Letting out some steam: My Full Feedback on Current State of High Fidelity

#83

So what you are saying is that we should 100% abandon all content created within the past 5 years because Oculus is coming out with an upgrade flight recorder I can stick on my face from anywhere? I think you are forgetting that THE COMMUNITY made most of the stuff in High Fidelity, and some of those people aren’t even here anymore! That’s a very selfish request to just say “Hey, we now want to abandon all usually performed actions and go to mobile. That stuff you sunk hours, days, months of your life into? Oh well! Commissioned someone? Too bad!”

Let me add more salt to that wound: I recently contacted a fantastic artist who made a fresh new avatar and it looks fantastic. The way they pulled off turning a high poly model normally meant for 3D printing into a game engine optimized model by recreating the entire topology by hand was both amazing to watch and really showed how far they were willing to go to meet the criteria requested. That entire project cost me about $510, was done in less than 5 days, and they have many more commissions coming in. This was finished roughly 1 month ago and is about 20k poly, which is pretty nice for the details that were added.

So basically, it isn’t mobile friendly, I didn’t make it, and to go back to the same artist and request an even lower poly version would most likely incur additional charges for having them go even further to lower an already well done topology. Just to support Oculus Quest. Same goes to my $800 primary avatar, and both would then look terrible when the majority come in on desktop PCs.

That’s wrong. That’s wrong on so many levels.

EDIT: Also to add, if we start making domains only for mobile devices, how do we enforce that? How do we set it up to only let Oculus Quest users to come in as to not be bombarded by a person wearing a Mathew or Pracella avatar (who are infamous for being stupid high poly)? On top of that, would it become expected for domain operators to run 2 versions of their own domain?

Versus if we had even BASIC LOD like specifying HIGH and LOW. HIGH would be our default and normal avatars while LOW would be our mobile friendly. That way, you could keep both and even start the process of adding LOD without adding additional work. And the kicker is you’d just need to tell the mobile device to read a different line on the FST and fall back if it can’t find it (to retroactively support already in place avatars).

2 Likes

#84

@FlameSoulis I feel to compelled to defend @IlanTochner because I met him IRL.

So:

I think this is an uncharitable interpretation of what Ilan is saying.

Ilan is as heavily invested as anyone here in the success of HiFi, through his company Kitely.
Perhaps he didn’t communicate clearly enough his bias and his business interests, which I have discussed with him personally so I feel more sympathetic towards.

Many people here have a bone in the fight, and every move made by HiFi has the potential to grow someone’s bone while breaking someone else’s bone. :stuck_out_tongue:
Just because person A presents a case for a move which will be better for him doesn’t mean he doesn’t care about person B and his success.

I was going to add my own 2 HFC here but I think I better put that in a separate reply.

2 Likes

#85

Not sure the future is to have low resolution content for Hardware that we want as small as possible and as wireless as possible. Not sure the future is to have the graphical power in those hardware.
I think that Cloud Gaming is going to be the solution. Oculus Quest might help to the popularity of VR, but is that really where we want to be? Maybe we should start to think about what will come next.

1 Like

#86

Oh crap this is much longer than I thought this would be. SORRY!
(I started writing about LOD and the challenges of the transition to mobile VR but it got away from me)

First I’ll explain my POV:

  • I’ve been a member of the “PC master race” tribe since about 1987.
  • I’ve been in VR since 2005.
  • I’m on record talking about the metaverse since at least 2012.

My views on the present and future of VR:
(I apologise, these are not arranged as logically as I’d like)

  • We’re still in a very early stage.
    We are still well within generation 1 of VR. CV1, Rift S, Vive Pro and HP Reverb, they’re all first gen. The same goes for all the software involved!
  • I think The Quest is also gen 1, but it’s also a mutant, a transition which creates difficulties for which many have no adequately prepared.
  • The Quest is amazing and it’s importance cannot be overstated!
  • I believe in the future most VR content will be “webified” or social VR based.
    Today 99% of the content is baked into applications, and users spend 99% of their time inside baked content.
  • I strongly believe in the future most content will be based on VR browsers/clients.
    The same goes for users/user time. Money too: right now basically 100% is spent on baked content, this will not be the case for long.

