Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

FFE D3D

43 views
Skip to first unread message

misiek82

unread,
Jul 4, 2006, 1:38:57 PM7/4/06
to
Hello,

Does anyone knows something about FFE D3D port? When it's gonna be ready?
Are there any compiled versions yet?

I don't know russian, so I can't understand anything what they're writting
on elite-games.ru forum :P Just watching screenshots ;)

Thanks

[jongware]

unread,
Jul 4, 2006, 4:41:57 PM7/4/06
to
"misiek82" <mg82.wy...@o2.pl> wrote in message
news:8nlpm0cilopm.vc8txc4z9p1h$.dlg@40tude.net...

DreamZzz keeps slugging code ... He's coagulating a small pack of likeminded
coders around him (all Russians) and set up a Subversion source site as well
as a discussion forum -- that's last week's news. Recent discussions on
elite-games.ru were on who would be the most appropriate person to approach
the Hon. John Jordan (presumably to clarify a particularly nasty piece of
Frontier code) and how to get rid of DirectX related drawing artefacts.
There are a couple of binaries roaming the 'web, but they are playable
alpha's rather than the colourful and speedy software of our dreams. As yet.
But the transparent atmosphere around the fully-textured planet already is
quite impressive...
Ah -- I *hate* to plug this. www.altavista.com/babelfish. There. It surely
helps if you are at least acquainted with the local vernacular and can
forgive a lot that is Lost in Translation.

[Jongware]


John Jordan

unread,
Jul 4, 2006, 9:11:06 PM7/4/06
to
[jongware] wrote:
>
> DreamZzz keeps slugging code ... He's coagulating a small pack of likeminded
> coders around him (all Russians) and set up a Subversion source site as well
> as a discussion forum -- that's last week's news. Recent discussions on
> elite-games.ru were on who would be the most appropriate person to approach
> the Hon. John Jordan (presumably to clarify a particularly nasty piece of
> Frontier code) and how to get rid of DirectX related drawing artefacts.

I'm not that hard to approach, although I can't see what I'd know that a
whole team of Russians don't. Use IRC rather than email, there's a link
on the JJFFE page somewhere. I can dig up a live translator if necessary.


--
John Jordan

AlexD

unread,
Jul 18, 2006, 12:58:43 AM7/18/06
to

John Jordan wrote, [jongware] wrote:

Stop writing and come in discussion! :) (John Jordan already came :) )

This post is address to jongware. :) We are wating for you. You can
really help us and accelerate development of FFE D3D. On site of John
Jordan you'll find contact (IRC). We are there almost every day (except
rest-days) . Now I trying to assemble something like "troops" on IRC
channel. John Jordan is already there. Anton Lindstrom gave his consent
to.
So, it is almost nothing to do: discuss the problems, find the
solutions and realize it in FFE D3D. :) As easy, as send two bytes... :D

[Jongware]

unread,
Jul 18, 2006, 11:37:29 AM7/18/06
to
"AlexD" <al...@mgpls.ru> wrote that John Jordan wrote, [jongware] wrote:
> Stop writing and come in discussion! :) (John Jordan already came :) )

-- if only the Elite Bar was for real! (Kool krater jooz for everyone!)

> This post is address to jongware. :) We are wating for you. You can
> really help us and accelerate development of FFE D3D. On site of John
> Jordan you'll find contact (IRC). We are there almost every day (except
> rest-days) . Now I trying to assemble something like "troops" on IRC
> channel. John Jordan is already there. Anton Lindstrom gave his consent
> to.

> So, it is almost nothing to do: discuss the problems, find the
> solutions and realize it in FFE D3D. :) As easy, as send two bytes... :D

You have my blessing to use anything you can grab from my site. There. (That
includes the True Love piccher decoder! Cheers all around, no doubt, it's my
#1 google hit page.)
As for tying in to any running discussions... ouch! My recent programming
efforts are fully focused on DirectX 3D -- as far a distance from my trusty
ol' Spectrum as one can imagine -- and personally, I'm not counting on any
breakthroughs there, as even displaying pre-cooked models from
DirectX-written games poses quite the challenge. My programming style is
based more on stubborn perseverence than on higher than average strokes of
genius.
I'll keep an eye on your efforts, though. Every modern hi-rez version is
welcomed by more and more fans, both old and new, which should ensure a nice
and steady buildup of pressure for Braben & co. :)

[Jongware]


AlexD

unread,
Jul 19, 2006, 12:19:08 AM7/19/06
to

[Jongware] wrote:

> You have my blessing to use anything you can grab from my site. There. (That
> includes the True Love piccher decoder! Cheers all around, no doubt, it's my
> #1 google hit page.)

We already use it. :) But, does not care, thanks for your blessing.
:)

> As for tying in to any running discussions... ouch! My recent programming
> efforts are fully focused on DirectX 3D -- as far a distance from my trusty
> ol' Spectrum as one can imagine -- and personally, I'm not counting on any
> breakthroughs there, as even displaying pre-cooked models from
> DirectX-written games poses quite the challenge. My programming style is
> based more on stubborn perseverence than on higher than average strokes of
> genius.
> I'll keep an eye on your efforts, though. Every modern hi-rez version is
> welcomed by more and more fans, both old and new, which should ensure a nice
> and steady buildup of pressure for Braben & co. :)

We are not geniuses to. :) We are working stubborn and hard to. And we
have a greate problem with a planets. It is moving not exactly we
want... It moves 30 times slower than other objects (e.g. starport).
But we are drawing planets excatly the same way as any other objects in
game... So, we stuck... Is there another special algorithm for planets?
Maybe you know, do you?

