Star Citizen Dev Responses – VR & Network Updates
Welcome to some more Star Citizen, today looking a CIG Dev Responses from the Spectrum Forums in Regard to VR Integration, NetCode Updates, New Cloth Tech & The Cooling Systems in game.
With VR we have a post talking about integration for VR in the wake of HL:Alyx.
Silvan CIG (Junior) Engine Programmer says:
“One of the reasons why i joined this company is to bring VR to Star Citizen. And there are also other people here in the company as dedicated about that as me.
Implementing VR into the rendering itself isn’t a big deal, but as others already have pointed out it’s having support from the game itself in terms of gameplay. I’m mostly concerned about tying it into our animation system and what locomotion we can offer. Nobody wants to have teleporting players in the verse right?
There are lots of unknown questions which need to be solved and requires a collaboration from various teams here.
I imagine we will implement VR in stages, e.g. for a first step just being able to view ships in the hangar or using it in a cockpit. But a full VR implementation similar to Half Life Alyx is a whole different story.
I can’t and won’t make any promises but my personal goal is to have something similar to Alyx :slightly_smiling_face:
Apart from that if we get a NerveGear in two years i will be the first one to implement that straight away. Something like SAO is my ultimate dream 🙂
Anyway, i just wanted to say that there are people here at CIG who are huge VR enthusiasts and we will do anything within our capabilities to bring it to Star Citizen. It will be a huge task, it will take a while, but i have no doubt we will get there.”
They then went on to say – “If possible i’d like to have full body ik support… huge fan of full body vr tracking in vrchat.”
I just wanted to briefly say a couple of things on VR as well. I do think HL Alyx is the best representative of a true VR game that makes full use of the Valve Index! It’s awesome BUT Squadron 42 and Star Citizen would need a lot of consideration in how to implement VR, location motion and movement, do you teleport, how do you jump and move around?
Getting aiming & interaction accurate, then how do you go from FPS to ship controls, if you are virtually moving a stick in a cockpit then tracking accuracy is an issue. Squadron 42 is much more sensible to get VR support than Star Citizen AT LEAST at first… the singleplayer more on rails known quantity nature of it just make it a lot easier to get integration BUT don’t expect VR support for a good few years yet AFTER they have their Vulkan API integration they can then make some plans for VR potentially BUT expect them to be LONG LONG TERM. The Longer they wait the better when it comes to devs skilled in VR & hardware available etc…
Garviel Asked a question
When the new cloth tech is finished, will it be used in props and assets like the prefabs that were being pulled in to build a point of interest? Will we get to see things like a tarp door blowing in the wind or sheets draped over large areas like tents? That little bit of movement would add some really nice depth to those points of interest.
Cory CIG replied
As you may have seen in the Citizencon 2949 demo, we’ve been exploring the use of physics cloth sim and other techniques for wind-responsive props.
You can break these down into two categories, the physicalised cloth sim assets, such as the rover tarp and lab coat and the vertex deformation based wind effects, as seen with our trees, crate tarps and wind flags.
The vertex deformation technique I mentioned above makes use of our asset’s vertex colour channels to set the intensity, phase and frequency of wind for each vertex on our objects. For instance we can have a cable dangling between two points, and mask out the wind on those attachment points so they remain grounded but the slack in the cable will sway with the wind. The main disadvantage of this technique is that it’s a fairly simple noise multiplied by the wind volume parameters. For objects such as flags, which need to droop realistically when there is no wind it doesn’t look that great when you’re not in a rapturous storm. However for trees, cables and other soft-body assets that are tied down it works well.
The cloth sim is certainly a more robust system and can handle a lot of situations with a greater degree of realism but I expect there is some added expense. To what extent? You will have to ask an engineer!
We will be leveraging everything at our disposal to deliver more movement and interest in our props and we will continue to experiment with the cloth tech again as it continues to mature.
Darj asked a follow up question
In the citizencon demo, the trees displayed a rather obvious repeating pattern in the sway. Is that easily addressed?
Also, since wind will be responsible for things swaying now, does that mean that plants in building interiors where there is no wind will stop magically swaying?
With Cory replying
the ‘noise’ itself is a part of the shader and as artists we can affect the intensity and phase of the noise, as well as blend between a high frequency and low frequency noise. I expect that any changes to the way this noise is being generated would be inherited by any object using the wind bending shader param (I’m an artist not a programmer so I might get told off for saying it’s easy!)
We saw some very cool tarp being moved off a Ursa Rover in the CitizenCon 2019 Demo and more recently in Inside Star Citizen we have seen deforming tarps in Prisons as well as plants that spawn bendy limbs when players are close that well… bend…
Very much looking forward to all of this coming together.
Coolers…. are they supposed to help with weapons overheating?
MGibson CIG gave some insight
“When thinking about how of the current cooling system works you need to split it into two separate parts the flow vs the pool.
The cooling rate is how much coolant that’s cooler can prove per unit of time giving you a total “pool” of “coolant” available, each item then has a maximum intake of coolant per unit of time, meaning it can only receive X amount of coolant per second, think of it like a pipe running to the item you can have a HUGE reservoir of cooling but only as much can get through the pipe as the pipe can take the benefit of a huge reservoir being, of course, you can cool a LOT of items or if you have some items with huge intakes you can meet those needs.”
It must be noted this is for the current system and not the final intended system for cooling. We need Physicalized Components and the Item Cache among other things really before we have all of these systems working a dialled in and in a more intended state.
Does SC use Network Compression for its game packets?
Clive Johnson CIG Replied
Yes, we currently use LZ4 for packet compression as it’s very fast and gives us roughly a 40% average reduction in packet size. I want to revisit this at some point and look at whether other compression algorithms can improve on this. Right now we can almost certainly make bigger savings by being smarter about what we send and how often.
When we do look at packet compression again Oodle is definitely on the list of options to investigate.
[Oodle is a real time compression solution for network traffic, greatly reducing the bandwidth required by game servers and is focused at solving issues with MMOs with high traffic.]
We actually discussed it internally about a year or two ago. My feeling was and still is that although static dictionary techniques could theoretically give us much higher compression it really comes down to how good the dictionary is. We’d have to train the dictionary with representative gameplay and it would be all too easy to bias that training such that the resulting compression doesn’t work as well as we’d like in all situations. With a game as large and diverse as Star Citizen the training data would need to be enormous, which presents problems in how to generate it without bias. Given where we are in development with new features and content being added every patch, as well as existing code being updated, we’d need to retrain the dictionary frequently. In other words, it’s not really the right time to try such an approach. Besides, we have a little thing called Server Meshing we’re focusing on right now. Finally, it’s not clear (to me at least) that a static dictionary can be large enough to handle the diversity of a game like ours and still provide good compression (larger dictionaries require more bits per substitution symbol, reducing compression efficiency). But it’s still an option I want to investigate when the time is right.
We do know that CIG have lots of Networking Tech planned this year, updates to SOCs, the iCache, Full Persistence, Server to Client Code Refactor and work towards Server Meshing.
CIG have also said – Inside Star Citizen is going on it’s 3 weeks break after this week BUT they’re still planning on releasing a 3.9 Pillar Talk – more info on that when they have it.