Replay issue

[PaperclipGames]

It has happened several times to me already, and I couldn't find a thread about it by skimming through a few pages.

Examples are my latest two replays I tried to save:

I think it's pretty obvious what's wrong with them. It has been happening on occasion before, but it looks like every replay I try to save now is showing this offset error.

Edit: It appears that the replay viewer tries to use speed 2.10 while it's actually 2.02. However, using the standard settings for replay viewing still won't fix it.

[Velocity]

Re: Replay issue

What engine were these saved in?

[PaperclipGames]

Re: Replay issue

I used the Beta engine, stand-alone, using Flash 9.
I am using a mildly old build though, so if it had already been fixed after the 16th of November, my apologies. It doesn't ALWAYS go wrong, just sometimes.

[Magic_V2]

Re: Replay issue

I read somewhere on here some time ago that speedmods should end in a 5 or 0.
I don't know the factuality of that statement, but using my settings (which are set to default 1.0 at the moment) your replays are working fine for me.
Is there any specific reason you're using 2.02? 2.00 or 2.05 may fix this bug, and it isn't really too noticeable a change.

Edit: Manually setting my smod back to 1.0 after changing it broke the replay again. I'm not sure what's going on here.

[PaperclipGames]

Re: Replay issue

I use 2.02 because I used 2.00 at first, went to 2.05 a bit later, then wanted to be inbetween and used 2.025. However, the replay bug seemed to occur very often with that speed so I just went with 2.02 - which apparently still produces errors sometimes. I'll try just playing with 2.00 and see if the bug still appears - thanks. :3

[Velocity]

Re: Replay issue

Your speed setting isn't liked by the engine. I can edit the saved speed on older replays, but if you use speeds that are 2+ decimal places, problems are likely to occur.

[PaperclipGames]

Re: Replay issue

I normally use speed ×2.02. It's been other values though, such as 2.15 or 2.28, but 2.02 is the speed I've been playing in for at least the last six months (and which were used to produce these replays). I don't mind much playing in 2.00 or 2.05, but I still think the replays should work fine in any speed. :P

[Velocity]

Re: Replay issue

It's more about how the math is done for moving the arrows, odd numbers cause rounding issues from movement.

310 x 2.02 = 626.2

626.2 / 31 = 20.87333333333333px each frame.

Flash is unable to move arrows on such a small level so it rounds to the nearest sub pixel, such as 20.8px.

The way the judging system works is based on distance from the receptors, but on extreme numbers like 20.87333333333333, the exact distance is never reached because of the rounding, and is never setup correctly. This makes notes unable to hit.

When the older engines were made, there never had user settable speeds, so extreme cases like this were never an issue.

If you've seen the R^2 engine, this isn't the case anymore since I compare how many frames it is before it reaches the receptors and not the position.

[Magic_V2]

Re: Replay issue

Your speed setting isn't liked by the engine. I can edit the saved speed on older replays, but if you use speeds that are 2+ decimal places, problems are likely to occur.

Do you think a short-term solution could be to have the replays force the smod to round to the nearest tenth -before- saving?
It works in my head, but there are plenty of illogicalities swimming around up there so .

[Velocity]

Re: Replay issue

The only real problem with that is only .0 and .5 are valid then too, and on some number even those are invalid. Ultimately there's no real solution to the problem unless the judging system is slightly rewritten, which is not something I want to do.

[PaperclipGames]

Re: Replay issue

Alright, fair enough, I'll just switch back to 2.00. I more or less just wanted to let you know about the existence of the bug.
Maybe it's possible to force the player to only enter valid numbers in the option box? That'd solve the problem immediately, too.