Omg! Why didn't I think of this sooner?! Put the upscaler in a breakout box!

Why does the upscaler have to he integrated in the headset? Why couldn’t a powered breakout box between the card and the hmd do the upscaling?

In this way the 8k can accept native 8k with two DP cables while the powered box does all the scaling. Sure, it would add to the cost, but the headset would be a native 8k headset with the breakout box scaling up whatever resolution is required. And because it’s powered you can basically put a linux system in there dedicated to upscaling.

And if GPU’S ever get the power to push 8k natively the breakout box would just act as a passthrough, or just connect the hmd directly to the card.

What do you guys think? Would you accept a further delay of the 8k and get a 5k while Pimax works on this solution?


Or even put two rtx cards in there using dlss to upscale to 8k!


The hardware is finalized, and shipping process begun.

Personally speaking, I’m happy with the current hardware’s reviews, and wouldn’t support any further delay.


How is making the upscaler external going to allow 2 DP cables? The upscale still happens before it’s displayed on the HMD. I’m confused about that. And, if the 8K right now barely gets by with a 1080TI… wouldn’t that seriously increase the GPU required to go full native on two DP cables? I would think if that was a really a viable solution then wouldn’t that be a better solution for the 8K-X?

1 Like

On one end there will be two synchronized DP cables. On the other end will be the input from the video card. The breakout box will be able to accept resolutions all the way up to 4kx2. Then the breakout box will upscale to the native resolution of the headset and sends the output to the two DP cables. The breakout box would also be able to handle doubling the frames if the input falls below 90 fps and other functions.

An external breakout box frees you from the weight and power constraints of the upscaler being in the headset. Heck, the wireless transmission module could even be in the breakout box this freeing the computer of that load.

So instead having an upscaler close to the panel, right where it should be, you will put it outside the HMD, add another DP bridge, another DP cable, and make the upscaler much more complex because it will need to operate on DP streams, instead of on “native” MIPI, without improving anything else.

EDIT: Besides, by removing the scaler and adding the second DP, you make 8K-X out of the 8K.


It’s basically be an external upscale for the 8kx if it’s placed on the person with a single cable or wireless going back to the pc. Actually I might want something like that if the wireless module doesn’t come to the 8kx.

Who says the upscaler needs to be close to the panels? It can occur in the breakout box. The panels then just become monitors.

I was thinking the upscaler would be close to the pc. It would be a mini-pc itself running Linux or something.

you would then need two cables to run from the headset to the breakout box in which case you may as well skip the box and do the upscaling on the gpu. the main advantage of having scaling in the headset (or close to it) is that you can reduce the bandwidth to something that can be transferred over a single cable or probably even wireless.


Well, the upscaler in the headset can’t do 90hz, can it? And doing the upscaling in the GPU requires a rather powerful GPU. In contrast, the brea6box would be it’s own self contained, powered system dedicated to upscaling. And because weight and power would no longer be a factor, a more powerful system could be developed for the breakout box to reach 90hz, regardless of input or GPU.

Logic. Any image transformation on DP stream would require two DP bridges around the scaler for decoding and re-encoding DP stream. This would add additional latency.

This is the reason why it is better to do it either at the source (gfx card) or at the panel. Gfx card would be better, but then we are talking 8K-X again.


How is the signal of the DP transmitted? And just how much latency would s breakout box add?

I get what you’re saying. If the upscaler is in the headset then it can send the signal directly to the panels, while the breakout box would have to re-trandmit a DP signal after upscaling.

I’m not married to two DP’s from the headset to the breakout box. It can be any type if cable or standard that can handle 4kx2. And the breakout box would be fast enough to make the latency negligible.

Right now, as far as I know, there is NO standard that can handle 4Kx2 at 80 or 90 Hz.

1 Like

Well then! It’s time to create one! Right @PimaxVR? :-Đ

Perhaps pimax is not very good at communication. But they are very very good at hardware tinkering. That’s why they managed to make the first good wide FOV headset which is better than any current competitor on the market.

If your idea would be this easy they would have done it already… :wink:


what about StarVR? at IFA in berlin it sounded like they have started production and will come up with delivery dates an prices in q4/2018 kind of the same as pimax, consumer market start after backers have the hmd’s (goal is christams) and no official price yet
and be honest if you dont have to pay for it, what headset would you choose 5k+, 8k or StarVR One?
after seeing mrtv’s comments after testing and having used one of the “pre” units in arcades myself i’m sure that the “One” is the one

but back on topic
there seems to be already a solution, the anx7539 will have a scaler (that for sure can be switched off and i’d guess will be able to handle more then one setting) and the 8kx is already planned with 2 of them and 2xDP
so there is no need do develop such complicated and expansive solution (that might fail in the end because of latency problems)
also such box will increase the wireless problem, if put that apart from the hmd’s wearer you have to deal with 2x4k@90hz, if its inside the hmd (like the anx7539) then you signal to transmit is 2x1440p@90 for now, its not clear atm when there will be a solution for 2x4k@90hz, for a price that a consumer might want to pay, if you do it with whats available (see tpcast or intel 60GHz + display link chip in htc’s solution) you might end up with more then the hmd will cost

also important (for pimax at least) is that is is something thats already available for a good price, any hardware development beside the hmd is done by partners, they dont develop (in perspective of complete R&D) solutions for wireless or eye tracking they will choose from whats available and if possible will outsource the production, just giving out the specs of the module (how big and how much current it can get from the hmd)

I want to wrap myself in a cocoon made out of upscalers. I will emerge as beautiful pixels.

1 Like