Hardware and Emulator issues.
3 years ago
South Carolina, USA

As most of you know, we've recently had results showing that emulator and hardware speeds differ from one another. Emulator speeds are, for the most part, found to be faster than original hardware. Emulator speeds also differ depending on which emulator is being used.

We do not currently know how much faster ALL of the emulators are or have "accurate" differences between emulators and original hardware. We do have some semi-roughly estimated time differences of some emulators though, thanks to @estacaco.

This thread will be to discuss what we all think we should do to remedy this issue.

Currently, we have a couple of options that we could take:

  • Add time to emulator runs (depends on what emulator was used)

  • Split the board to contain both "Original Hardware" and "Emulator" tabs; each with their own "Easy", "Normal" and "Difficult" categories.

Please keep in mind that we will be taking the option that is mostly agreed upon. The board will not stay as it currently is so your voice is needed in this discussion.

If anyone has any other options that we could take, we would love to hear them from you. We're trying to find the most fair, reasonable, and accurate solution to take.

Something to take into consideration while discussing is that other Genesis boards on the SRC website are also going through this same issue, at the moment. Some of them have come to their own conclusions on what the best course of action was to take.

Also, if anyone would like to assist in finding the correct timing method for emulation, I will create another thread where we can post our findings.

Thank you all for your patience.

Edit:

Form: https://forms.gle/NKXw9r7bmwDcYtbX6

Edited by the author 3 years ago
Tomarrow and Bronkel like this
Nova Scotia, Canada

Given that we have a pretty equal ratio of runs that are on original hardware and runs that are on emulator, I think splitting them would be a good choice. Adding time to runs might turn into a funky situation if different emulators would require different amounts to be added

Bronkel, Insomnimatic, and Sarabi like this
Orlando, FL, USA

Personally, I lean more towards a board split. I think it would be a better choice than trying to figure out the time to add when we can just have it split while being 100% sure what is accurate. I don’t believe having a time that we THINK is right and fair by almost guesstimating is better than just knowing what we’re putting on the board is fair for everyone. It would never be 100% accurate as some have shown with adding time.

I’ll add more when more people respond.

Edited by the author 3 years ago
Bronkel, SaveTheWorld_JG and 2 others like this
Victoria, Australia

With testing that has been done so far what are the time differences like?

Insomnimatic and Bronkel like this
Hesse, Germany

I for myself wont bother much about a split board. Why? Because a leaderboard is mostly there (for me) to better judge my own time (and to see other fantastic runners, learn new things and get better). How good my time really is I can only ascertain within a context. But to be able to set a run in a context times must be comparable. It seems like it's, unfortunately, not the case with hardware and emulators. So yes, split boards makes sense.

Sure, it looks nicer if all is in the same table, but only aesthetic reasons don't apply much here as an argument I think.

I do have the same question PWD has. Just because I'm curious about if there are notable differences between some emulators as well and that leads me to a wish I want to add: That we at least make emulator runs as comparable as possible. That means to maybe even ban (the most) inaccurate emulator(s). Also that we maybe use the "Resources" tab to guide new runners to emulators that should be used and help them start going (and add that it in the rules of course).

Canada

Alright so, the only emulators we should consider using is Retroarch using the Genesis Plus GX core or Blastem core, or Bizhawk with the Genesis Plus GX core. those 2 cores are accurate in game speed. The only issue that you seem to run into consistently with every emulator is that scenes load/transition faster. ALL emulators that I have tested have loaded 1 maybe 2 frames faster (1 frame = 1/60th a second or 0.0166666667 seconds) than hardware (I've compared both original carts and Everdrive with no differences in load or game play) added time would technically be reliant to how many screen transitions the player has in their run. Essentially for each death a player has in a run they would need to add that 1-2 frames of time for each one to account for death scene transitions (exception being pit deaths as it never changes the scene). The time difference is 0.04s per scene change/transition (rounding up to 0.02 seconds a frame and assuming 2 frames every load instead of 1-2) from what I've found. There's 12? if I'm not mistaken scene changes/transitions in a normal deathless run totalling 0.48 seconds (actually less but we're rounding up and using worst case). so all in all less than half a second in any WR or even top 10 run. Rounding up to .5s and adding .5s to all Emu runs would make them more accurate to hardware and even give hardware advantage over emulation.

On another note, all DCG (Disney Classic Games) will have to be separated to its own board regardless of a split or not as you can't change it's emulation or cores so the game as a whole will run faster (game play and load times).

