[Heather Kush]

Hear, hear :whisper:

I guess I can't appreciate it because I am a noob. I wish I was more comfortable with it like you guys. But I think you can understand me, it is really hard to read.(if those were in HLSL, I think previous OBGE team would be successful with their projects.)

@HawkleyFox,
I want to try it, but don't think I can do the implementation myself. (Hehe, my actual goal here was to find recruits for Morrowind/MGE. But it worked the opposite way.)

[laila hassan]

(Hehe, my actual goal here was to find recruits for Morrowind/MGE. But it worked the opposite way.)

Aren't pkeen, phal, hrchIforgottherest and the other folks working on MGE anymore ?

[Jade]

Aren't pkeen, phal, hrchIforgottherest and the other folks working on MGE anymore ?

MORE is better. There is a 6 month gap between revision 118(10-2009) and revision 119-120(4-2010) that made me a little worried. Now the development is faster than ever though with a general optimization. Still I would be glad to see new developers and shader writers for Morrowind.

It appears we will drop SM 2.0 support(which I have no objection at all) and maybe even go for the DX 10 direction. :hubbahubba: So this is time to decide where to go with MGE next, and a great time to shape that direction.

[Gwen]

...
It appears we will drop SM 2.0 support(which I have no objection at all) and maybe even go for the DX 10 direction. :hubbahubba: So this is time to decide where to go with MGE next, and a great time to shape that direction.

Perhaps you can apply some of that MGE brilliance to Oblivion I know that Oblivion's not as well-understood as Morrowind yet, but can you see a future OBGE utility like MGE is now, a nifty external application, that can support DX10, etc.? I've followed MGE since its infancy, even when I'm not playing Morrowind at all, simply because I so admire the project and how it's run. I would love to see OBGE move more toward MGE's model of development and execution

[Kate Schofield]

... a future OBGE utility like MGE is now, a nifty external application ...

That's next on my list.

[Kate Schofield]

That's next on my list.

This is what's going to keep me going until TESV. I should print off pictures of all your avatars and hang them up on my wall, strung together into a banner. Or is that just creepy...?

[marie breen]

As long as they're not photos of real people I wouldn't call it creepy. Hmmm, why did your question make me think of Caldera?

[Emma Louise Adams]

I just installed OBGEv2 and clicked on the item in my miscellaneous inventory. It told me to close the inventory for the OBGE menu to appear, but after i closed it, nothing happened. What am I missing?

[Damned_Queen]

I know this is old news by now, but I just want to mention that depth of field (and I guess probably godrays as well, though I've not tried to enable them at the proper times) are NOT working for me, I guess because of my ATI GPU.

[Danial Zachery]

Could someone provide me a list of what MGE's front-end does ? I haven't got the game on me.

[Lyndsey Bird]

I know this is old news by now, but I just want to mention that depth of field (and I guess probably godrays as well, though I've not tried to enable them at the proper times) are NOT working for me, I guess because of my ATI GPU.

Hmm, I've got an ATI GPU too, so that can't be it. :shrug:

I'm in the middle of a rewrite using some of Crysis' ideas anyway, so we'll see for the next version.

[NAtIVe GOddess]

Yeah, I too noticed that the latest versions of DoF and godrays didn't want to do anything. Could be an issue with what Catalyst versions we're using, perhaps. :shrug:

[carley moss]

Is there a way to force Depth of Field to focus on your character while in 3rd person mode?

Currently my character flips in and out of focus when I crouch or run with a weapon drawn.

[Shianne Donato]

Is there a way to force Depth of Field to focus on your character while in 3rd person mode?

I second that! The feature I've been asking for ever since the original DoF mod.

[Marie Maillos]

Is there a way to force Depth of Field to focus on your character while in 3rd person mode?

Currently my character flips in and out of focus when I crouch or run with a weapon drawn.

I think that may be possible, but it would have to be using that method shadeMe suggested (and was working on?) that used GetCrosshairRef and some Oblivion scripting before passing the distance/position or something on to the shader to do its stuff. Obviously we know what the ref is, but it's a matter of whether we can get the distance to it from the camera, or it's centre position on the screen, or something.

I'll look into it at some point, seeing as it's a fairly important thing.

EDIT: I've got a problem here which I'm wondering anyone can help with:

I'm trying to store depth information (I've done stuff to it) in the alpha channel of a pass, and then use it in the next pass. However, it appears that for that next pass the alpha channel has been set to 1.0 for all pixels, overwriting my stored info.

