Hello

We are thinking with our team of making "2.5D" point and click adventure game (for PC mainly). 2.5D means pre rendered 3D scenes (so they are just 2D background) (created mostly in 3ds max) but with real time 3D characters ("low poly" around 15K-20K polygons?) moving on the scenes.

I would like to know how much additional work would it be to create a full real time 3D scene instead of just pre-rendered 3D background?

Like pre-render game like Black Mirror 3 in comparison with games like the company Tell Tale does or how the game "Life is Strange" was.

I guess i mean how much additional time would it take to "optimize" the original hipoly scene for 3D realtime game use- that means- create low poly versions of objects, proper unwrapping (not just UVWing)+; baking maps (normals etc.), maybe create LODs etc.

How much additional work would it take? If the hi poly scene creation takes 100% of time, what would take to "optimize" it for real time useage- 150% of time?

ironically it will take you the same time to make pre-rendered scene as it will take to make a full game ready scene.

For a game ready scene the flow will be like this:

Tessellation mapping. -> quick block model -> High model -> Low model -> UV unwrap -> Baking textures -> Texturing = Low poly model -> Build level using low poly models

For a pre-rendered scene it's like this:

Tessellation mapping. -> quick block model -> High model -> UV unwrap -> Texturing = High poly model -> build level with high poly models -> render -> Map level to a plane.

As a rule of thumb it will take your 3D modeler a day to make a model. Things like chairs and desks will be fast, with the 3D modeler making several in a day other models will take much longer with the most complex model taking four days.

Once your art style is clear to the modeler they will know where time can be saved.

In the end it will depend on what you want, pre-render scenes can be rendered at a high quality except you lose in effects like shadows and has to be viewed from mostly fixed angles. Game ready scenes can be viewed from anywhere and is easier to do collisions and most game engines is optimized for it, shadows and lighting will work as it does in most games.

The larges deciding factor here will be what style you want, also if you plan on using a PBR engine like Unreal you could build very realistic game scenes.

No i dont really thinks so that you are correct, so i will rephrase my question:

Lets say that i was doing a scene (a room(a "rich room" with plenty of different objects- easy and quite diffucult to model)) that should have been used only for render videos, not ingame real time gameplay/engine running. So everything is quite high poly, no Lods, created, not a lof of unwrapped objects (just simple UWV mapping) etc. Simple a scene that supposed to be rendered into a rendered video on some render farm. Lets say i already spent 30 hours doing it. Now the art director or someone comes in and says: "Hey, we decided that besides the render video, we would like to have this room/scene also playable ingame, by the player, real time..." So you have to create now a low (or at least lower) version of the entire scene/room to make it playable in real time on PC (lets say on Unity or Unreal engine). That means creating normal maps, Lods, Lowe resolution of high objects, baking maps, etc. etc.

How much additional time you will have to spent on it? (besides those 30 hours already putted in). Additional 10 hours? Give me your best experienced guess...

(Thank you, it will help me a lot)

How much additional time you will have to spent on it? (besides those 30 hours already putted in). Additional 10 hours? Give me your best experienced guess...

I will admit, having 3D objects that you need turned into game ready models is faster, however not by so much as you will think. In truth it only saves on the designing and planing steps.

With out seeing the scene I will say a 30 hour scene(Four working days) will take 16-18 hours more(Two working days).

However making the scene from the beginning to be game ready will take about 30-39 hours(Four to five working days).

When you think of a 3D model and how long it will take to make it, that time is almost always wrong. When ever a client hires me I double the time I think it will take and there where times where I had to work a extra day or two.

Look, if you want to confirm the time it will take do the following:

Build a 3D chair for a game with textures and every thing, note down the time it took you to do it. Next delete that chair then build a new one for render, when the new chair looks acceptable when rendering note that time.

Last, turn your ready to render chair into a game ready model and note down the time.

Doing it yourself is the only way you will really get a feel for the time.

How much additional time you will have to spent on it? (besides those 30 hours already putted in). Additional 10 hours? Give me your best experienced guess...

I will admit, having 3D objects that you need turned into game ready models is faster, however not by so much as you will think. In truth it only saves on the designing and planing steps.

With out seeing the scene I will say a 30 hour scene(Four working days) will take 16-18 hours more(Two working days).

However making the scene from the beginning to be game ready will take about 30-39 hours(Four to five working days).

When you think of a 3D model and how long it will take to make it, that time is almost always wrong. When ever a client hires me I double the time I think it will take and there where times where I had to work a extra day or two.

Look, if you want to confirm the time it will take do the following:

Build a 3D chair for a game with textures and every thing, note down the time it took you to do it. Next delete that chair then build a new one for render, when the new chair looks acceptable when rendering note that time.

Last, turn your ready to render chair into a game ready model and note down the time.

Doing it yourself is the only way you will really get a feel for the time.

Thank for your reply, it is actually helpfull (im not used to received concrete helpfull answer on this forum, so your response was a nice suprise :-).

And yeah, some personal test would be the best aproach.

Thanks again for reply.

If anyone else has an experienced gues how much time would it take, according to his/her own experience, step forward :-)!

Thank for your reply, it is actually helpfull (im not used to received concrete helpfull answer on this forum, so your response was a nice suprise :-).

This is only because you are still a bit inexperienced, even if you don't understand what people are saying now as you progress it will all make sense.

If anyone else has an experienced gues how much time would it take, according to his/her own experience, step forward :-)!

I am taking a few days of to work on a portfolio piece, I could record my working times for you if you are interested.

Thank for your reply, it is actually helpfull (im not used to received concrete helpfull answer on this forum, so your response was a nice suprise :-).

This is only because you are still a bit inexperienced, even if you don't understand what people are saying now as you progress it will all make sense.

If anyone else has an experienced gues how much time would it take, according to his/her own experience, step forward :-)!

I am taking a few days of to work on a portfolio piece, I could record my working times for you if you are interested.

No its not because of inexperience, its because are generally "stupid" or unwilling to provide any numbers/estimations even if you descrivbe the situation/problem quite good and clearly :-) and clearly state that you want just some ROUGH estimation, theyll spent instead 2 pages arguing that "it depends" instead of providing one post with number estimation :-). I get this problem a lot. For example i found out that most usefull forum on graphic design is http://polycount.com. Short, straight, to the point and HELPFULL advices. Not like on other forums where the answers are just stupid, of the point, obvious advices, offtopic, or fighting.

You provided straight meaningfull information, thanks for that :-).

If you are doing some high poly model with low poly version than yes, why not, if you could bother i would like to see some real world numbers.

By the way, for time recording/measurement i love to use "Free Stopwatch".