Edited by the author 3 years ago
8BitsOfJoy, Insomnimatic and 3 others like this
Victoria, Australia

In terms of not splitting those results are very manageable, especially when we’re still only counting by seconds, half the time when adding the 0.5 time to emulated runs there would technically be no change to the time (with how it’s displayed on the board).

If it wasn’t a static number of difference, or even if that number was a few seconds higher which I could see annoy people by having to add to there times possibly taking away spots higher on the board because of a penalty system basically, but a consistent 0.5s, I feel is a very small margin to split over.

I can probably be convinced either way really, but atm I’d prob have my vote in keeping it together.

Florida, USA

If Retroarch with Genesis Plus GX core / Blastem core and Bizhawk with Genesis Plus GX core are that accurate, then I also feel that a board split is not really necessary, as long as those are the required options. In my personal opinion, if 1-2 frames per transition is the only real difference, then that doesn't seem significant for this kind of run.

United States

console lag test

emu lag test

It's impossible to account for the differences in how the emulator produces lag compared to console in this game. There's no way to accurately add a perceived difference.

Canada

Bat/sprite count was larger in your hardware test than emu test.

Victoria, Australia

It’s been a while so I’ll post my thoughts again. Regarding lag it is hard to compare 100%, but if we can’t accurately prove its within this game I think we have to assume a difference doesn’t exist. There’s only a few areas in this game that lag does come into play and now I’ve played emu & hardware for an extensive time I can say I’ve never noticed a difference. That doesn’t mean there isn’t one, just my personal experience. If it was a noticeable difference I feel like we would have noticed it ourselves in the time we’ve all spent watching others play different versions. And if it does exist so minimally in the few sections that the lag does effect us I don’t see it being an issue worth debating over.

Again if it can be proved and we can detail the time differences and line all the sprites up etc so it matches that would be best but at the moment the only game specific results we have to base our thoughts on is Esta’s testing. If others want to test as well to compare results that would be great but at the moment that’s the most detailed conclusion.

And again my opinion is there’s no reason to split the board based on those results. And the emulators allowed be restricted to Retroarch using the Genesis Plus GX core or Blastem core, or Bizhawk with the Genesis Plus GX core.

8BitsOfJoy likes this
Victoria, Australia

Sooooo, it's been over a month on this since the discussion was started in discord. Can we put it to a vote if there's no other opinion's/info? (also there's the other forum topics that could be finalized)

Orlando, FL, USA

I'm gonna say yeah. Seems as the discussion is pretty much dead. I would say we should come to a decision here real soon, gonna wait on Insom to put final say on this and figure out best course of action. Now would be the best time, as the board seems to have calmed again.

But in the meantime, I will say there is no need for a board split if the time is less than a second or a second. Especially if we're not using milliseconds. If it is more than 1 second, then maybe so. (Had a change of heart after further thought on it.)

8BitsOfJoy, capnredbeard and 3 others like this
South Carolina, USA

Alright, this has been a long time coming. I do apologize for letting this topic hang for so long as I have been quite busy lately and have had a lot of things come up. However, this thread / topic seems to be dead at this point so let's move things forward.

This is a topic we seem to be divided on and some have changed their mind on what we should do at this point. Since this is the case, we should use another form/survey to get the general consensus of the runners on what we should do with the board.

Please make sure you take the time to weigh out the options and the option that you choose is, without a doubt, what you would like to see happen. I can imagine this topic will be brought up again at some point, but the results from this form should reflect what the current consensus is.

Please remember that at the end of the day, we're a bunch of people that geek out over a lion tossing another lion from a cliff. Don't let the results of this possible change affect that we're still a community that loves this game.

Form: https://forms.gle/NKXw9r7bmwDcYtbX6

Victoria, Australia

Hello, its me. I guess we can close this off? Over 2 months of discussion, coming up to 6 weeks of voting.

By unanimous decision... boards stay combined and the introduction of banning use of Emulators that do not / can not use the Gen Plus GX or Blast Em' Core.

Would a/the mods like to adjust the rules and conclude this?

ZeldasKitten likes this
South Carolina, USA

My apologies everyone, I will be closing the poll and making the changes to the board. As @PlayingWithDick stated above, the board will not be splitting, however, we will be banning the use of emulators that do not use the Gen Plus GX / Blast Em' Core and adding 0.5s to all emulated runs.

Due to the current world record run being performed on original hardware, the previous runs will be grandfathered in.

Thank you all for your patience and understanding during this time.

Game stats
Followers
90
Runs
347
Players
55
Latest threads