Now some more practical and probably painful things to say:

If you have invested in content for PC VR and resent being forced to transition to mobile VR:

  1. I sympathise, honestly.
    You have the right to feel exasperated.
    I also happen to think HiFi devs have NOT been preparing for this transition well enough.
    Frustration is definitely an appropriate emotion here.
    You have every right to tell the devs they need to make the case for you to transition and to help you do it.
  2. (But) You’re not really “forced” to make any changes.
    The number of PCVR users will continue to rise and you have no obligation to support all systems.
    Maybe some of your effort has been wasted, maybe it hasn’t.
  3. In the long run content creators have to make a choice: is your content accessible on all platforms or not.
    Not very long ago websites had to adapt to mobile access. I mean the OLD mobile access. 176 pixels wide phone screen access with d-pad navigation.
    Things have improved since, but you still have to build two websites, more or less.

Cross contamination!!!*
I don’t have a better word for it so this I’m naming this set of issues “cross contamination”.
I will be using a slightly less user friendly language in this section.

Problems:

  • Low end users want to access high end domain.
  • Low end users want to use high end script/items.
  • High end avatar user meets low end user.
  • Users of high end domain leave to hang out with their low end friends. “slumming”
  • Cap for number of users in event/space because of low end users.

From a content creators POV these will translate to:

  • I’m losing visitors (and revenue).
  • I’m losing market customers (and revenue).
  • I’m getting negative feedback from high end users, low end users, or both.

The challenges for a content creator are:

  • What do I do with my existing content?
  • How and for whom do I design my future content?
  • Where do I invest my resources?

I don’t have any good answers here, but I do know some things are not the answer:
Ignoring or delaying is the transition to mobile VR is not a good answer.
Hoping some technomagic or a huge sum of money will make this problem go way is not an answer.

Now I have a few more things to say and I’m afraid I’ll have to make yet another separate reply, but a very short one.

2 Likes

#87

Addendum:
@FlameSoulis I really value your analysis of the problems involved.

Even if cloud or edge computing ever happens for VR, it’s too far off to consider it a solution.
It’s not going to save anyone involved today from making difficult decisions today about what to do tomorrow.

Great news!!!
OrbusVR will be available on the Quest!!!
https://uploadvr.com/orbusvr-reborn-coming-to-quest-pc-vr-release-date-revealed/

1 Like

#88

That’s a straw man argument if I ever saw one. Where did I say anything even remotely close to that?

What I said is that it’s important to enable the creation of domains that can work on Oculus Quest class devices and that if that requires accepting limitations for what can be used in those domains then that’s a price worth paying. Nothing about that limits what you can do on other domains that aren’t designed to work on this class of devices.

Like it or not, even with LOD support (by the platform and the content) there are still going to be a lot of extra optimizations that will be required for getting a domain to work well on low-power devices such as the Oculus Quest. It’s going to be a while until the platform includes those capabilities and it would be a mistake to delay support for Oculus Quest until you can have domains that can automatically make the most of both desktop-driven VR and devices such as the Oculus Quest. Until that happens you are likely to have domains that are targeted and optimized for one class of devices or the other. (And as an aside, that wouldn’t be an issue for how our organization-focused solution works.)

1 Like

#89

Well I guess I should also add a bit more to it: just to put some things into perspective, I’ve always done what I could to improve user experiences even down to optimization. Unlike most avatars running around, mine is ‘baked,’ meaning that my avatar was shoved into the oven and given KTX auto-decimated textures that are set up to talk natively with the GPU of both PCs and mobile. This also means that my avatar uses less bandwidth and is less taxing on most people’s systems. Combining this with asking my own server to provide any compression assistance when available to further reduce bandwidth costs on the network due to optimizations on the GPU side, I have been always wanting to make it so my avatar plays as nicely with the platform as possible.

The problem is exactly what you pointed out: High Fidelity really didn’t prepare for this at all and rather than fixing LOD, just dropped it. The result is that creators cannot appease both parties when the options were available in the past. Heck, I’d love even an option for objects to say “Hey, this is REALLY optional. Only show if the system can handle it.” but even then, it’s just not available.