Does anyone know if this is a problem with my coding, or is it just how OBGEv2 does things? The RGB channels work fine for storing info, but I can't use them for the method I'm trying to use.

Oh, and if anyone's got a good explanation on how to downsample an image to 1/16 the original size, that'd be great. I think I got it, but then the next pass keeps everything downsampled for some reason. But I'll figure that one out later, the alpha problem is the big one.

EDIT 2: As an incentive, here's some screenies of a slightly hacky workaround to my issues:

http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-41-49-01.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-19-50-25.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-19-43-83.jpg

It's about half way there. Besides the above issues, I've also got to figure out how to get a decent circle of confusion from the variables I have.

EDIT 3: I just scoured this thread and the last for screenies to upload to Nexus. Unfortunately, about half of them were corrupt - if you've got any screenshots that you've taken recently, either post them here for me to upload to Nexus, or upload them to the Nexus page yourself. If uploading straight to Nexus, please try to follow the naming pattern of what shaders you've got active in the screenshot.

[Hairul Hafis]

I think that may be possible, but it would have to be using that method shadeMe suggested (and was working on?) that used GetCrosshairRef and some Oblivion scripting before passing the distance/position or something on to the shader to do its stuff. Obviously we know what the ref is, but it's a matter of whether we can get the distance to it from the camera, or it's centre position on the screen, or something.

I'll look into it at some point, seeing as it's a fairly important thing.

I haven't tried it out yet. This should change by the next week though.

EDIT: I've got a problem here which I'm wondering anyone can help with:

I'm trying to store depth information (I've done stuff to it) in the alpha channel of a pass, and then use it in the next pass. However, it appears that for that next pass the alpha channel has been set to 1.0 for all pixels, overwriting my stored info.

Does anyone know if this is a problem with my coding, or is it just how OBGEv2 does things? The RGB channels work fine for storing info, but I can't use them for the method I'm trying to use.

The ColorWriteEnable render state should be set to RedGreenBlueAlpha to allow writing to all channels. Which makes me wonder if OBGE parses state directives and applies them - I'll take a look at the code.

[Céline Rémy]

I haven't tried it out yet. This should change by the next week though.


The ColorWriteEnable render state should be set to RedGreenBlueAlpha to allow writing to all channels. Which makes me wonder if OBGE parses state directives and applies them - I'll take a look at the code.

Great, thanks. No rush on that GetCrosshairRef stuff though.

Here's a couple of other DoF screenshots:
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion2010-04-1723-26-58-10.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion2010-04-1723-26-26-47.jpg

[Alex Vincent]

No rush on that GetCrosshairRef stuff though.

I hope you realize what you just did ...

Here's a couple of other DoF screenshots:
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion2010-04-1723-26-58-10.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion2010-04-1723-26-26-47.jpg

:thumbsup: Impressive, to say the least. The CoC could use some work though, as you stated before.

[offtopic]PS : Would you folks know of any reasons as to why Oblivion (even the vanilla build with nothing more than the main master enabled) would crash on load ?[/offtopic]

@Ninja, on the matter of downsampling: Have you tried moving your downsample pass to the top of the compile technique ? A hierarchy like this:
p1 { downsample }
p2 { grab relevant data }
p3 and onwards { process thisframe and subsequent lastpass textures }

[jeremey wisor]

[offtopic]PS : Would you folks know of any reasons as to why Oblivion (even the vanilla build with nothing more than the main master enabled) would crash on load ?[/offtopic]

I had that problem when I reinstalled Oblivion after a drive wipe and OS reinstall. For me, Oblivion was reading a specially configured Oblivion.ini that had been backed up and restored onto the refreshed computer with references to DarnUI fonts and other resources that were no longer present. Reinstalling DarnUI or removing those references fixed the crash-on-start problem I was having.

Who knows if it is the same, or similar, situation for you though. But, this might give you a place to start.

[Sophie Morrell]

Best bet is to delete your ini and start fresh if you crashing with vanilla oblivion.

[Carys]

I had that problem when I reinstalled Oblivion after a drive wipe and OS reinstall. For me, Oblivion was reading a specially configured Oblivion.ini that had been backed up and restored onto the refreshed computer with references to DarnUI fonts and other resources that were no longer present. Reinstalling DarnUI or removing those references fixed the crash-on-start problem I was having.

Who knows if it is the same, or similar, situation for you though. But, this might give you a place to start.

Got it fixed by reinstalling DarnUI. Many thanks.

