Project D3D11Client Development

So my apologizes to you guys with slower systems, but it's time for this client to move on...
First you tell people with DX10 cards, that they will be supported by the client, and then you say that you (most likely) drop support when there wasn't released a stable version yet. And don't tell me my GTX 285 (DX10) is slow. GT 520/530/545 (DX11) are much slower. :dry:

D3D10Client anyone?
 
First you tell people with DX10 cards, that they will be supported by the client, and then you say that you (most likely) drop support when there wasn't released a stable version yet. And don't tell me my GTX 285 (DX10) is slow. GT 520/530/545 (DX11) are much slower. :dry:

D3D10Client anyone?
Hold on - I didn't tell anything about dropping DX10 support! It still will be supported as soon as their performance is good enough. Infact I've put A LOT of effort in ensuring that new pixel shader-based HDR engine will work on DX10-level hardware. I was only talking about Intel integrated graphics, and that is because they are just way too weak, but not because they are not DX11-compliant....

Again, let me stress this out - DX10 WILL be supported by the client, albeit sometimes with a bit degraded functionality (i.e. less detailed terrain - nothing can replace hardware tesselation). Infact I'm planning to invite someone with good DX10-level video card into closed alpha test so we could ensure that the client works fine not just with latest-and-greatest, but also with greatest-from-near-past hardware.
 
First you tell people with DX10 cards, that they will be supported by the client, and then you say that you (most likely) drop support when there wasn't released a stable version yet. And don't tell me my GTX 285 (DX10) is slow. GT 520/530/545 (DX11) are much slower. :dry:

D3D10Client anyone?

You are not really calling the Fermi cards slow are you? The 285 is a great card. I had one, and love it, and got all I could out of it, but when I jumped up to a Fermi, it blew the 285 out of the water.
 
We have all spent pretty good money on our video cards. My 4770 is DX 10.1 yet it just "worked" for me and I saw little need to consider replacement until recently. Yet I don't expect game devs to hold off for my sake.

Besides I believe we were referring to weak onboard graphics. DX11 fallback modes give alot of leeway on older powerful cards yet for onboard it will be a mess.
 
The latest version lacks binary, I will build and upload it tonight.

As for integrated video cards, I'm afraid current features and effects bring client beyond their performance envelope, so I might be forced to drop that support... Will see what I can do, but if this support will preclude some important features, I will drop that. As I've mentioned before, D3D11Client is supposed to be at cutting edge, and inegrated videos aren't quite there... At the end of the day, you can't play Crysis on integrated video, can you? :)

Of course that drop won't mean, that I will take measures to specifically prevent the client to be run on that HW. It just will be sloooow...

Ah well, I guess I'll just have to wait until I get Win7 on my main rig, then.
 
BTW just an apology. What I stated earlier was as indicated an opinion but I obviously don't want people to think I was saying DX10 support was out.

My opinion on DX10 support is that an addon dev is most likely going to jump to DX11 from DX9 so I personally don't see the point. However that is my opinion only and Asmi obviously has a far better vantage point over the entire situation.

I called for DX10 support back in the day because it was the most realistic path in my opinion. And that the limited OVP development resources of the time were better dedicated to that. However, Today in the situation where DX11 and DX9 development are proceeding quite rapidly and DX11 hardware quite cheap in my opinion DX10 becomes quite the lesser priority. And I believe this will seriously manifest itself in addon development.
 
And that the limited OVP development resources of the time were better dedicated to that. However, Today in the situation where DX11 and DX9 development are proceeding quite rapidly and DX11 hardware quite cheap in my opinion DX10 becomes quite the lesser priority. And I believe this will seriously manifest itself in addon development.

The emphasized term above made me laugh. It sounds like developers are resources. We are people.
Disclaimer: I know you didn't really meant it this way, but I just wanted to point it out.

On the subject of what direction D3D11Client development should go: it is up to the folks that actually code it. Those who do not contribute (like me) can suggest something, but that's it.

However, in this project (OVP), everyone has the opportunity to fork whatever code base he want and change it to his liking. There is no "official" team for it or somesuch. I have the strong feeling that this is the notion some people have in the community, like e.g. "DX9 is done by jarmo, OpenGL is done by Artlav, and now we have DX11 assigned to asmi, don't you dare stepping on one of those toes!". This is simply not the case.

Whoever is interested in coding on OVP clients: please do so! Now! Yes, you can do it without much hassle, no bureaucracy involved. Don't ask for permission, it is GPL. If you think D3D11Client should focus on DX10 and you are sure about it because you already know how to do it, please download the open sourced code base and hack away...

You may wonder why I'm thinking like that, when it all seems so obvious. If you check the beginning of this thread, you'll notice that D3D11Client's current lead developer was unsure how to contribute his work. Just imagine what we'd have missed if he never decided to join in here.

OK, enough preaching for today.

regards,
Face
 
And don't tell me my GTX 285 (DX10) is slow. GT 520/530/545 (DX11) are much slower. :dry:

So.. wait. It means my GeForce 9800gt doesn't support DX 11 too:blink:? Awww :compbash: But I think it would support DX 11 functions. I was playing Battlefield 3 on it, and all features such as tesselation worked(but at 15 fps).

How much would be tesselation accurate in Orbiter? Is it officially confirmed that new Orbiter will support terrain collision detection? Would 3D terrain include oceans too? What about FSX like trees, buildings, cities? Clouds?
 
No offense but did he not say "high end?" That means he is thinking pretty far ahead of the 9800GT.

