What happens if you don't hit 90 FPS?

[mabsey]

As per the title really, when CV1 is released, if I don't manage to hit a Constant 90 FPS and say I hit 80 FPS, whats going to happen, no low persistance and experience broke with Judder? Will there be no 75 FPS fall back mode, so it runs like the DK2 for example?

The more I'm thinking about this, the more I don't want to go spending another £600 on machine upgrades just to get a better experience on top of the price of the CV1, I've got a I7 with a GeForce 970 GTX now and there are some titles I'm struggling to get decent frame rates without turning the settings down. I feel I've gone from being on the fence, to really wanting the CV1 to being back on the fence again.

Mabs

[Anonymous]

You know ~ there isn't anything stopping the community from making such a benchmark ~ just saying. DX12 is starting to show up in supported APIs as well. Simply make a benchmark that test a few things and report back base on DX11 or 12.

The only problem is how much "stuff" do we put in there and say you must meet 90fps for it to count. At least with PS4 we know the limits ~ so I guess you could just in theory create a demo that pushes the recommend min hardware to the point of 90 fps ~ but as with anything ~ it'll only recommend because software is not all created the same.

[Anonymous]

Ronson, when has Oculus ever said that there would be no 75Hz fallback? I've never heard or read anything supporting that statement.

[RonsonPL]

"jngdwe" wrote:
Ronson, when has Oculus ever said that there would be no 75Hz fallback? I've never heard or read anything supporting that statement.

Just my observations:

1. Oculus don't want to invest in PC VR quality, but....
2. They also have invested already, and talked much about "poisoning the well".
3. 90Hz display is useless if Oculus allows the devs to fall back to 75, since 90% of them will use it to increase the profits (ancient PC owners would buy their game).


They broke lots of their words, but I don't think they're able to do such ridiculous thing.
When Palmer was confronted with the criticism of their lousy CV1 display, his main argument was it's great because it's 90Hz.
If Oculus decides to allow 75fps VR on CV1, Palmer would made a fool of himself.
No. I really don't think such scenario is possible.

Maybe there will be some legacy modes, with a big message board shown while user launches a DK2 app or something, but in new games? I don't think so. Not on CV1 and DK2 will be not important for the majority after CV1 releases.






Mradr
I'm not a person you should ask about how such system should look like.
I have not enough knowledge about coding, about different types of calculations being required for this or that. About their respective latency, that looks like this, this, and that.

But what I do know, is that it's a very simple thing to do over a 2-3 years, for a company that has so big funds as Oculus has. If they just start to create this now, it might be a little complicated, but nothing you can't deal with when you have just one good coder and half a year until CV1 debuts.


Benchmark would have to be standarized. Made by Oculus and then connected to the rest in every VR game/experience that has Oculus support.
The features I would consider:
- forcing the settings and blocking them accordingly. PC cannot maintain the res? Allow only lower. Cannot maintain the draw distance/LoD settings? Lower them. Cannot be quick enough for physics calculations or something else, that is crucial for the game - block the game altogether.
- not too long pass. If it's mandatory for everything, then it cannot be 5 minutes long.
- It could be useful to focus on hardware that is much faster than the minimum specs.If, for example, two games report that this PC's overclocked CPU and high-end GPU beats the benchmark for the minimum, scoring 200+% of the required score, then PC could be permanently relieved of benchmarking in any upcoming game. That would be connected with Internet and if some AA/AAA game released that changes the landscape, update from Oculus would be downloaded, resetting the "green light for all".
- disabling the blockage should require basic PC knowledge, so no casual could alter it. A few files copied from here to there, a registry value change etc. Alternatively it could be unblocked after Oculus software forces the user to watch instructional video and then makes sure he didn't just walked to the kitchen while it was rolling. Video clip would have to be idiot-proof and this might turn out to be the biggest challenge :mrgreen:
- a CPU-Z like or AIDA64-like check could be helpful. No need to test a machine that has no chance for passing it.
- A check to see if anything's clogging the CPU/memory in the background, is also not that hard to do. Antivirus software can do it, Windows' Task Manager can show CPU usage. It could be done. And if it was, it could be automatically started whenever game encounters below-limit performance.
For example: we're sorry, but because of our concern about your health, we stopped the game since it produced too low framerate. Please wait 10 secods while we're checking if it's not a software running in a background.
(then)
It seems .... uses lots of your PC performance in the background. Please turn it off and you'll be able to return to the game.
(I'm not going to think about cheaper Windows 10 versions that take away the control over updates from the user. I'm not even thinkg it, see? I'm not thinking about it! I don't even know what I was not thinking about! 😉 )


It would require more work and effort (OK, lots of it), but it could be great if finally we got some standarized performance comparisons. It's about CPU, it's version, the memory configuration, chipset and so on. It would be nice to create a base of every possible configuration and from that - it could be easily accessible data that for example AMD FX-8350 with memory working at "x" speed, can achieve this much calculations of this sort, and this type of calculation will have "y" latency.
I watch PC world since decades, and disagree with all that PR BS about "thousands of possible configurations - impossible to make a game that works on so many machines" we're fed recently when a cheap, outsoruced to some Uganda-placed company, crappy port receives the gamers backlash. Take CPUs released in recent 3 years (AMD) and 6 years (Intel). It's not that much. This is a major problem with latency. GPUs are more easily scalable, more predictive for the devs.

