TalosXT

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

Hello everyone,
 
I recently watched the RED camera video and it was nice since i stopped wondering what camera LTT uses. However a question answered and two takes its place.
 
1. From my understanding, LTT always likes to take it to the extremes so why not go for an ARRI ALEXA XT, ARRI ALEXA MINI, or even ARRI AMIRA?
Correct me if I'm wrong but XT as I have heard is around $70,000 USD or ARRI AMIRA is which is around the same price of a RED. Plus you don't have to deal with RED charging all those insane prices for cables and what not. I'm only saying this because I think ARRI would have been a better choice for LTT because they are always featuring people (Linus himself) which on ARRI would have the skin looking softer which is one of ARRI's winning points. I always thought RED was used for more action oriented or open environment shooting but maybe that's just me being affected by RED's marketing direction, design and naming choice. 
 
2. Why did LTT continue using RED despite all its problems?
I know they made a follow up video with one of their camera operator explaining some things, but to me it just didn't make sense. The guest shooter in the video said its better for post because you can reframe and downsample (which you still could do with an ARRI) and don't get me wrong, but that just doesn't make sense. As far as I can see, LTT is not doing action sequences with characters shooting each other up while the camera man is playing fidget spinner with the camera.
 
Also Linus, if you ever read this thread, it would be amazing to see YOUR opinion about RED after all these videos. Did you change your opinion or do you still think it wasn't the best choice? 

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

While you guys been testing games and making videos and forcibly proving/denying the problem to each other, Me and @SpeedUp10  were actually investigating the issue. Please stop testing games and making videos and start researching. For once we actually know what the problem is however the solution or what causing it is unknown.
 
First of all, Please Please and Please stop making general names for this problem. "Texture Flickering" or "Jagged Edges" will only attract people whose problem can be fixed by reinstalling driver or windows or people who wanna prove to us nothing is wrong. 
 
So here is the investigation:
We did a bit of research here and there and found out that this problem is something that is happening once there is a camera motion. meaning that a fixed frame is rendering perfectly. therefore we concluded that this is because of some sort of calculation being run in GPU and is being messed with and it takes a few seconds to re-adjust. After a bit of more researching and cross-solving, we drew some connection to LoD Bias. Now LoD bias is a series of calculations that aid the GPU in rendering each object/texture specific to their location in relation to player. The values being yielded by theses calculation also aids the GPU in draw distance and shadow quality (Depending on Shader being used by the engine). Now the APIs (DirectX, OpenGL, VulkanRT) are responsible for this. So I went after articles written by shader programmers wo use these APIs to specifically find the problem since blaming LoD Bias is again too general (LoD bias has alot of more calculations). So we came to this conclusion that this is actually a really common problem in shader programming and thats why simply searching "Texture Flickering" on google wont get you to this. We tried alot of things, even went for RF Noise or radio interference or testing game enigine based LoD bias. Also a clue that helped was that this is all about shadows. this may also come as bit of shock but shadows are what we should blame the whole thing on, and you maybe say that im crazy why should I blame shadows on jagged edges or popping but bare with me im going to prove why.
 
So guys, to shorten this up, the problem is not one but 3:
 
1. Shadow Acne (Main one)
Another common artifact found in shadow mapping is shadow acne, or erroneous self-shadowing.  This generally occurs when the texel depth in light space and the texel depth in view space are so close that floating point errors incorrectly cause the depth test to fail. There are a few ways to address this issue.  It’s so prevalent, that most graphics APIs provide a means to instantiate a rasterizer state that includes both a depth bias and a slope-scaled depth bias.  Essentially, during shadow map creation, these values are used in combination to offset the current value by a certain amount and push it out of the range where floating point inaccuracies would cause inaccurate comparisons.  One must be careful when setting these bias values.  Too high of a value can cause the next issue to be discussed, peter panning, while too low of a value will still let acne artifacts creep back.
 
2. Peter Panning
It’s frustrating when introducing a fix for one thing breaks something else.  That’s exactly what we can potentially end up with when we use depth biases for shadow maps.  Peter Panning is caused by offsetting the depth values in light space too much.  The result is that the shadow becomes detached from the object casting it. Sometimes the depth bias is so large that it pushes the shadow away from the caster, causing them to appear as though they could be floating. 
 
3. Perspective Aliasing
These types of artifacts are perhaps the simplest to alleviate.  Stair-like artifacts outlining the projected shadows are generally caused by the resolution of the shadow map being too low. Unfortunately, increasing the resolution will only get us so far.  Even at high resolutions, if the viewer is close enough to the receiving surface, tiny stair-like artifacts will still be noticeable along the edges of projected shadows.  The solution to this is to use a technique called percentage closer filtering (PCF).  Instead of sampling at one location, this algorithm samples several points around the initial location, weighs the results that are shadowed versus non-shadowed, and creates soft edges for the result. 
 
