Dolphin, the GameCube and Wii emulator - Forums

Quote: Likewise the majority well be unconcerned about advanced AA.

Even these people will be concerned as soon as you also consider performance.
Other forms of AA can provide about the same or even better results with less performance penalty in many cases. 8xSGSSAA VS 9xOGSSAA comes to mind, for example.
You might find yourself at the point where another AA solution fights AA to the point where you dont notice it anymore, but at less performance cost.

Quote:Anyway, you completely missed the point I was making about your being in the minority of gamers.

(10-03-2013, 04:30 AM)Shonumi Wrote: [ -> ]ven when people are vocal about certain changes, the devs still consider how useful it is to the average Dolphim user. If they don't.believe it is useful to all or most, it doesn't make.

Is this what they say?

I have been active in the emulation scene for a long time, or the gaming scene in general.
Do you know OpenRA?
I have written a whole new User interface over the time of several months, for exactly three users!
I got no money out of this, so i did not care if there are 3 or 300 users using my interface.
All i wanted was to make an already great game more perfect.
I've written many Shaders for console emulators just for one person: Me!
Cause i wanted a specific effect and was able to do it.
Do not underestimate this: Some people just love games or consoles.

So who says there aren't any Dolphin Devs now or in the future (pls consider the staff is always changing) who think the same way?

(10-03-2013, 04:30 AM)Shonumi Wrote: [ -> ]It was that such a minority is unconcerned with high DPIs and resolutions as they still want advanced AA. Likewise the majority well be unconcerned about advanced AA.

How do you know what majority or minority think or who is in majority or minority? I don't mean to pick a fight or anything but I can't know what others think if they don't say. And I believe people who visit and/or post in forums frequently are only minority of all users.

...I think people who're so obsessive about AA that they (think they) still need "perfect" image quality on a high-DPI screen need a break from staring at displays. The average gamer (i.e. someone who cares more about the game itself than the fact that their GPU is fill-rate-limited) will be unable to discern between even 4x and 9x SSAA on a higher-DPI display.

More to the point, I have an HTC One (4.7" 1080p screen -> ~469 dpi) and a copy of Mupen64Plus AE, which can output at the screen's native resolution but has no AA options. I actually have to *try* to see any aliasing artifacts on this thing. 9x SSAA would easily obliterate what little visible aliasing there is. On the other hand, aliasing is painfully visible on my 3DS (3.5" 400x240 screen -> ~133 dpi), to the point where I can certainly see a need for exotic AA options.

I calculate some PPIs just for your information. A 27" monitor at 1920x1080 has only 81.59 PPI. For 4K (3840x2160), the PPI is still only 163.18. If I understand correctly, even on a 4K monitor, the aliasing artifacts will be painfully visible according to your information.

To get to the point that you have to try to see aliasing artifacts (469 PPI), the resolution needs to be more than 8K (7680 x 4320). For 8K, the PPI of a 27" monitor is still only 326.36.

It will take quite a while until there is a graphic card that can run games at full speed on Dolphin at 8K resolution.

469 dpi is for a phone that's within six inches of my face. You're not gonna be nearly as close to your 4K 27" screen, so the DPI requirement is quite a bit less. See how Apple varies the DPI for its Retina displays (330 for iPhone, 264 for iPad, 227 for 13" rMBP, 221 for 15" rMBP) based roughly on average expected viewing distance and screen size.

Also, yeah, I know it's gonna be a while until displays and GPUs are good/available/inexpensive enough for this 9x SSAA utopia thing to materialize.

>3 pages of WoT posts

Jesus you guys have been busy. Should I even go back and respond to any of that?

Gabbyjay Wrote:Even these people will be concerned as soon as you also consider performance.
Other forms of AA can provide about the same or even better results with less performance penalty in many cases. 8xSGSSAA VS 9xOGSSAA comes to mind, for example.
You might find yourself at the point where another AA solution fights AA to the point where you dont notice it anymore, but at less performance cost.

