Following up on Melee 120FPSAbout a week ago, I created Faster Melee - which is a community for players who want to push Melee further. For players who actually want Melee HD to be a reality rather than just talking about it. For players who want to have the least amount of input lag as possible - especially for online play. So I posted a write-up going over the the process of how Dolphin 120FPS came to be and what makes the lag reduction possible. But of course, we ran into skeptics that didn't hesitate to call the process a placebo.
And of course most of these doubters did not even try testing the build from actually playing it. All they would do is mess around in Dolphin and say nothing was happening in game. They really just misunderstood the situation. Of course Dolphin would reveal the moves are still coming out on the same frame. What they weren't acknowledging though is that frame comes out faster. They would go frame by frame or something and give Dolphin inputs artificially and say, "look, it's still the same".
For the people who refused to actually test it by playing it and completely disregard their bodies ability to sense lag, we came out with a test that would show the 120FPS build being one frame faster. But of course, the skeptics again criticised the test saying that we cherry picked our data to get that result and that Dolphin can only handle inputs serially and in regards to which one happened to lag less - that was just random. So I repeated the test a bunch of times to confirm and could not get a different result. The 120FPS build was always coming out ahead. Clearly the deniers didn't care though. They wanted more.
So yesterday I made a 30min test video showing pretty clearly that console, Dolphin, and Dolphin 120FPS all have different amounts of lag. I used a Canon 7D shooting at 60FPS at 1/500 stutter speed, ISO 12500, F/4. I used 4 different lights to properly light the scene. It was crucial to properly light the monitor while also capturing my finger pressing the button without it being too dark. I think it came out perfect for what I was trying to achieve. I purposely went through the footage pretty raw so no one can say I cherry picked the data to get my results and stayed consistent through out the process on how I determined when the button was fully pressed. The Full Video is 30min long because of this. If you trust me that I'm not cherry picking results, I made a Summary Version of the test that's only a little over ~4min posted below. (Lag Results: Console 3f, Dolphin 2f, Dolphin 120FPS 1f)
Shortly after posting this however, critics immediately pointed to my Sewell Wii2HDMI adpater being the reason why my console result lagged more and not because of the console itself. Even though it has already been proven that Wii2HDMI adapters like the Sewell don't add lag and basically act as passthrough devices - I had to make ANOTHER test showing that I still got the same result with a regular 60Hz CRT. So if you want to see that test, here it is below.
Based on these tests, it actually reveals online play could be pushed further - perhaps all the way up to 66ms (buffer 4) for players with 60Hz monitors and 83ms (buffer 5) for players with 120Hz monitors. I'll have to do another test video with the buffer at 16 in the 120FPS build (since that's the buffer 4 equivalent) and see if that's equal to console.
Outside of the doubters calling my initial test pseudo science and people saying it's all a placebo, we run into a potential trade-off issue regarding the way slowing down the game handles inputs. There is some evidence from StabbyMcKniferson that the current 120FPS build could be dropping inputs or ignoring multiple inputs being pressed per frame. While I couldn't replicate this result in real world testing (meaning not artificially inserting inputs via Dolphin), I decided to revert the in-game 4x Input Polling feature. Some testers however swear that it was better online and was worse offline. We're definitely going to experiment more with this and try to figure out what is actually going on when we use this process. But for now, we're going to go with the more pure way of simply letting Dolphin be the force to speed up the game and the game itself only has to slow itself down.
So today we have a new v3 release that updates Dolphin, cleans up all the Gecko Codes, reverts the 4x input polling (still possible to enable for testing), and enables widescreen by default. Yes, widescreen is going to be enabled. We fixed the offscreen bubbles from being off (Thanks Dan), and now there's not really a good excuse to not use it. The slight pillar bars are totally normal by the way. It happens with 4:3 too in the new Dolphin builds because Dolphin now accounts for overscan. Remember, this project is about progressing Melee to Melee HD - a goal we are actually serious about rather than just talking about it. In terms of actual gameplay, it's definitely the superior option competitively. For the record it only adds 15% more to each side of the screen so I really don't want to hear about it making that much of a difference. I heard an argument about it taking away the skill of being good with the offstage bubbles and actually already touched on this here:
By that argument it should be a square aspect ratio or less because that increases the skill cap. It's just not true. If you're on Dreamland still on stage and the other player gets off screen because of 4:3 - that is not increasing the skill cap. You should be able to see your opponent on stage at all times. As for the off stage scenarios, you don't gain vertical advantage only horizontal (which is exactly what we'd need). Saying "being able to make decisions based on info off screen is part of skill" is a LOWER form of skill. You know what's a better form of skill? Actually fucking seeing your opponent and hitting them based on your hand eye coordination and not "guessing" based on only bubble info. Not to mention - it's only like 15% bigger on each side. Which means you get the stage camera situations fixed, you have the doubles situations fixed, you keep the vertical situations the same, and you still have plenty of off screen bubble situations where that "skill" you're referring to would still exist. You act like it completely eliminates bubble situations when it doesn't at all.
So then the argument doesn't even become about what's actually better competitively - becomes about trying to maintain the current status-quo because "this is what tournaments use". Here at Faster Melee, we're not about that life. We want to advance things. However, people completely forget that a lot of TOs DO allow widescreen! It's been a feature in 20XX for people who bring Gaming Monitors to tournaments for a while now. Using Wii2HDMI adapters allows us to use Gaming Monitors without the lag and a lot of TOs are cool with them. People totally forgot that WHOBO 2014 USED WIDESCREEN - and it was a freaking MLG qualifier! Guess what - no one cared - because it's not actually making things worse. Not to mention, for PM Netplay - literally everyone uses Widescreen and it's never been a problem despite tournaments still using 4:3.
Almost everyone uses a 16:9 monitor so this is great in terms of Dolphin compatibility. If you're using a 4:3 or 16:10 monitor, it should still scale correctly so no worries. I also use a 4:3 CRT monitor and can say I'm fine with having letterboxing. Widescreen is awesome if you give it a real chance and acknowledge if Melee was made recently, it 100% would be widescreen 16:9 - just like every other modern game. So let's stop pretending 4:3 is better. Give it a chance and actually try it. I heard the argument that it ruins your 4:3 game but I don't agree with that at all. If anything it helped my 4:3 game because I have a better understanding of how far the blast zones actually are and how far the characters are going out. I can imagine easier where they actually are from seeing it after playing widescreen.
Let's talk about how to find matches on the Faster Melee Discord. Assuming you followed all of the steps in our guide, you might noticed we now have regions added to the end of our names. This only appears in our Discord server so no need to worry about your name looking funny elsewhere. We also made it so players and testers that are active and playing or looking for a match will be in the netplay voice channel. If you don't wish to use your mic that's fine - just hop in to show you're active.
Make sure to download the latest v3 build from our #updates channel in our Discord and read the #welcome channel to understand more about finding matches and using the Dolphin 120FPS build. The v3 build also includes all the latest Dolphin fixes here.
Wrapping up, in one week we have gotten nearly 3 thousand unique users on fastermelee.net, almost 750 users on our Discord Channel, over 60,000 impressions from my initial tweet, and over 580 up votes on Reddit. If you want to try playing online with console-like input lag - Become a tester and join our Discord!
Share this on Reddit here!
Posted on 5/12/2016 by: TruckJitsu