Is the Marvelous window wider in DDR A compared to DDR 2014?

Don't mean to rain on anyone's accomplishments, but just wondering, given how obtaining MFCs in DDR A seems to be much easier compared to DDR 2014.

Or did the removal of 'magic BPMs' just really have that big of an effect? (I suppose one way to test this would be to pick something with a magic BPM, play it a bunch of times on both 2014 and A, and see if there's any discernible difference.)

On that note, how did DDR A get rid of magic BPMs anyway? Is it like ITG now where all timing windows are millisecond-based rather than frame-based?

As far as I've been told by people, the Marvelous window remains at +/-1 frame on DDR A, as it had been in previous games. All other windows have remained intact.

However, the polling method significantly improved with the new software. "Magic BPMs" used to exist because of the way the system used to estimate where the windows are for each note, and at 60 frames per second, there would be 3,600 frames per minute. It used to be that a window started on a whole number frame and ended on a whole number frame as well. This meant that unless the song's BPM was a factor of 3,600, the windows would move slightly from note to note due to rounding. This is why it was likely to drop a Perfect even if you were consistent throughout the entire song if you were playing a song outside of the "magic BPMs".

On DDR A, the windows are relatively more nicely fixed, although the polling continues to be frame-based. So players, not having to deal with fluctuating windows anymore, can MFC more songs just by keeping consistent. I'm not the expert on how it runs the frames, though, since this is all second-hand info.

So in summary, what you're seeing now on DDR A is a bunch of songs getting MFC'd that wasn't likely to be done before only because the windows moved from note to note on the previous mix. It still takes an incredible effort for newer MA players to MFC their first song, as always.

---

This is still all second-hand and may have inaccurate info, but from what I've gathered, the game visually runs at 60 frames/second while steps are polled at a faster rate (I believe either 150 or 300 frames/second). In 2014 and earlier, if you stepped on the "border" of the marvelous/perfect window, the game randomly assigned the judgment, while in DDR A, the game will automatically give you the better judgment. The existence of "magic BPMs" is due to aligned quantization of frames with the song's beat, as NewbStepper said. With that explanation, however, I suspect that there are still very slight differences even on DDR A in certain BPMs being easier to MA than others, but there's no definitive way to confirm this without actually seeing the game's data for ourselves - the explanation I just gave is probably a simplified version of very complicated stuff that's going on underneath the surface (for example, another factor in all this is the operating system/type of cab e.g. SD cab, X-cab, or white cab, since the cabinets themselves actually run the game slightly differently).

As a side note, from talking to a couple others, MAing previously magic BPMs on DDR A is slightly easier than it was on 2014. Which makes me kind of surprised that Unreal hasn't been MFC'd yet on A given it was 1 perfect from an MFC on 2014.

In my experience, MAing songs with magic BPMs is no easier on A compared to 2014-.