New Release D3D9Client Development

[jarmonik]

Is there any work on going to improve the atmosphere rendering?

Unfortunately no. The atmosphere rendering should be complitely reimplemented. The current one is a very simpple physics model that's been pushed to fovor orbital rendering with the cost of twilight/sunset/rise conditions.
Current focus is pretty much in a development tools regarding surface bases, virtual cockpit lighting, mesh utilities. But, of course, a new atmospheric model would be important.

 

[cristiapi]

Older version

I installed the version 25-Feb-2019, but it's not good for my hardware. Same problem with D3D9ClientR3.5-forOrbiter2016(r1061).zip.
Is there an even older version? Probably I need a very old version, but I forgot the working version.
I don't need the most up to date version, I just need a reasonably stable version.

EDIT: I found my post #4176 on Sep 2016, hence a version published at that time should be fine.

 

[kuddel]

I do not "collect" old versions, as we are pretty happy when those "extinct"
The hardware requirements did not change much over the whole project lifetime (as far as I'm aware), so I guess you might just miss some (VC-) runtime(s)...

I think VS 2012 redist (32/64bit?) is needed, but you might have to try others (2015, or 2013)[*]

If you really want to take the shot at an "old version" I would step back in this thread from the date you've mentioned until you find a post by Jarmo or me that had a ZIP attached.

Out of curiosity:
- Do you have an XP machine?
- Did you re-install your System? You might have forgotten to install the DirectX 9 runtimes...(happened to me a couple of weeks ago too, thinking: "Why doesn't this start at all?" :facepalm
- Does the log (Orbiter.log and/or Modules\D3D9Client\D3D9ClientLog.html) tell anything?


[*] For all those runtimes: All in One Runtimes worked fine for me more than once.

 

[Sbb1413]

I installed the version 25-Feb-2019, but it's not good for my hardware. Same problem with D3D9ClientR3.5-forOrbiter2016(r1061).zip.
Is there an even older version? Probably I need a very old version, but I forgot the working version.
I don't need the most up to date version, I just need a reasonably stable version.

EDIT: I found my post #4176 on Sep 2016, hence a version published at that time should be fine.

Why need D3D9 on Orbiter 2016 at all? I am not using D3D9 on Orbiter 2016 but on Orbiter 2010.

 

[cristiapi]

If you really want to take the shot at an "old version" I would step back in this thread from the date you've mentioned until you find a post by Jarmo or me that had a ZIP attached.

Good! I'll do it.

Out of curiosity:
- Do you have an XP machine?
- Did you re-install your System? You might have forgotten to install the DirectX 9 runtimes...(happened to me a couple of weeks ago too, thinking: "Why doesn't this start at all?" :facepalm
- Does the log (Orbiter.log and/or Modules\D3D9Client\D3D9ClientLog.html) tell anything?

Windows 7 64 bit. No significant changes since 2016. The problem is that with the new D3D9 version I see the FPS display (RivaTunerStatisticsServer) flashing near the center of the window, like D3D9 is repeatedly opening and closing a new window. Anyway, is not a problem, an old version will be good for me.

---------- Post added at 09:42 ---------- Previous post was at 09:39 ----------

Why need D3D9 on Orbiter 2016 at all? I am not using D3D9 on Orbiter 2016 but on Orbiter 2010.

Because Orbiter 2016 (and 2010) looks much beautiful and runs much more smoothly.

 

[Marijn]

I am still using D3D9 R2.1 because it seems a more stable version to me. Recently I tried the most recent one. One of the things I ran into was the FPS counter of my screen capture software being displayed in the top-left corner of the right MFD instead of the corner of the screen.

 

[Sbb1413]

Because Orbiter 2016 (and 2010) looks much beautiful and runs much more smoothly.

My FPS was reducing when I was using D3D9 on Orbiter 2016. Since then, I am using the native client and works smoothly. Although, I am using D3D9 R17 on Orbiter 2010.

 

[cristiapi]

I see the opposite with my hardware, but I have an old GTX 660 Ti. I also use a fps limiter (200 fps).

 

[jarmonik]

Windows 7 64 bit. No significant changes since 2016. The problem is that with the new D3D9 version I see the FPS display (RivaTunerStatisticsServer) flashing near the center of the window, like D3D9 is repeatedly opening and closing a new window. Anyway, is not a problem, an old version will be good for me.