However, I think your analysis is faulty: the problem is what is a low end domain and what is a low end user’s system? For example, my own domain is under 25MB in size. Small in bandwidth, and low on polycount. But what if someone builds a large city full of optimized models meant for low end systems? Now the domain isn’t a low end domain. So domains are supposed to now be small? Operators have to run multiple worlds because of lower end systems? That defies virtually all that made High Fidelity unique (domains can be f***ing huge).

The issue is that the systems were designed one way and now suddenly want to change directions, and that change just invalidates so much that was done so far without providing tools to improve compatibility. Heck, I had to include a link to my own compiled oven tools just in hope that creators do use it to improve optimization on top of helping to bug stomp issues that would impede on both server operators and users just for using them (https://github.com/highfidelity/hifi/issues/14086).

Even then, the platform doesn’t even run well on the intended systems! Ever since v0.79 and onwards, performance has just plummeted. Putting the same engine on a mobile system isn’t going to help matters, let alone with VR redrawing everything twice. If the Oculus Quest can party with everyone else like it was rocking a GTX970 or something, then it may stand a better chance, but as someone rocking a RTX2080, performance in High Fidelity just isn’t good, and those exact issues are going to carry over to any other system, low end, mid end, my end, etc.

I said it once, and I will say it again: High Fidelity has a reputation problem. The reason people are excited for things like VRChat and Beat Saber coming to the quest is because they were successful on their intended platforms, and have confidence the developers will make any sacrifices needed for the mobile/quest version not impare the original experience (Rez Infinite, by the way, did a fantastic job of this with the PC v Google DayDream version, as someone who has both (yes, this includes Area X)). High Fidelity has yet to prove it can provide a good, smooth experience on just one system. How am I (or anyone) to be confident that the migration to another system will work just as well?

1 Like

#90

You’re right, I completely ignored the question.

It’s no secret HiFi lacks the tools top manage this.
I managed to block this out in my mind, but when I showed up a few months ago I was very surprised, even shocked HiFi has no mechanism to prioritise what the client downloads.

I can’t say about this anything you don’t know already.

Regarding performance:
Again, I’m totally with you. I have a 1070 and i7 and if I didn’t have a serious interest in what’s going on here I’d have uninstalled more or less immediately.

0 Likes

#91

Can someone link me to where the HiFi only supporting mobile experiences rumour came from? Thanks

1 Like

#92

You’re forcing my hand: I’ll have to submit an article to https://www.vrdizzy.com/ along these lines.

1 Like

#93

Ironically, the “not-our-business-avatar” seems to be the biggest concern for performance.

2 Likes

#94

As a “prospective dev” that have been looking heavily into the hifi documentation to learn how this all works.

I have come to the conclusion that security is seen as a second class citizen. Allowing arbitrary users to run code on other arbitrary users “unless, the domain owner happens to disable it” is not and will never be a valid security model. If users need to interact with each other for whatever experience is built, that should happen via clearly defined standard interfaces, or via the domain owner’s server.

Anything custom should be explicitly whitelisted (by trusting certain scripts, code signing, certain licenses, users), and if something can be executed on the script owner and only render the result for other people ingame. that’s how that should be done.

When it comes to the hifi repo, I am also sad to see a CLA (which maybe I just dont understand), that makes it a big hurdle for developers to get involved. I certainly have vowed to never sign a CLA, after seeing the many projects be ripped from the community and re-licensed.

If you look at other successful open source projects you will often find an extremely active dev community that guides new users through whatever thing they’re doing. Which is sorely lacking here.
There’s a spirit that is missing from hifi, it doesn’t feel open source. it feels like one of those fake commercial open source projects, that are using the term more for marketing than actually wanting to be open source.

I have no doubt in Phillips intentions, I even agree with him that multiplayer VR spaces is one of the most important thing to come in newer human history.

That makes it even more important that hifi doesn’t become an implementation, it should strive to become a standard.

