seem to agree on this point.

In any case, your definition does not make sense in some respects. First, it doesn't lead to a consistent definition of oblique projection. As you note, you end up with different numbers of "diagonal" axes in different types of oblique projection and there is inconsistency among them. Second, you end up with the absurd situation that the type of projection changes depending on where you are sitting! If you sit at 45 degrees to the paper (or turn the paper 45 degrees), "straight" lines become "diagonal" and vice versa.

As I said, I'm no expert, but I really think you've got the wrong end of the stick.

[catmeows]

Obligue/cabinet etc shows most important side of object as perfect scaled 2d image of object - x and y axes are perpendicular. The second most important thing is scale on third axe - it supposed that you, as engineer, will be able to measure "depth" of object. Typical ratio is 1 or 1/2 of scale used in front of object.
The angle the third axe meets XY is not so important but due the way how you actually do drawing with pencil, to minimize error, you should choose 45 degrees or something similar. 60 or 30 works too.
What I would like empaphise is that obligue & family are very real way how to describe 3D object. It is not supposed to be nice, it is supposed to be acurate.

[Rorthron]

Obligue/cabinet etc shows most important side of object as perfect scaled 2d image of object - x and y axes are perpendicular. The second most important thing is scale on third axe - it supposed that you, as engineer, will be able to measure "depth" of object. Typical ratio is 1 or 1/2 of scale used in front of object.
The angle the third axe meets XY is not so important but due the way how you actually do drawing with pencil, to minimize error, you should choose 45 degrees or something similar. 60 or 30 works too.
What I would like empaphise is that obligue & family are very real way how to describe 3D object. It is not supposed to be nice, it is supposed to be acurate.

Yes! Two perpendicular axes, plus a third at any angle = oblique projection.

[catmeows]

I'm glad you are happy but honestly I think this discussion Is little bit pointless. We are talking about technical drawings while games use artistic interpretation of technical projections and in case of ZX, limited by resolution and most often also by hardware specs (byte boundary, attribute size etc). That's why I think "isometric games" is quite good and actually, using different term would defeat purpose od database - to find things, at least for users without education (be it either formal oř informal) in subject.
My nightmare is that someone would come with idea of building 'proper' hiearchy for sport or management or strategy games.
And last thing - db is used by international community of users and many of them are not native speakers. I'm fine with using english as lingua franca but I definitely don't have vocabulary as big as a native speaker. So I would like to see db tags in rather limited english without too much extravaganza.
Just my 2p.

[Rorthron]

Actually, I agree. As I've said, personally I'd be happy to stick with "isometric", but if people really want to switch to technically accurate terminology, we should use technically accurate terminology! Introducing complicated new labels that are also wrong is the worst of both worlds.

[Joefish]

It seems to me that some have seized on the 'Military Projection' diagram purely to confuse and disrupt the argument over the use of 'Oblique Projection' to describe a view like Rasputin!

[Rorthron]

Incidentally, just as [mention]Ralf[/mention] suggested at the start of the thread, Wikipedia says that "isometric" in computer graphics has a different meaning from in technical drawing, and encompasses, at least in principle, oblique projection:

"Despite the name, isometric computer graphics are not necessarily truly isometric—i.e., the x, y, and z axes are not necessarily oriented 120° to each other. Instead, a variety of angles occur; some form of parallel projection, such as dimetric projection with a 2:1 pixel ratio, is the most common."

https://en.m.wikipedia.org/wiki/Isometr ... e_graphics

[Joefish]

The Wikipedia page on the isometric view in game is utter rubbish.

[Einar Saukas]

Not in the way you describe. The Wikipedia definition of oblique projection does not refer to angles relative to a reference outside the projection.

From the Wikipedia page about oblique projection:

"The term cabinet projection stems from its use in illustrations by the furniture industry. Like cavalier perspective, one face of the projected object is parallel to the viewing plane, and the third axis is projected as going off at an angle"

"In the military projection, the angles of the x and z-axis and y and z -axis are at 45°, meaning that the angle between the x-axis and the y-axis is 90°. That is, the xz-plane is not skewed. It is rotated over 45°, though."