[Anonymous]

"thousands of possible configurations - impossible to make a game that works on so many machines"

Actually, I hate to be "that guy" but it was kind of true back in the day. Each card had something different the other did (nivida was known for having some weird extra features to keep a head) but no one else could use from the bit level, to resolution, to extra features ~ and vis vers for other cards at that time. It was changing so much that there were more than just 2 big players.

Now a days it's not as bad as it used to be. Instead of having 6 different card makers ~ we're left with 2 super BIG card makers that normally are about the same in terms of features and performance. With that said, I feel it's not as bad as it once was and they really don't have a reason to say anything like that "unless" time is a factor. Just because a new release card comes tomorrow allowing 120FPS ~ doesn't mean Oculus should create something for it before it's release ~ it's still always best to create what is available now and allow time to pass to take advantage of faster cards.

Why? Because it takes "time" for newer stuff to grow out to the rest of the world. For example, you still can't really find a Fury X card as the time of writing this. This means ~ I would except that I wont see as much content to show off with my new card yet:) That also means I don't except that some other old game is going to look really good on it either just because the card is stronger to display more ~

[RonsonPL]

Aaaand back to off-topic I go again 😉 :

Exactly, back in the days, it was more difficult, and yet, devs didn't complain, made proper PC games and it worked fine.
Now we have x86, well-known GPU equipped consoles, and some PR f...s from Ubisoft and the like, dare to say the reason for not working, pre-alpha version released as final, or the poor performance, or a year of delay to console versions, is because there's thousands of different PC configs, and optimization is ultra-mega-difficult. Bullshit.
I remember when games had different sound engines for different sound cards, there were separate versions of games for CGA or EGA and VGA cards. We had simultanously 386,486, co-processors and no co-processors, then we had Glide, DX, OpenGL and sofware modes at the same time. And yet, proper PC games were created. And now we have just a couple of AMD and Intel CPUs, ultra-slow consoles to port from, and they (publishers) utilize the lack of knowledge among gamers and (that's scary) some game journalists, to sell such PR BS to the public, as a valid explanation of them shitting on the PC version and expect people to pay full price for it (attitude like outsourcing AC:Unity and Batman:AK to the cheapest developer on Earth, because PC gamers will buy it anyway).



About the top-end cards. I don't say we should see a AAA game designed in such a way that while it's released now, it REQUIRES Radeon Fury or 980ti to run properly. But if CV1 is to be released probably a full half of a year from now, and when it happens, next generation of GPUs will be just around the corner, and there's noone with a gun pointed to publisher's head telling him there must be sales to 10% of the initial price a few month from the release of the game...
That's another matter.

[Anonymous]

I would consider it being on topic because it goes one and one with FPS, but how games are ported is another story all together xD (Not a fan of ported games actually - they always feel like they are lacking something off and on. Kind of like a movie and a book - just stuff missing you know?)

So I think we agree that we should have something to see how well a PC needs to perform for a VR game before allowing the green light to say it should run the rift. I'm pretty sure we can throw something together as a community base off the recommended and have it test base off recommend values.

I don't think they should allow 75 fps fall back either unless there is real reason to and I don't see the reason to allow such a fall back anyways. Nothing truly happens if you fall below the 90 fps - we just don't recommend falling below that value in the first place. Plus, I don't remember them saying there was a 75 hz fall back to be honest ~ I remember them talking about it - maybe - but most of the time I hear it has been from forum members.

we're sorry, but because of our concern about your health, we stopped the game since it produced too low framerate. Please wait 10 secods while we're checking if it's not a software running in a background.

That's fine, but I would really consider that back burner for atm. There really isn't anything "unhealthy" about it, but I know what you are saying. There will be times tho that some software might push the game a bit hard ~ so this might cause some unwanted problems as well if you think about it xD

Either way! Get a PC that can support the CV1 - it's not like you have to upgrade the whole PC - just upgrade the graphics card and maybe power supply.

[fitzsteve]

On the DK2 I've noticed when the frame rate drops it normally halves to 37 so it's a judder fest, this is especially noticeable with project cars, the head tracking gets very juddery, it's not a nice experience. I always aim to get games over 100 so that when I do drop frames it still stays above 75