Or not. Consider that there are those that will stick to using resolution alone. We already get a lot of users here who do just that (even when their GPUs are capable of doing more). Obviously they're not as sensitive to aliasing as others such as enthusiasts or even those with a moderate need for graphical enhancements, but there are a lot of people who would forgoe AA entirely or like myself, stick to high resolutions and MSAA. It's not a given that performance is part of the equation either. Case in point, I stick to MSAA not because it's cheap on GPU resources in comparison to SSAA, but because it's good enough for me.

Gabbyjay Wrote:Is this what they say?

Pretty much. Hop on the Dolphin IRC channels and ask them yourself. One of them (degasus) already mentioned that RG/SGSSAA would be a useless enhancement. If they can't see how something is useful to more than just a handful of people, they're not inclined to work on it. A couple of my suggestions have been shot down by the devs personally on that basis. It happens countless times here on the forums. And if what you want causes GUI clutter, neobrain himself will hunt you down. :p

Gabbyjay Wrote:Do you know OpenRA?
I have written a whole new User interface over the time of several months, for exactly three users!
I got no money out of this, so i did not care if there are 3 or 300 users using my interface.
All i wanted was to make an already great game more perfect.
I've written many Shaders for console emulators just for one person: Me!
Cause i wanted a specific effect and was able to do it.
Do not underestimate this: Some people just love games or consoles.

But you're one person, not a community project composed of several members with differing opinions. If all it took were someone willing and able enough to add a feature to Dolphin, we'd probably have more options of all sorts available. Even for the devs, they have to convince their peers that Feature X isn't something completely mundane and worth having in the project. The devs have to agree that it's an idea that benefits the project and its users (or sometimes just the other devs, e.g. debugging tools). If you are a Dolphin dev, you could commit new features to master and merge branches left and right, but you'd be hard-pressed to keep such commits from being reverted if they don't follow some consensus among the other devs. Like a lot of FOSS projects (with multiple members that is) the whole process is akin to a meritocracy.

Gabbyjay Wrote:So who says there aren't any Dolphin Devs now or in the future (pls consider the staff is always changing) who think the same way?

And who's to say it will change at any point in the future? As long as delroth is Dolphin Dictator, please try not to be too hopeful :p

xemnas Wrote:How do you know what majority or minority think or who is in majority or minority? I don't mean to pick a fight or anything but I can't know what others think if they don't say. And I believe people who visit and/or post in forums frequently are only minority of all users.

You can decide what the majority thinks by the volume (or lack thereof) of discussion about a certain subject. This is how historians decide what most people in the past believed about a subject, and what a lesser number of people believed; they analyze the available literature to get a fuller view of the dialog of the time. To do something comparable, we need only view the number of responses in online communities about something, in this case advanced AA. There has not been a great number of people asking for advanced AA here or in any of the other emulation communities that know about Dolphin. It's not that you can't tell what they're thinking because they've said nothing; saying nothing has told you that they're not interested in advanced AA, at least not enough to bring it up in discussions (hence the lack of said discussions).

For the Dolphin forums specifically, consider that a vast majority of our users are not frequent members; they're just random users who pop in here with a question or two then leave for the most part. The Dolphin forums only has a limited handful of day-to-day members. Give or take 20, that's a drop in the bucket from our 11,132 registered users, and even more so for untold Guests and lurkers. What's more, the average post per user is around 25.38 (although this is undoubtedly skewed thanks to heavy posters like myself and NaturalViolence), which again suggests that many people aren't frequent day-to-day posters, just your run-of-the-mill user who asks a few questions, suggests a feature or two, then gets back to playing until a problem occurs. A number of users also sign up only to suggest a feature. All it takes is a post of two to suggest that Dolphin use advanced AA, and if it were something a majority of users wanted, we would have seen more activity in this regard. But we haven't, and that strongly suggests that advanced AA doesn't matter to a lot of people. If not that case, then it certainly indicates that the issue doesn't matter enough for them to voice their opinions (as many have yet to do).

lol natural... no need for this, as most of this discussion was only about a purely fictional setting:

What if we HAD high DPI screens.

But the reality of the situation today and for many years to come is the unfortunate situation that DPI grow fastest on the smallest display devices (smartphone and tablets), much slower on monitors and televisions, and slowest on the largest display devises like projectors, where high DPI would be needed most.

