Test project

[erico]

A couple years ago, there was a job we did here that required many 3d car models, the agency simply gave us money to buy them from what was available  instead of modeling them ourselves, which would sure cost more and take more time. Within the lib we acquires, which I just remembered to have them here, I thought to make a couple sprite renders should you like to test.

Here they are, sprites are 214x90 in size, I dumped an image from one of your videos to try and match the perspective.

Personally, I find them very ugly and I particularly dislike in rage pre-render gfx as it was used on PC throughout the 90ties.
They seem so out of life to me, but maybe it is my formation with lotus, jaguar, crazy cars 3 and xtreme racing yelling.
My dump opinion aside, see if they can serve your testings better than the ball.

[MrTAToad]

The minimum resolution is 640 x 480, and will scale up as resolution increases.

At the moment, every status display fits nicely in the base resolution.

All graphics have to be done using polyvector as the sprites colour can be changed (especially the players one in a tunnel), so a loading system (that can load single sprites, animated sprites in multiple files, animated sprites of fixed cell size and possibly animated sprites who's size and position arent fixed need to be loaded and displayed.  Then I can convert the players display function to use polyvector and see how well it animates.

Polyvectors have to be used for clipping and scaling (I soon remembered that the function can do that after posting and wanting a clipping function elsewhere), so a graphics processing type needs to be developed now.  I've started on the loading system but nothing has been tested yet.

[MrTAToad]

A couple years ago, there was a job we did here that required many 3d car models, the agency simply gave us money to buy them from what was available  instead of modeling them ourselves, which would sure cost more and take more time. Within the lib we acquires, which I just remembered to have them here, I thought to make a couple sprite renders should you like to test.

Here they are, sprites are 214x90 in size, I dumped an image from one of your videos to try and match the perspective.

Personally, I find them very ugly and I particularly dislike in rage pre-render gfx as it was used on PC throughout the 90ties.
They seem so out of life to me, but maybe it is my formation with lotus, jaguar, crazy cars 3 and xtreme racing yelling.
My dump opinion aside, see if they can serve your testings better than the ball.

I do like them - do you have any more of these?

[erico]

There is about 200 cars, if not way more, on many different 3d complexity levels.
Some are useless, some are really pro and have billions of points, I got 2 examples that are more or less low poly.
I never went through all of them myself.

The 2 pics I did is a result of a simple scene made to render the rotation on that perspective.
The actual render was 2x bigger, I scaled it down on photoshop.
I also had to clean up a bit of the surfaces, like the main color, reflections and specially the glass.

There isen´t much of photography going on at them as that takes time and the way they are surfaced is really crap.
The convertible ones also would need a human inside.

Glad you like them, I prefer the low poly ones from the pages before.

[matchy]

No more prototype ball vehicle because those test car tiled sprites are perfect. 

[MrTAToad]

There is about 200 cars, if not way more, on many different 3d complexity levels.
Some are useless, some are really pro and have billions of points, I got 2 examples that are more or less low poly.
I never went through all of them myself.

The 2 pics I did is a result of a simple scene made to render the rotation on that perspective.
The actual render was 2x bigger, I scaled it down on photoshop.
I also had to clean up a bit of the surfaces, like the main color, reflections and specially the glass.

There isen´t much of photography going on at them as that takes time and the way they are surfaced is really crap.
The convertible ones also would need a human inside.

Glad you like them, I prefer the low poly ones from the pages before.

Could you make them available or is their copyright status dubious or public use ?

[erico]

The more complex ones I believe are really un usefull to you unless you are really proficient into a 3d package, you would also need a heavy machine to deal with it and lots of time to make it work out. These I´m not sure what the copyright stated when we bought, I´m not sure it was a buyout or that only those who actually acquired it are entitled to use.

The more low poly ones, like the couple I used here, are ok to distribute but not ok for commercial use and are from different artists.
These we got as references and placeholders, I think there should be no problem sending them to you.

They are about 120 files in lwo format (lightwave).
I think I also have some libs from VUE and other 3d softwares we acquired, but I´d have to search and check.

I thought I had your email around here, pm me it and I send you the models I believe are ok.

You can also check the mesh database to see if there is anything interesting (there should).
https://www.rocq.inria.fr/gamma/gamma/Database/index.fr.html

[matchy]

There is about 200 cars, if not way more, on many different 3d complexity levels.
...