[Rachel Cafferty]

EDIT: I've got a problem here which I'm wondering anyone can help with:

I'm trying to store depth information (I've done stuff to it) in the alpha channel of a pass, and then use it in the next pass. However, it appears that for that next pass the alpha channel has been set to 1.0 for all pixels, overwriting my stored info.

Does anyone know if this is a problem with my coding, or is it just how OBGEv2 does things? The RGB channels work fine for storing info, but I can't use them for the method I'm trying to use.

Oh, and if anyone's got a good explanation on how to downsample an image to 1/16 the original size, that'd be great. I think I got it, but then the next pass keeps everything downsampled for some reason. But I'll figure that one out later, the alpha problem is the big one.

EDIT 2: As an incentive, here's some screenies of a slightly hacky workaround to my issues:

http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-41-49-01.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-19-50-25.jpg
http://i617.photobucket.com/albums/tt255/WrinklyNinja1/Oblivion%20DoF/Oblivion2010-04-1722-19-43-83.jpg

It's about half way there. Besides the above issues, I've also got to figure out how to get a decent circle of confusion from the variables I have.

EDIT 3: I just scoured this thread and the last for screenies to upload to Nexus. Unfortunately, about half of them were corrupt - if you've got any screenshots that you've taken recently, either post them here for me to upload to Nexus, or upload them to the Nexus page yourself. If uploading straight to Nexus, please try to follow the naming pattern of what shaders you've got active in the screenshot.

Normally OBGE uses a R8B8B8X8 (8 bit no alpha) texture for the intermediate frames. I think I tried a R8G8B8A8 (8 bit with alpha) texture but I couldn't get it to work. Changing the OBGE.ini setting to use a 16 bit or 32 bit frame buffer texture should enable you to use the alpha channel.

I'll have another go at getting the alpha channel working for the 8 bit frame buffer though.

I was thinking. Enabling vertex shaders for the effect files seems to be causing a lot of problems. Do you guys want it to be an opt in feature? i.e. an effect file will use the old fashioned way of rendering the shader unless you have a line saying bool bEnableVertexShaders = true; in your global section of the effect file making it use the new way of rendering the shader.

I also must get round to getting OBGE's shaders to render before Oblivion's shaders therefore minimising any nighteye / HDR issues.

[djimi]

Hmm, I've got an ATI GPU too, so that can't be it. :shrug:

I'm in the middle of a rewrite using some of Crysis' ideas anyway, so we'll see for the next version.

Good to hear. But just to confirm, I too have the same lack of Godrays problem.

I just recently downloaded the three main packages from TESNexus. And the v4 Godrays etc.

http://i618.photobucket.com/albums/tt261/ADParker314/ScreenShot35.jpg?t=1271600826
http://i618.photobucket.com/albums/tt261/ADParker314/ScreenShot37.jpg?t=1271600856

[Chavala]

@Ninja, on the matter of downsampling: Have you tried moving your downsample pass to the top of the compile technique ? A hierarchy like this:
p1 { downsample }
p2 { grab relevant data }
p3 and onwards { process thisframe and subsequent lastpass textures }

That works, but unfortunately I'm trying to create a downsampled pre-blurred image with the alpha channel containing relative depth information, which I then combine later with thisframe to create the DoF blur. Or that's how I'm reading what I have to do. I'm not really clear on the need to downsample and pre-blur the image. :shrug:

On a different note, has anyone come up with a good linearisation of depth? I'm currently using HawkleyFox's code:

float4x4 m44proj;static const float nearZ = m44proj._43 / m44proj._33;static const float farZ = (m44proj._33 * nearZ) / (m44proj._33 - 1.0f);	float depth = pow(abs(tex2D(DepthSampler,UVCoord).x),0.05);	depth = (2.0f * nearZ) / (nearZ + farZ - depth * (farZ - nearZ));

But as good as it is, it still leaves much of the depth detail relatively close to the player. I'm not sure on what the transformation matrices hold, but shouldn't there be something in them that can be used to get a linearised blur?

[El Goose]

Im not sure if this has been asked before, but can the SSAO effect be applied on npcs and objects only? Im currently using ssao_perf.fx (due to the dialogue lag issue with ssao_test.fx), but using it with the Optimized Distance mod made the landscapes appear a bit weird, like shaded patches of triangles on the ground. Also water surfaces and foggy days can look pretty glitchy with ssao_perf.fx.

But anyway my main question is... is it possible in any way for SSAO to be used on only the actors and objects, while forced off for landscapes?