leecher1337 / es1969 Goto Github PK

Midi implementation in win2k kernel mode + user-provided file is a rather dangerous combination.

Mostly because the midi implementation doesn't sanitize the bank format. At least, I've gotten the NT4 extracted implementation to crash program with garbage data(reading bank out of a naked malloc, lol).

Adition to suggestion, user mode component should send FM writes one buffer at a time(flushing) to reduce context switches.

Of course, as a bonus, this lets the user diddle with the implementation.

I know the original driver doesn t do it. But would it be possible to implement some things like chorus, vibrato rate or reverb type or channel s master volume (maybe filter resonance isn t possible in theory).

Sorry to bother but I already bought the card but unable to use it. I wish I knew earlier that there is no official x64 driver. Here's the dump file.
121923-24882-01.dmp

Still trying to build the driver myself but the learning curve is quite steep.

Noticed you used these FNUM values, presumably copied from the original driver:

I believe those were calculated with fs=50000 (instead of 49716), so the pitch is about 10 cents off. Not immediately noticeable, but might become an issue if someone uses this for music production and expects it to be in tune with other sounds.

The following numbers will be more accurate:

517, 547, 580, 615,
651, 690, 731, 774,
820, 869, 921, 975

On the other hand, this could clash with any fixed-pitch voices in the existing patch banks, so changing it might not improve the general use case. But maybe that's only percussion. You decide :)

This is compilable for 32 bit win2k/xp, yes?
Might be interesting to try in place of original.