It is understandable that you wont turn on 9xSSAA on a Smartphone with FullHD-resolution.
But please consider people at the other end of the spectrum, most projectors will be stuck at 1080p for some years to come, until 4K projectors can be bought for reasonable price. And even then it's only 2160p stretched over several meters height, thats nowhere near as high DPI as a smartphone.
And unlike in smartphones or computer monitors, home cinema projectors are kinda tied to the resolution the movie-industrie offers, and that will be 1080p for quite some time and afterwards 4K even longer.
It will take a looong time till you see really high DPI on projectors, so every available AA-option is welcome, even if considered overkill by certain users. ;-)

---

Quote: It's not a given that performance is part of the equation either. Case in point, I stick to MSAA not because it's cheap on GPU resources in comparison to SSAA, but because it's good enough for me.

And there are users, where 1xnative is good enough for them...
That does not change the fact that more AA-Options give you more flexibility in terms of quality, but also in terms of performance. 2xSGSSA for example has the same EER as 4xSSAA, but at much lower performance cost. Depending on the game, it may be a good option, especially if you need to get your 60 fps and your system is already on the limit. The same can be said about 9xSSAA. More flexibility in options is a good thing.
The graphic card manufacturers provide a vast array of AA-options - for a good reason!

Quote:If they can't see how something is useful to more than just a handful of people, they're not inclined to work on it.

And you're actually working against including/keeping a feature here. Good job man and thanks!
Then there better be more people speaking up for good image quality.

Quote:But you're one person, not a community project composed of several members with differing opinions.

No, that's not true.
There's a huge community behind this project, and many devs with different opinions.
I did it nevertheless, one of the advantages of open source projects.

Remember the option at stake in terms of Anti-Aliasing (custom internal resolutions) is a feature everyone agrees that would be nice to have, maybe some consider it not really neccessary, yet it does not have a disadvantage either - but it needs some work to be done. (We already got almost every other feature i can think of that today's hardware supports in terms of AA.) There are no disadvantages, and i am sure if there is somebody willing to do the work, the other devs won't complain about it.

Quote:And who's to say it will change at any point in the future? As long as delroth is Dolphin Dictator, please try not to be too hopeful :p

Hopeful for what?
Remember what this discussion was about: Do any AA-options other than 9xSSAA make sense (once really high DPI screens are here)?
Now i dont have to remind you that we already HAVE lots of other options: We can choose between many MSAA options, we got Downsampling, we got SGSSAA, we got coverage samples, we got high internal resolutions for Supersampling, we got things like FXAA, SMAA and the like.
This is fun and this is how it should be: We can use (almost) any AA-technique available.
As long as Dolphin stays compatible with them and can make use of them.
They do not even have to be included in Dolphin itself, as long as you can use the drivers to make em work, i am thankful.
So why should I hope delroth has a car accident in the future? If Dolphin continues like this and no one thinks "hey this guy in the forum said any AA other than 9xSSAA is useless, so let's drop em" i would buy him a Pizza any day.

What i would like to see, tho, would be a "custom" option for internal resolutions, as PCSX2 does have.
That would help improve my cartoon shader.

Quote: You can decide what the majority thinks by the volume (or lack thereof) of discussion about a certain subject.

This is plainly wrong.
I really hope i won't need to give an explanation for this! I really do.

Gabbyjay Wrote:And there are users, where 1xnative is good enough for them...
That does not change the fact that more AA-Options give you more flexibility in terms of quality, but also in terms of performance. 2xSGSSA for example has the same EER as 4xSSAA, but at much lower performance cost. Depending on the game, it may be a good option, especially if you need to get your 60 fps and your system is already on the limit. The same can be said about 9xSSAA. More flexibility in options is a good thing.
The graphic card manufacturers provide a vast array of AA-options - for a good reason!

It's not a given that flexibility is part of the equation either. Note that the point I was making, the "it's good enough for me" argument has been a consistent idea I've explained. Again, the majority of users aren't concerned with advanced AA. Internal Resolution, MSAA, SSAA, and whatever else your GPU drivers can force, have offered all the flexibility and performance many want. If people wanted more, they would have made it clearer by now.

Gabbyjay Wrote:And you're actually working against including/keeping a feature here. Good job man and thanks!

