@flyinghead Could be more YUV corruption issues maybe - the title screen seems to be a looping FMV video that plays over and over. Certain parts of this video are rendered correctly, but at specific intervals it glitches out.

Glitches -

Happens with both Reicast OIT and non-OIT.
Yep, happens on my fork as well. Could be a DMA timing issue or some sync problem.
@flyinghead @twinaphex Maybe the yuv dma is not being correctly terminated, or its detection if faulty. It uses the ch2-dma. If you look at the page 57 of that doc i've sent you.
Actually it's a bug I introduced myself... Sorry
See https://github.com/flyinghead/reicast-emulator/commit/f18aa63a670248ff6595e37f3eb3940f295afa56
Great job @flyinghead :1st_place_medal:
Yesterday i compared spg from reicast master and libretro. There are some differences on CalculateSync but i'm not sure which one is more correct, or not being correct works best on reicast
Thanks!
There is still a remaining problem though: the first FMV video hangs at this point:

It happens both in upstream and in the libretro OIT core.
The game doesn't hang and pressing start leads you to the title screen. After a while the first FMV video is played again, but doesn't hang this time...
@flyinghead talking about a cliffhanger. :rofl: This with dynarec? Any difference with interpreter? If not it may be a DMA issue.
Same with generic recompiler
Some of the SPG changes I carried over from MAME. The only thing I happened to notice as a result was that I got better frame pacing in the core vs. before.
Anyway, awesome commit. I will be sure to backport soon.
Backported the PR. Seems to work fine now.
Weirdly enough though, the FMV intro doesn't hang for me at all.
Try to see if the EG hack is being applied to your copy of the game -
https://github.com/libretro/reicast-emulator/blob/master/core/rom_luts.h#L49
Indeed, it wasn't applied. And that fixes it. Thanks!
If you had to add another entry, let me know what line you had to add so i can add it to the list.
I'm running the USA release but it's been "renumbered": RDC-0117
Hmm sucks, from looking at it, it seems that this serial number corresponds to another game -
RDC-0117| The king of Fighters '96 Collection (NEO4ALL RC4)
Can we find some other heuristic to more accurately identify games by? Seems a lot of the GD-ROM rips were renumbered/mislabelled.
Well I confirm its id is RDC-0117 although it should be RDC-0120 according to a reference I found by googling...
Using the Software name would be fine in this case: "BOMBERMAN ONLINE"
OK, let me see if we can maybe catch both of them at the same time through a software name lookup instead.
I will close this for now since the main issue is fixed. Reminder to self: I still need to add the heuristic for detecting both versions of Bomberman Online.