Could you make them available or is their copyright status dubious or public use ?

Quite relevant for test project, not. 

[MrTAToad]

Its always good to plan ahead...

[matchy]

Understandably as I have a similar issue with content vs functionality. To foresee, perhaps creating these in run-time from base 3d models to produce 2d sprites (ie. vector splicing) or even over-layer 3D.

For now, something more sensible in the next test video would be nice and I'm hoping as I'm just not a fan of the racing ball/sphere.

How important is it we worry about brand name cars in our games?

[erico]

If I understood correctly the content vs functionality, yep it always a chicken egg thing since some gfx can impose on code, like fine tunning the perspective and sometimes the other way around.

Lately I have been doing a lot of low res black and white graphics as place holders since they do hold a bit of the final idea´s intentions, I´m growing really found of them.

[matchy]

erico, Have you tried to produce animsprites with 3D to 2d? Do you have 3d models of cars, bikes or vehicles just to test with? I'd like to reconstruct one but need a frame to design by I guess. 
The vehicle overlay function is really separate to how the track works and is displayed, even lighting. Presumably there are only left and right turning frames like in your tile-set, but maybe also jump and bump (yaw and pitch frames). Alternatively rather than a bitmaps, a smooth turning 3D motorcycle like Road Rash 3D would compliment the track more. (but then there's the rider animation).

[erico]

I never tried to produce 2d out of GLB´s 3d, I have very little to no experience in 3d with GLB as of yet.
It is easier for me to render in Lightwave and export to Photoshop.

When I did the settings up there, I first thought to output car´s full rotation and some flipping for crashes, as well as a bunch of sets for going up and down the road, but at a resolution above the classic 320x200, the amount of frames needed for a simple car turn goes really high. That is why I thought about doing low poly 3d pixel art cars in the first place, to get rid of 300+ sprites per car. In fact, the 3d low poly cars works ace if your game is low res to boot.

I then took some time to examine my fav racer of old age and checked how they did it, specially on a time that scaling sprites was not common, it is a lot of work or you get hard restrictions with high amount of sprites. I still believe using 3d to try to display 2d styled cars is the best compromise, a 200km/h crash that simply slows you down does not impress on current age, and a flipping crash like outrun on higher res then 320 with just a couple of frames, going over a road that perfectly smoothly moves about, sounds too out of place.

I think what limits most the use of 2d is to want to have anything higher res.
Yep, the rider´s animation is a very complex job, you will need some sort of rigging to boot.
I can send you some models should you want, pm me your email.

edit: just noticed that the beadle up there looks strange with such thin tires, that would need some editing so that it feels more in place.

[MrTAToad]

Understandably as I have a similar issue with content vs functionality. To foresee, perhaps creating these in run-time from base 3d models to produce 2d sprites (ie. vector splicing) or even over-layer 3D.

For now, something more sensible in the next test video would be nice and I'm hoping as I'm just not a fan of the racing ball/sphere.

How important is it we worry about brand name cars in our games?

Better not get your hopes up then!  I've got a partially animated ball that I originally planned to use (at the moment I've only processed 3 frames as I need to make sure my graphics system works), although if there are some cars that can be used (the problem will be animation), I may use them.  Whether I will keep it is another matter, but the important thing is to make sure the animation system works...

The reason for using a ball is you can jump (its slow at the moment), and would be odd with a car 

I've started intergrating the graphics system into the program - the first thing to do is update the static object type to have animation data in it...

[erico]

I was wondering, if a truly retro racer is what you are looking for, you could work with a base resolution of 320x200 fat pixel, like amstrad and c64, so actually 160x200. Probably use a 16 color palette preferably Dawbringer´s and no transparencies.

I came across this mock up by José Pereira and added my renders just to check, with a little clean up it could work nicely.
At this res, you would need at the most, 7 sprites to make the car turn so there is a lot of reduction on the art workload.

Gotta love the low res.

Lately I stumbled by that amiga racer, http://distrita.com/amiga-racernews/
and while I don´t want to criticize that nice effort, I fell this is a perfect example of mixed styles that are fully incoherent, and therefore looks very ugly on my books. It is what we (me, Ian) were discussing on the other job thread.

A truly retro racer, with those aesthetics, and improving/updating on the formula would be just perfect for me.