TL;DR Stop looking for proofs and videos and game problems and start looking for a fix for Shadow Acne. Trust me when I say this. You fix Shadow Acne your problems will be solved forever. How to fix it? that's the million dollar question.

So guys, here's an update after a week of researching:
 
First of all thanks for those who submitted the D3D AF Tester results. I also grabbed some off of internet to see it running on different cards.
 
Second of all, It appears all of us are facing different issues here. Similar symptoms but different kind of problem. But the funny part is that all the problems are affecting LoD Bias and ultimately messing with draw distance.
 
I fixed Witcher 3 a few days ago by forcing nvidia inspector to turn off LoD bias Driver Control, meaning I gave all the LoD calculation responsibility to the Witcher 3's engine. IT FIXED IT (except the CSMs). This means that the game engine is doing it's job correctly however the GPU was interfering with it, and instead of aiding the game in the LoD calculations, it messed it up resulting in shimmering, flickering, Alaising and ultimately CSMs. Now in order to prove the theory, I installed Stonghold Crusader 2 (2014) which uses a cheap engine that rests on Hardware based post processing. The problems were back again.
 
Now why do I think it's different for everybody? This is where D3D AF tester comes in play. let's look at @Zemzemma's. As you can see, the hardware based Anisotropic Filter is smoothing the beginning of the tunnel but as it progresses through it appears as it is warped horizontally. even though the LoD bias is set to -8, the filter is clamping it to 0 (transition from smooth to jagged) but beyond that it can not process. That should not be happening. for me it's a different case. the filter is allowing the the negative bias to pass through even though I've forced it to to clamp to 0. *** NOW this can be because of many different reasons. These LoD problems have DIFFERENT sources and are yielding DIFFERENT RESULTS. But ONE THING COMMON between the results is that it's messing with the draw distance calculations.
 
I'm willing to bet my money that this is the root to all of our problems. I will try to continue researching even though I may not have enough time since I'm a student but ONE thing i'm damned sure of is  that there is NO MAGICAL SOLUTION for everyone. Some people can fix it with driver's install, Some with UPS Stabilizer, Some with building a new PC, or some with neither. This is why we can't get to a solid solution for years.
 
@iiNNeX The problem that you're facing is similar to CSM. It's hard to tell since it's a top down camera but that's because the LoD bias cannot tell how accurately to draw the pixels so the computer is making a safe assumption that the shadow is very far and renders it in low quality (AKA realtime low rez Lightmap bakes). It make's it better when the game is dynamically lit (games with a day night system). Have you tried forcing the driver to stop calculating the LoD bias and let the engine do it? (If the engine has that option built in).

So guys, here's an update after a week of researching:
 
First of all thanks for those who submitted the D3D AF Tester results. I also grabbed some off of internet to see it running on different cards.
 
Second of all, It appears all of us are facing different issues here. Similar symptoms but different kind of problem. But the funny part is that all the problems are affecting LoD Bias and ultimately messing with draw distance.
 
I fixed Witcher 3 a few days ago by forcing nvidia inspector to turn off LoD bias Driver Control, meaning I gave all the LoD calculation responsibility to the Witcher 3's engine. IT FIXED IT (except the CSMs). This means that the game engine is doing it's job correctly however the GPU was interfering with it, and instead of aiding the game in the LoD calculations, it messed it up resulting in shimmering, flickering, Alaising and ultimately CSMs. Now in order to prove the theory, I installed Stonghold Crusader 2 (2014) which uses a cheap engine that rests on Hardware based post processing. The problems were back again.
 
Now why do I think it's different for everybody? This is where D3D AF tester comes in play. let's look at @Zemzemma's. As you can see, the hardware based Anisotropic Filter is smoothing the beginning of the tunnel but as it progresses through it appears as it is warped horizontally. even though the LoD bias is set to -8, the filter is clamping it to 0 (transition from smooth to jagged) but beyond that it can not process. That should not be happening. for me it's a different case. the filter is allowing the the negative bias to pass through even though I've forced it to to clamp to 0. *** NOW this can be because of many different reasons. These LoD problems have DIFFERENT sources and are yielding DIFFERENT RESULTS. But ONE THING COMMON between the results is that it's messing with the draw distance calculations.
 
I'm willing to bet my money that this is the root to all of our problems. I will try to continue researching even though I may not have enough time since I'm a student but ONE thing i'm damned sure of is  that there is NO MAGICAL SOLUTION for everyone. Some people can fix it with driver's install, Some with UPS Stabilizer, Some with building a new PC, or some with neither. This is why we can't get to a solid solution for years.
 
