Calm and civil discorse about update 31-07


#241

Remember just because your paranoid it doesn’t mean they are not out to get you…


#242

They literally wrote in the update that the testers will be receiving the new m2 headsets…and they haven’t given green light for m1 which is why they made an m2…


#243

OK. But if they are problems again we are back at January again. For Pimax it would be not so bad because now they can sell all HMDs to business partners and we backers are the stupids again. Sorry for my negative attitude but I fear this situation is coming…


#244

You ain’t gonna do a Fourier transformation based scaler at 90Hz at 2xWQHD resolution.
Just drop the thought.


#245

But how can you scale an image if you do not have all the pixels available prior to starting the scale. How can any scaling algorithm work without having all the data adjoining it. Simply duplicating the same pixel around it without changing the colour data would produce a horrendous up-scaled image wouldn’t it?

Hmm, maybe that is how hardware scalers work? On the most basic of level.

Will have to do more reading.


#246

No her answer is simple. Until the Testers can Test the m2. The answer is not available.


#247

Testers are apart if the Testing process. Without them Testing the M2 yet. Pimax cannot answer since this has not been completed.


#248

The scaling is very simple, assuming it’s the same as the 4K. Upscaling doesn’t need all the data, just 2 lines (the current one and the next one, so it has all the neighboring pixels it needs). Why? To reduce memory needs, simplify the chip (which makes it cheaper), and reduce latency (it only needs to wait for the next line, not the full frame).


#249

There is. SteamVR render scale does that automatically since a while ago if you don’t chose to override it.


#250

Cheap n cheerful aye :smiley:
So that would probably be a bilinear filter in the Pimax 8K upscaler?


#251

Thanks for the info.


#252

It’s likely. I don’t know for sure, because the blending algorithm wasn’t specifically mentioned.


#253

We dont know but im sure its a fast one and no Lanczos for latency reasons.


#254

Got ya,

While researching how these scaler chips upscale I came across a new breed of scaler in the new Samsung Q9S 8K set. Now that has what appears to be a revolutionary scaler. It uses Artificial Intelligence, well, a learning algorithm to constantly improve itself based on what you watch. I think it probably uses RAISR,

To side track a bit more, I read that poor poor performance games (that drop to 45fps) can be brought back to as-fast-as-possible fps by turning off SteamVR’s Reprojection
SteamVR Settings>Configuration>Performance>Allow Reprojection

I don’t know if Brainwarp works on Pimax 8K yet but I wonder if that setting has an effect.


#255

The SteamVR reprojection methods has always been a bit broken in my opinion. Especially the Asynchronous, as the dropped frames are very inconsistent and juddery. Like watching a 24fps movie on 60Hz display without proper telecine. That’s very easily spotted in free locomotion games where even a game running in 88-89fps looks extremely jerky.

Im not allowed to talk brainwarp yet :frowning: sorry


#256

This is where Oculus seem to hold the crown?. I take it ATW and ASW are still king when it comes to fixing judders.


#257

Definitely. ASW is what all headsets needs. WMR headsets have its own motion reprojection similar to ASW that is almost as good (with some more artifacts only).


#258

wait, but according your YT “vive pro rant” they didnt offer you free Vive Pro


#259

Yep, HTC totally ignored me lol. I was offered by someone else a while after that video :slight_smile: But I could not find a reason to accept it. And I dont need it.

If HTC doesn’t want me to have it, then I wont use it.


#260

I take it these are not opensource algorithms? Everybody has to recreate their own version without stepping on somebody else patent. If so, I hope Pimax have a Carmack clone lol