As for a "pre-cooked models from DirectX-written games" it is only
additional ability of new engine. Now we use it for testing
productivity of engine. In most cases there are self-written
interpreter of FFE models (your site helps very much :) ).

And one more, FFE D3D it is not just "modern hi-rez version". It is
something more. ;) There are many other (not graphic) changes. ;)

Christophe Cavalaria

unread,
Jul 19, 2006, 1:45:35 PM7/19/06
to
AlexD wrote:
> And one more, FFE D3D it is not just "modern hi-rez version". It is
> something more. ;) There are many other (not graphic) changes. ;)

Well, looks good but since it uses the D3D crap, it is worthless for me.
Couldn't you do an OpenGL version instead ? :p

AlexD

unread,
Jul 20, 2006, 12:10:00 AM7/20/06
to

Christophe Cavalaria wrote:

> Couldn't you do an OpenGL version instead ? :p

I think, no... for a while. :)
GL version is develop in other theme. :)

Look this:
http://groups.google.ru/group/alt.fan.elite/browse_thread/thread/aa783685b420cd04/da4d75b23fa410dd

AlexD

unread,
Jul 20, 2006, 12:32:44 AM7/20/06
to

[Jongware] wrote:

> and personally, I'm not counting on any
> breakthroughs there, as even displaying pre-cooked models from
> DirectX-written games poses quite the challenge.


By the way, we need you not like a programer, but like a consultant.
FFE's algorithms and mehods Сonsultant. This is what we need. We'll
very much obliged to you, if you'll answer some our questions...

[Jongware]

unread,
Jul 20, 2006, 4:50:42 AM7/20/06
to
"AlexD" <al...@mgpls.ru> wrote in message
news:1153369964....@p79g2000cwp.googlegroups.com...

[Jongware] wrote:

>> and personally, I'm not counting on any
>> breakthroughs there, as even displaying pre-cooked models from
>> DirectX-written games poses quite the challenge.
>
>By the way, we need you not like a programer, but like a consultant.

>FFE's algorithms and mehods ?onsultant. This is what we need. We'll


>very much obliged to you, if you'll answer some our questions...

The idea of my web pages was to write down everything and anything I know
about the game! The mesh viewer uses a few more tricks but those are just
for display purposes--nothing to do with the actual game.
As for the rest--it appears DreamZzz already knows more of a few parts than
I do, f.e. the planet tesselation. The graphics part of my web pages serve
more as an addendum to the star/planet generator physics than as a separate
project, so I doubt I could help you guys out on that. (And in the real
world I suffer from limited available time, no summer vacation for this guy
this entire decade...)

[jw]


John Jordan

unread,
Jul 21, 2006, 6:11:00 PM7/21/06
to
AlexD wrote:
>
> We are not geniuses to. :) We are working stubborn and hard to. And we
> have a greate problem with a planets. It is moving not exactly we
> want... It moves 30 times slower than other objects (e.g. starport).
> But we are drawing planets excatly the same way as any other objects in
> game... So, we stuck... Is there another special algorithm for planets?

Yes, but that's not why you have the problems. First, let's explain why
the planet moves slower:

FFE's physics code handles object positions in 64-bit integers, with a
precision of 1 unit = 1 millimeter. For speed (ha), for each object the
rendering code scales these down to 32-bit integers, and decreases the
object's scale by the same factor. Drawing an object at "correct"
position and radius is equivalent to drawing it 32 times closer and 32
times smaller. Because these values aren't used for depth sorting in
FFE's renderer, this works fine. Because the positions are scaled down,
far objects such as planets will appear to move slower if you simply
watch the 32-bit position.

In fact, the visible planet-expansion effect in D3DFFE was caused not by
this, but by the additional scale factors applied by the D3DFFE code
for positions >16000.0 (the y-coord of Hope at the default start is
-18529). I assume these were intended to keep objects within the finite
z-buffer. There was no corresponding factor applied to planet radius, so
the planet was dragged visibly closer to the player.

Having disabled the z-buffer code, I started running into FFE bugs. It
turns out that the 32-bit position is only accurate to around 7 bits,
due to terrible matrix orthogonalisation code (see F1519/20). Having
replaced this, I found that the player orientation often flipped upside
down. Turns out that the object rotation code is a very nasty hack (see
F585 docs, cphys.txt) and cannot handle accurate orientation matrices.

After replacing the orientation code, planets (well, Hope - I didn't
bother generalising the radius calc) finally stayed exactly where they
should.


The z-buffer issue is something that has troubled space-sim writers for
as long as I remember working on space sims. For the general case, there
are no easy solutions. In FFE, object radii rarely intersect, so simply
rendering each object in depth order and using a separate chunk of the
z-buffer for each (or clearing, with modern hardware) would work for
most cases.


--
John Jordan

Graham Thurlwell

unread,
Jul 19, 2006, 12:59:18 PM7/19/06
to
On the 18 Jul 2006, "[Jongware]" <sorry@no_spam.plz> wrote:

> "AlexD" <al...@mgpls.ru> wrote that John Jordan wrote, [jongware] wrote:
>> Stop writing and come in discussion! :) (John Jordan already came :) )
>
> -- if only the Elite Bar was for real! (Kool krater jooz for everyone!)

You can't beat Brown, IMHO.

--
Jades' First Encounters Site - http://www.jades.org/ffe.htm
The best Frontier: First Encounters site on the Web.

nos...@jades.org /is/ a real email address!

0 new messages