@iiNNeX The problem that you're facing is similar to CSM. It's hard to tell since it's a top down camera but that's because the LoD bias cannot tell how accurately to draw the pixels so the computer is making a safe assumption that the shadow is very far and renders it in low quality (AKA realtime low rez Lightmap bakes). It make's it better when the game is dynamically lit (games with a day night system). Have you tried forcing the driver to stop calculating the LoD bias and let the engine do it? (If the engine has that option built in).

Hello everyone,
I guess im gonna join your club. here's my story
 
Bought this rig November 2015
Asus Z170 Deluxe MOBO
Intel Core i7 Skylake 6700k
32GB Corsair 2666 Ram
2x Asus 980Ti Strix SLI
1000w EVGA 80plusTitanium PSU
 
Everything was working fine until I purchased GTA V on 27th of June 2016 only to get this magical bug, virus or whatever the heck it is (yeah @iiNNeX I read your 98 page thread on nvidia's forum). Now I remember getting Witcher 3 Blood and wine on 3rd of June and playing it until 12th of June and I remember it was gorgeous and not a single flicker or aliasing or white pixel was to be seen. So something happened between 12th of June of 27th of June that triggered this problem for me. I've had this PC for nearly 7 months and It's the first time it happened. The only thing that happened between this period was getting a PS4 and connecting it to the same power extension cord that my PC is connected to. But I doubt it's anything power related since my PS4 is working perfectly (Just played Uncharted 4). Also tested with different monitor and tv.
 
Things I tried:
Reinstalling Drivers
Reinstalling Windows (This was the time where I had to cross out any software or driver related problem)
Switching the GPUs place (top went to bottom and bottom to top)
Disabling SLI
Plugging to different power outlets
 
Now, we all have nearly the same symptoms, I sat down, gathered all the evidence there is and started analyzing. Here's the results:
In most games, (Including Witcher 3) the textures doing the flickering can be categorized in 3 different groups:
1. Decals, Vegetation, Transparent objects (Glass, Fence)
2. Geralt's armor, some baskets, some types of cloth
3. Metallic props, water reflections (White Pixels)
 
Now if any of my 5 years of studying game art assist me, the first group has one thing in common and that is alpha mask, which are assigned to Alpha channels of albedo or diffuse textures specially in PBR practices.
Second group I'm guessing is Micro-Details usually stored as micro detail normal maps or heightmaps which only renders fully when player is close to the object.
Third is because of reflection. The white pixels are actually light reflections, which was very common with games using specular maps (which is really stupid compared to metallic/roughness workflow). Now this is where it gets weird; I'm pretty sure games like witcher use PBR (metallic/rough) and not old school specular stuff so its definitely something to do with gpu trying to render that.
 
Now I'm not a hardware guy and I'm just trying to connect the pieces here so correct me if I'm wrong. When I was testing I saw the same thing you guys been saying. The GPU is only trying to perfectly render only 2-3 meters from the player and as soon as I get far more than 3-4 meters from an object (Lets say thatch roof on a house or a basket), The ailaising looks the same as if it was 500 meters away (One thing that can support my claim is the CSM - Cascade shadow maps - around geralt where a light shafts disappears near him as he get 3 meters away). I know how CSMs work and this should not happen that close to Geralt. So the PC (Our mystery piece of hardware I dont know which that is responsible for this) cant tell the difference from an object too close (but still 3 meters away from player) the same as the object that is too far away, causing jagged edge as the pixels trying to adjust their location causing the famous flickering or pixel shimmering. Let me reword it better. It's like a situation where the game can only render close as in 3 meters perfectly and after that it starts rendering everything else crappy.
 
@iiNNeX I read all the things you tried on the other forum, your mystery virus theory may not be all that wrong. however "Virus" may be the wrong term for it. I a newbie when it comes to BIOS and hardware but I have a few theories:
1. Something triggered a change in vBIOS or MOBO's BIOS or even both where it changes a few parameters to the driver every time the OS changes or new driver is installed. Like I dunno maybe overheating that forced the GPU to go into safe mode disabling some capabilities of it (Which I'm not sure if BIOS is writable during use or even at all. Maybe it was a blue screen or a hard shut down).
2. Power, (I don't think the system would be that weak). Maybe some other Environment changes like ambient temp?
3. Ok dont laugh at this but I believe a spider went into my gpu and got fried, taking something with it while dying. (Srsly this is what i wanna believe rather than a hysteria or issue with game engine or etc.)
 
-Elijah D.