February 22, 2019

Realizing HoloLens Field of View (FOV)



My posts about FOV are not done. I still have at least two more, this is one of the two. Why? Because people care, and because I have something to say. For reference, please check out my other posts:

Lets get right to it. As mentioned before, I was convinced the HoloLens’ FOV has become a problem, and Microsoft should do anything possible to fix that before the product goes to the market. In the meantime, I suspect that the limitations that constrain the FOV may not have an easy fix. I mean there are clearly ways to extend the FOV, but trade offs may make it not worth while. The likely scenario is: It will get fixed to some extent, but not restoring to its January level.

Realizing it is an issue, I do feel it is over-blown in the media. It’s disappointing only because we know what it could do three month before. If out of nowhere, someone showed you a device like this, everyone would still be blown away. It’s still undeniably magical. Like what they said, the thing exists itself is no short of a miracle.

So, what if there is no major fix? What if the product is release as is now? What kind of experience can we get?

First, we need to understand, FOV is not the only factor. Distance play a big role in regard how much you can see. If you imagine you are looking at the holograms anywhere beyond 10 feet, that wouldn’t be a problem. I believe most game setups are not close to the player, unless you have to kick something by feet, or grab something by hands, that figure better be small. If you are in a virtual concert, you expect the holo-people all around you, that wouldn’t be possible. The Mars exploration scene wouldn’t be possible, because when you step on the hologram, part of it would be gone.  If you pin holograms all over on the wall, you won’t see them all at once until you step 15 feet away. Minecraft would be hard too.

But many other things are still fine. The Skype plumbing would work as intended. AutoCAD on a desktop while a hologram 3D model beside you will still good. If you use HoloLens as a HD TV, it will still work perfectly. You don’t watch HDTV 5 feet from your eyes, right? Trimble construction example will still be realized well. The medical professor can still show human anatomy in 3D hologram in class, just don’t sit in the front row. Most of the holo-ideas will work as expected. Virtually anything that is fine with a distance would still not be affected that much.

But we are not satisfied with that, we want to see more. In my next post, I will show you what can be done to achieve that.




HoloLens – The Matter of FOV




Reports have come out from BUILD conference focusing on concerns over HoloLens’ FOV (Field of View, or Field of Vision). I first got the message from social media, and I was in shock at first. For a moment, I thought it was the same crowd who complained about FOV when HoloLens was first announced. I understand those who complained usually based their opinion on comparison to VR head sets. There weren’t a concern because they are different technology, and targeting different use cases. Then I read a number of post online, from the same people who were shown the demo last time. I have even checked with friends. I am convinced that there is an issue to deal with, big or small.

Regarding FOV, the feelings are universal, and almost identical: it’s just too small this time around. I looked back at their previous posts, the opinion toward HoloLens FOV are mostly favorable. Most of them felt it’s not great but fine. Of course, bigger would be better.  It has never become a barrier for a great user experience. Honestly, after I have written the post “Analysis of HoloLens FOV“, I have never considered this a problem.

Many people I have read online are surprised. No one had expected this. It is certainly not because the technology is not capable, we have seen how it worked. It should get better over time not the other way around. Now many have realized the setup in two cases are different: First time they were strapped with a laptop. This time it was on a less powerful Atom CherryTrail PC, according to report. Based on that, it can be one of 3 things: 1. The light engine (generating and projecting lights) component has changed to fit in the small chassis. 2. The light engine is the same, but less powerful CPU to power it up, which limits the among of lights it can process. 3. The GPU that takes care of creating the initial image, though that image is never rendered. I am no expert, but if have to take a guess, I tend to think #3 is less likely, because the GPU does the same thing as it does on any other PC. Generally speaking, today’s GPU has no problem dealing with 1080p videos. I think #2 is most likely.

I know there must be difficult decisions behind the scene, but what makes it hard to believe is they (Microsoft) appear to take chance to see how people react. There is strong evidence of that: 1. Obviously, they don’t just find this out today. They must have known it long before today. 2. They have marketing people ask your expectation when you come in, then ask again afterwards. They are testing water. That shouldn’t have happened. If it is really like what people have wrote today, that’s too obvious, there is no need to test water like that. It’s a dangerous act they shouldn’t ever tried.  At least, it damps enthusiasm among everyone who are watching the development. The impact can be more than that.

In any case, I would expect an explanation coming out in any form. People need to be convinced there is a fix. Related question is: how would that affect the  release date. One thing is almost certain: HoloLens can not be released in this state. The reason is obvious.

The positive side is we knew it worked.