I am still using D3D9 R2.1 because it seems a more stable version to me. Recently I tried the most recent one. One of the things I ran into was the FPS counter of my screen capture software being displayed in the top-left corner of the right MFD instead of the corner of the screen.

I have collected a list of older builds from R2.1 to R3.4 into the first post. What is the latest build that works properly ?


Also if you change "PresentLocation" parameter from "1" to "0" from D3D9Client.cfg, does it make any difference in R3.9 release ? Also have you tried without post processing effects ?

---------- Post added at 21:06 ---------- Previous post was at 16:56 ----------

There is a new build published in an attempt to address the "framerate display" issue. If the problem still exists then reply to the questions above.

 

[Marijn]

I'll check R3.9 and report back. Thanks for your efforts.

Edit: R3.10 I see

 

[Abloheet]

My FPS was reducing when I was using D3D9 on Orbiter 2016. Since then, I am using the native client and works smoothly. Although, I am using D3D9 R17 on Orbiter 2010.

What graphics hardware and system configuration do you have?
I get 20-50fps using built-in dx7 client in Orbiter 2016, and 100-500 fps using dx9 client.

Dx9 client is better than built in dx7 in almost every way, only missing features being solar-arrays, monorail and hangrail base objects, and dropped support for rendering irregularly shaped celestial objects using .msh files

 

[Donamy]

My specs are i5 6500 GTX 970

Full ISS

 

[cristiapi]

There is a new build published in an attempt to address the "framerate display" issue. If the problem still exists then reply to the questions above.

I downloaded the 6-Sep-2019 version; same problem.

I downloaded a very old version (2016) as suggested by Kuddel; problem solved, for me.

 

[DaveS]

Unfortunately no. The atmosphere rendering should be complitely reimplemented. The current one is a very simpple physics model that's been pushed to fovor orbital rendering with the cost of twilight/sunset/rise conditions.
Current focus is pretty much in a development tools regarding surface bases, virtual cockpit lighting, mesh utilities. But, of course, a new atmospheric model would be important.

First of all, a belated thank you for this reply. Another thing that should be on that list is the implementation of a more realistic handling of reflections. With that I mean that there should be several "environment cams" and not just one so you get proper reflections.
I've attached a screenshot that shows the limitations of the current implementation of a single "environment cam". The green circles shows what shouldn't be seen from that particular angle (part of the payload bay) and the red circles show broken/interrupted reflections of the vertical stabilizer in the two forward radiator panels.

 

[kuddel]

I downloaded the 6-Sep-2019 version; same problem.

I downloaded a very old version (2016) as suggested by Kuddel; problem solved, for me.

Can you post the exact version that is working?
Out of curiosity I would like to check whether that version was possibly the last version without SSE/SSE2 enabled by default...(just a guess)

 

[cristiapi]

I see "R1 [Aug 30 2016]".

 

[kuddel]

I see "R1 [Aug 30 2016]".

Thanks for the info, unfortunately that didn't shed any light onto your issue...
That build already had the SSE2 enhanced instruction-set enabled, so that's not responsible.

cristiapi, would you mind running a release one version younger than the working one?
That one should not work, but I hope it generates some log data (Orbiter.log and/or Modules\D3D9Client\D3D9ClientLog.html).
If you would be so kind and post those "crash"-logs (as a ZIP preferably).

 

[jarmonik]

New binary for testing the recorder FPS issue

Here's a new binary for testing the reported issue with screen recording utilities. It's a DLL only to be intalled on a top of 3.10 in /Modules/Plugin/

 

[jarmonik]

Thanks for the info, unfortunately that didn't shed any light onto your issue...

I have a feeling that the issue could be related to a commit r.1019 in which case 3.2 should work and 3.4+ not but that's just my best guess. If that's the case then it should be fixed from the package/attachment above. I am not really expecting anything to show-up in the logs but who knows.

---------- Post added at 20:57 ---------- Previous post was at 20:51 ----------

With that I mean that there should be several "environment cams" and not just one so you get proper reflections.

I have some plans for an advanced mesh editor and it would also include the possibility to setup and configure multible "env cams" to improve the reflections. It's a bit difficult to configure them properly without a proper editor application.

 

[kuddel]

I assume this one above (post #5003) is for Orbiter 2016, right?