It seems all of them are defined relative to the viewing reference axes.

If you sit at 45 degrees to the paper (or turn the paper 45 degrees), "straight" lines become "diagonal" and vice versa.

Rotating images may not look different on paper, but it looks completely different on screen, especially using low-res graphics.

My point is, W*H*B graphics approach is similar to "isometric-like" games such as Ragnar, but very different from "cabinet-like" games such as Ball Breaker and Sophistry. It makes sense to put the first 2 games in a group and the last 2 games in another group.

IMHO our goal should be finding definitions that best fit these groups that make sense for us, instead of the other way around.

Based on everything discussed so far, I think "Isometric-like 3D graphics" (with a link to this page) and "cabinet-like 3D graphics" (with a link to this page) are the best choices.

[Nienn Heskil]

The Wikipedia page on the isometric view in game is utter rubbish.

And so is this thread

[Rorthron]

From the Wikipedia page about oblique projection:

"The term cabinet projection stems from its use in illustrations by the furniture industry. Like cavalier perspective, one face of the projected object is parallel to the viewing plane, and the third axis is projected as going off at an angle"

"In the military projection, the angles of the x and z-axis and y and z -axis are at 45°, meaning that the angle between the x-axis and the y-axis is 90°. That is, the xz-plane is not skewed. It is rotated over 45°, though."

It seems all of them are defined relative to the viewing reference axes.

The sections you quote are from the definitions of "cabinet projection" and "military projection". Wikipedia makes it very clear they are subcategories of "oblique projection". As I explained in my previous post, you are committing a logical fallacy in inferring about the superset from the subset.

If you sit at 45 degrees to the paper (or turn the paper 45 degrees), "straight" lines become "diagonal" and vice versa.

Rotating images may not look different on paper, but it looks completely different on screen, especially using low-res graphics.

These are terms that apply to paper drawings. If you're applying them in a way that doesn't work on paper, then you're misapplying them.

My point is, W*H*B graphics approach is similar to "isometric-like" games such as Ragnar, but very different from "cabinet-like" games such as Ball Breaker and Sophistry. It makes sense to put the first 2 games in a group and the last 2 games in another group.

Except that W*H*B seems to use oblique projection in my opinion.

IMHO our goal should be finding definitions that best fit these groups that make sense for us, instead of the other way around.

Agreed, but the discussion arose out of a desire to use more accurate language. I am not sure I see the benefit of swapping one inaccurate term for another.

Based on everything discussed so far, I think "Isometric-like 3D graphics" (with a link to this page) and "cabinet-like 3D graphics" (with a link to this page) are the best choices.

"Oblique projection" seems a much more accurate (and elegant) term than "cabinet-like". Why not just use the proper name for the category, rather than using the name of a particular subtype and adding "-like"? If we were categorising different Spectrums, we might say 16K, 48K and 128K are all types of "ZX Spectrum". We wouldn't call a 48K Spectrum a type of "16K-like ZX Spectrum". Personally I think "cabinet-like" is horrible!

But it's not worth fighting over.

[ZXDunny]

I can't believe we're going to effectively prevent anyone from searching for a 3D game in the database because they're not design nerds and can't remember the precise term for what they see in their fuzzy memory of a game they last played in 1986.

[Vampyre]

I can't believe we're going to effectively prevent anyone from searching for a 3D game in the database because they're not design nerds and can't remember the precise term for what they see in their fuzzy memory of a game they last played in 1986.

Completely agree.

[4thRock]

As ZXDunny points out, we should use common search terms.
Most of them are not scientifically correct but that's what people know. And they relate to gameplay style, not graphics.
For example we have Arcade games that only came out for the Spectrum. People understand it refers to gameplay

The actual graphic style can be mentioned on the comments field.
Seems like the best place for an expanded analysis. I'd welcome that.

[MatGubbins]

it's a first person 2d isometric shoot-em up text adventure utility