What does it matter if you perceive me to be working against your interests (advanced AA in Dolphin) if it's my honest opinion that Dolphin doesn't need it?

Gabbyjay Wrote:No, that's not true.
There's a huge community behind this project, and many devs with different opinions.
I did it nevertheless, one of the advantages of open source projects.

You didn't read the context of that correctly. Your examples in the previous post demonstrated how you by yourself easily made changes to efforts that were solo projects (dependant on no one else but you). You had all the say to bring about whatever changes you wanted to. Even if someone said "hey, I think you should do this", it's of no bearing; you could ignore them, and they couldn't revert your changes. On the other hand, Dolphin is a community-based project. You're one person; whatever changes you propose has to reach some sort of consensus among users and especially the developers.

Gabbyjay Wrote:Hopeful for what?

Hopeful for more advanced AA. Degasus (one of our main gfx devs) has no interest in this. Delroth (tongue-in-cheek Dolphin Dictator) doesn't seem to take advanced AA seriously either, else he wouldn't have hinted at its being a sort of "placebo". Neobrain (our other gfx dev) routinely broke SSAA on DX9, you can't trust him to fix things, let alone add new AA methods (not really, I jest about this). There's hardly any momentum behind it, and evidently it's going to take some effort to change their minds, unless you or someone else joined and actually started implementing advanced AA (which delroth hinted someone should do as much if there's going to be all of this talk).

Gabbyjay Wrote:As long as Dolphin stays compatible with them and can make use of them.
They do not even have to be included in Dolphin itself, as long as you can use the drivers to make em work, i am thankful.

Paul and I say that advanced AA beyond 9xSSAA will more or less become a moot point. You are not so certain that will be the case for others, and certainly not for yourself personally. That's our positions if I'm correct. That was a lengthy sub-discussion in the thread's larger topic (implementing RG/SGSSAA in Dolphin) which we seem to be returning to.

Gabbyjay Wrote:They do not even have to be included in Dolphin itself, as long as you can use the drivers to make em work, i am thankful.

That's another point against including it in Dolphin then. Why duplicate the work unless you can do better than the driver's implementation of advanced AA? Again, for many, IR, MSAA, and SSAA provide a good-enough experience, for IQ, performance, and flexibility. For the enthusiast, are not they (yourself included) more likely to have GPUs and drivers capable of advanced AA in the first place? If so, than it seems like advanced AA should be left to the drivers as it is today. If certain drivers don't provide advanced AA (hello Nvidia Linux drivers, MSAA or FXAA only) gamers should bring it to the attention of the company.

Gabbyjay Wrote:So why should I hope delroth has a car accident in the future?

I don't know, why would any one ever hope that? Where did you get that idea anyway? No one ever said the best way to get rid of delroth was with a car anyway (he's stronger than that). At any rate, Paul is next in line to the Dolphin throne, and he's much worse than delroth by a country mile (not really :p).

Gabbyjay Wrote:This is plainly wrong.
I really hope i won't need to give an explanation for this! I really do.

You kind of need to. It's bad form in any discussion to say "you're wrong" and not give any reasons. Even if you think the reasons are plain as day, obvious to you does not mean obvious to everyone else.

Again, the majority of users here on the forums do not care about advanced AA enough to ask for it in Dolphin. This is a fact. If they cared enough to ask for it, then they would have asked already. Thus far, they have not. Aside from a handful of threads here, not many have even discussed advanced AA in Dolphin. The volume of discussion informs you about what people are thinking. This is a method that's been used countless times by historians, as I noted previously. If you think this methodology is incorrect, you're going to have to contend with a lot of seasoned professional researchers. As it relates to our humble emulation community, only a few are talking about it; everyone else, by that measure, hasn't even felt the need to discuss it.

(10-03-2013, 12:30 PM)Shonumi Wrote: [ -> ]It's not that you can't tell what they're thinking because they've said nothing; saying nothing has told you that they're not interested in advanced AA, at least not enough to bring it up in discussions (hence the lack of said discussions).

Thanks for you reply. Just wanted to know how you think. I disagree that saying nothing means they aren't interested but we can have different opinions. It would be quite chaotic if someone talked about something then 1,000 users just posted to say "I agree".