When reading about domains for example, it’s not made clear to me why place names are a high fidelity inc thing. Why isn’t this just normal domains, why can’t I use the global standard for domain name resolution. Come to think of it how does the hifi currency even work, is it just minted by hifi? that’s not very decentralized…

how does place name resolution work, probably just a high fidelity inc thing yet again.

If one really wants to have custom donains, that should be done like namecoin or the ethereum name system.

The overarching feeling is that the network isn’t uncoupled from hifi at all, it’s barely decentralized.

If we really are building the next frontier of humanity, this stuff needs to be able to exist even if a single company bites the bucket. It needs an active community, and contrary to popular belief. “Build it and they will come” doesn’t actually mean “Make something behind closed doors and dump tars of source code, and they’ll show up and work for us for free”.

The relationship between the community and hifi seems to be that of a customer/userbase and the platform owner, rather than the healthy flat structure of what foss is all about

Why isn’t hifi part of GSoC?

Why isn’t there an official chat room where devs hang out. preferrably on a platform with the same values as what hifi is supposed to be (irc, matrix)

Why isn’t there a list of easy to grab issues to work on for newcomers.

Why is the build pipeline this bad, pulling in dependencies from the web in about every stage of it?
Why isn’t there a linux build?
Why does the build scripts use an external package manager???

I’m sure I came off quite harsh in this post. But that’s how it is when you come in with the expectation of something great, but which then turns out to in reality be quite far away from ideal.

The product doesn’t need to be amazing or perfect, it is under development. But it needs a soul and community engagement. lest it be no better than VRChat

7 Likes

#95

Regarding that. You can use normal naming conventions for this. For example, if you have a DNS record listed to say mydomain.com, and that server the DNS record is set to is also the High Fidelity domain, then putting in mydomain.com into goto will actually work. Same goes for IP addresses.

That being said…

There is a telegram and… that’s it. There used to be a Gitter and the old in world chat system was IRC based, but those both have more or less been left behind.

2 Likes

#96

There’s also the unofficial Discord server. People do talk somewhat on there.

I’ve not been given the relevant info about the Telegram channel, tho, even though I’ve seen it mentioned in passing occasionally, so I’m not on it. 0o

1 Like

#97

The HiFi Telegram channel is pretty dead. The unofficial Discord is a lot busier. I have read through all the comments on this thread and I must say that a lot of long time users and supporters of the platform seem to be pretty upset and frustrated about communication with the company and not having their feature requests and bug reports being acted upon. What do people think is the ONE THING that Philip Rosedale and his team need to do to turn this situation around?

0 Likes

#98

Also, why is there no official Discord server for High Fidelity?

0 Likes

#99

Definitely been wondering the same thing in regards to an official discord server. As far as what HF can do, I think (and I am only representing myself in saying this) that two things would help significantly:

  1. Focus on the community and make them and their feedback a real priority- half of the things @Dandellion and others have said are things that Hi-Fi can easily implement and that have been asked for some time, such as a proper communications channel, settling for real on an issue tracker and removing the ones that aren’t being used, and even Linux builds packaged with the Steam release instead of having to be built by users manually. The biggest common concern I’ve seen in this thread is that no matter what Phil or HiFi says, the vast majority of users are convinced that their feedback doesn’t influence the development of the app in any way shape or form- and that’s not conducive to keeping people invested, even people who’ve been using the app for years.

  2. Embrace the open source model. Many others have talked about this more than I will, but High Fidelity feels open source in name and not much else- it’s one of the key advertised features on the site, and certainly HiFi is more open source than, say, VRChat or Sansar- but it doesn’t carry home many of the key benefits of open sourcing, many of which could really help bolster the development of this app, and allow the community to influence the development not just in feedback, but also in practice.

Just my two cents, but I think seriously special effort needs to be taken to assure users they do matter or they could very well jump onto another, more open-sourced or decentralized social VR bus when it comes rolling around the corner. (Or I guess Hi-Fi isn’t really considering itself a social VR app anymore?)

1 Like

#100

There is now a proper community owned and operated discord server.

You are invited to join the Federated Hifi Users Discord!

https://discord.gg/pVrfqE

8 Likes

#101

That discord link is invalid.

1 Like

#102

Discord sucks in many ways. You found one.

2 Likes