openmw.org

AnyOldName3 wrote:snip

Finally someone who knows what he is talking about... I'm really tired of reading that the human eye cannot see more than 30 or sometimes even 20 frames, just because movies are filmed at 24. There may be people who believe they cannot see a difference between 30 and 60 fps or maybe they really can't but these people tend to claim nobody can. There are more than enough websites where everybody can make the comparison for himself so yes, I wish they would just stop. But this is the internet, after all.

Anyway, of course, more than 60 is usually not necessary for games, which is why the VSync setting exists (well, that and to remove screen tearing because of the frame rate not matching the screen's refresh rate), but of course, for benchmark purposes, it is important to know how many frames can be calculated, even though they cannot be displayed because the screen only works at 60 Hz.

That being said, I am amazed by the performance of the newest version! And it looks just like vanilla Morrowind, but with much better water! Seriously, the water is amazing and also so much more performance-friendly than before. Can't wait for the next, even better versions!

You speak as there is lot of heated discussion on the Internet... well, it might be possible.


About the point, I think the confusion comes from making the wrong questions and having implicit assumptions.

As well explained in 100fps website.

This questions are all different:
- How many frames per second can the human eye see?
- How many frames per second do I have to have to make motions look fluid?
- How many frames per second makes the movie stop flickering?
- What is the shortest frame a human eye would notice?

Besides, in video games there is also the factor of response time as Scrawl mentioned.

ezze wrote:You speak as there is lot of heated discussion on the Internet... well, it might be possible.

Yeah, there is. But then again, about what is there no heated discussion on the internet?

Anyway, yes, you are right about the wrong questions and assumptions. Frame does not equal frame and yes, seeing does not work in "frames", anyway. But the other point is: this thread is not about how many frames can be seen or displayed, but how many can be calculated, in order to see how the engine performs on different systems. In the end one would most likely play with VSync on, but the maximum possible framerate is still useful information.

I'm currently running 0.38 and these are my specs:

Intel G3258 3.2GHz Dual Core
8GB DDR3 Ram
AMD R9 270X 4GB
Windows 10 Pro
SSD

Game Settings
Resolution 1920x1080
Texture Filtering=Trilinear
Anisotropy=16x
View Distance=Far
Water Shader=On
Refraction=On
Texture Quality=High

I was able to get an average of
70 FPS walking around Balmora (exteriors)
390 FPS walking around Balmora (interiors)

Very promising considering my PC is definitely not very powerful, yet got significantly better performance with OpenMW. I'm looking forward to the shaders getting added into OSG. Hopefully this information is beneficial for you all and let me know if I missed any information you'd need!

>AMD CPU+GPU work better in vanilla than OpenMW
My friend is pretty assblasted about that.

TechNoirMK wrote:>AMD CPU+GPU work better in vanilla than OpenMW
My friend is pretty assblasted about that.

Yeah, well, let's hope this is going to improve in the future.

Another AMD+AMD here. I get really bad performance with the reflective water (even if the shader file itself is nearly emptied) but with texture water fog is applied wrong. Getting a pixel-shaded water setting that doesn't set up reflections would be great, at least until the vertex-shaded water is fixed for fog.

wareya wrote:Another AMD+AMD here. I get really bad performance with the reflective water (even if the shader file itself is nearly emptied) but with texture water fog is applied wrong. Getting a pixel-shaded water setting that doesn't set up reflections would be great, at least until the vertex-shaded water is fixed for fog.

reflective water causes the renderer to redraw the portions of the scene that reflect twice no matter what the shader code does with the final pixel. so performance, unless more sophisticated optimizations are implemented, will generally always be quite a bit slower with scenes with reflective surfaces.

I know that reflection is expensive outside of the shader.

i'm happy for you. for those who didn't already know, they do now.