As for how important DX11 tessellation is. In the Orbiter dynamics topic I believe he stated that they have gotten the resolution rendered down to a level of a tenth of a meter

They have blown past the limits of 32 bit memory...

Quite important I think :)
 
No offense but did he not say "high end?" That means he is thinking pretty far ahead of the 9800GT.

As for how important DX11 tessellation is. In the Orbiter dynamics topic I believe he stated that they have gotten the resolution rendered down to a level of a tenth of a meter

They have blown past the limits of 32 bit memory...

Quite important I think :)


High end is quite a bit beyond the 9800. Think 570Ti and above. Not saying that is what will be required, but that is probably what you should shoot for. Look over at AVSIM and see what all those guys are using for FSX rigs. That is what high end is.
 
And usually when so much data for terrain and texture work is involved. Having a great deal of fast memory is quite important to avoid one hell of a bottleneck.

Disclaimer: I actually have no idea how they are handling GPU memory management. My statement is just a general one on the game industry at large.

In said situations my 512 meg 4770 (120bit GDDR5) and 9800s (256bit GDDR3) will choke. (Saw it happen in X-plane 9 & 10 ) Compared to said GTX 570 above with 1280 megs 320 bit GDDR5.
 
Guys, please calm down. I've promised that DX10 will be supported, and so it will be. I'm trying to write code in a way that it would work on DX10, but I don't actually test it until I stabilize it on DX11 - simply because my video card is DX11-level. But there won't be a public beta release until I ensure it works on DX10 - I've got 275 GTX in my old desktop so I have ability to test it, it's just "first things first" - first I get it to work at least on a DX11, then I move on to DX10.

Now, back on topic - I've made entire engine HDR-aware, also tuned exhaust and base lights a bit - Cras please take latest and test.
 
Last edited:
Give me a sec to get the screenshots up, but in short, the white washed Earth is solved. Gonna need to tweak the settings for the HDR stuff to get it right, but gotta love where this project is heading with all this stuff. If you are able to get the "planet shine" to work, with that beautiful glow from the atmosphere, it is going to be be awful good and realistic to look at.
 
Give me a sec to get the screenshots up, but in short, the white washed Earth is solved. Gonna need to tweak the settings for the HDR stuff to get it right, but gotta love where this project is heading with all this stuff. If you are able to get the "planet shine" to work, with that beautiful glow from the atmosphere, it is going to be be awful good and realistic to look at.
Forgot to mention - pls go to client settings and re-select HDR engine. I've removed "No HDR" option so the order of engines has changed. And if there is anything wrong with the image (which is kinda possible since there were a lot of changes) - pls post screenshots so I could take a look & fix.
 
here are the pics.

DGIV standing at KSC near sunset
https://dl.dropbox.com/u/68333863/TerrainTest/008.PNG

As you can see, that white washed Earth is now gone, looks much better
https://dl.dropbox.com/u/68333863/TerrainTest/009.PNG

The Sun is just HUGE when on the surface of the Earth. It looks like it is going Nova.

https://dl.dropbox.com/u/68333863/TerrainTest/010.PNG
https://dl.dropbox.com/u/68333863/TerrainTest/011.PNG

Some on the ride up to orbit
https://dl.dropbox.com/u/68333863/TerrainTest/012.PNG
https://dl.dropbox.com/u/68333863/TerrainTest/013.PNG

a strange effect with the horizon when it is on the dark side, as you can see how it goes really hard and looses that atmospheric haze
https://dl.dropbox.com/u/68333863/TerrainTest/014.PNG

And some shots of the stock DG flying around Brighton Beach

https://dl.dropbox.com/u/68333863/TerrainTest/015.PNG
https://dl.dropbox.com/u/68333863/TerrainTest/016.PNG
https://dl.dropbox.com/u/68333863/TerrainTest/017.PNG
 
The Sun is just HUGE when on the surface of the Earth. It looks like it is going Nova.
I've made some tunes - please download latest and double-check. Also check this horizon issue.
 
alright, took some new pics with that lasted build.

Still not there yet. The best way I can describe it really is that it just doesnt look right. I am not sure if it is just too intense, or what, but looking at the first pics of the DGIV sitting on the SLF runway, the atmosphere just looks alien. There is this green, the Sun is still quite big, yet it is only the bloom. The Sun itself is actually very hard to find, because it gets dimmed out. Also, the top of the sky goes very dark. It just doesnt look natural yet, some further tuning is in order in my opinion.

https://dl.dropbox.com/u/68333863/TerrainTest/018.PNG
https://dl.dropbox.com/u/68333863/TerrainTest/021.png
https://dl.dropbox.com/u/68333863/TerrainTest/022.png

The horizon in orbit is also very bright, and I think that is what causes that dark horizon thing to show up, that the Earth gets too much light on that end. Based on the D3D9 client, which has the shadow darken out the earth well before the horizon in the same situation, the light bloom lights up that area, and makes that hard horizon appear.

https://dl.dropbox.com/u/68333863/TerrainTest/024.png
https://dl.dropbox.com/u/68333863/TerrainTest/025.png

another with a really bright Earth, https://dl.dropbox.com/u/68333863/TerrainTest/023.png
 
Yes, the Earth views are a little strange...

But the Moon pics posted above are amazing. That terrain associated with collision detection would justify a "high end" graphic card IMHO :P

(and what is "high end" today will be outdated in 2 years anyways ;))
 
Back
Top