Hi guys,
As you know, KdL and HRA!-san work really hard towards MSX FPGA improvements, but, they have their own goals (which if I'm not mistaken, is to have a full Turbo-R on OCM-SM / Second gen OCM devices (SM-X, SM-X Mini and SX2). On the other hand, there are quite a few things I would like to see implemented that perhaps doesn't align with their final goal (as that will occupy space on FPGA that might be needed for Turbo-R related stuff, 22K cells seem like much, but OPL3 and OPLL already use quite a lot of that).
So, to not bother them, I'm doing parallel builds, I did 3.9a and now 3.9b, incorporating features I did for myself and perhaps someone else will like. Of course, I keep communications with KdL (i.e.: he has picked up my OPL3 fixes), code is on Github so anything there can be picked up by him for OCM-PLD and by HRA! for his DE0 project, if they see fit.
Those versions were fully tested on SM-X by me and SMX-HB by Fabio Santos and Agenor (that is something that perhaps only Brazilians will have, it is a 1.5gen device that has 15K cells, more than Zemmix/OCM but less than SM-X/SX2, that is a replacement board for Hotbit MSX computers made in Brazil and using the keyboard from it and having the slots aligned with the case, quite nice), SMX-HB versions are a little bit limited (no OPL3 as its FPGA doesn't have enough cells for it, and you have to choose to have either paddle emulation or more sound chips). I don't think there should be any issues on SX2 and SM-X MINI, but I would recommend that if you don't have an ALTERA USB BLASTER to recover using the JIC from the original 3.9 OCM PLD from KdL, you wait until someone else test it, as if it for some reason doesn't work, you will need an USB Blaster to recover using JIC file. 
There is nothing for Zemmix Neo and OCM owners, sorry about that, but it is really difficult (if not impossible) to add new features as those devices are being used to its limits and I don't have a first gen device to test...
Now, introduction made, warning made, where can you get OCM 3.9b by me? In my github:
https://github.com/ducasp/MSX-Development/releases/tag/OCM3....
And what new features awaits those that decide to venture on OCM 3.9b?
- All devices except SMX-HB: OPL3 fixes that now allows proper usage of VGMPLAY 1.3. Original OPL3 code didn't have interruptions enabled, and VGMPLAY 1.3 uses the OPL3 timer feature to generate interrupts to have a high resolution timer, no interruptions, meant that VGMPLAY was expecting an interruption rate WAY FASTER but it was only getting the regular 60Hz interruption from VDP, causing slow performance. Also, the timer was not properly working, causing wrong speed playback (better, but still wrong) once interruption was fixed. Fixed that too.
- All devices except SMX-HB: OPL3 workaround for missing channels. If a song played a channel only on the right channel, that channel was lost. This is due to the fact that OPL3 is a resource hog and to have both left and right audio channels being sequenced requires more cells than second gen devices have. I've made a workaround that allows the FPGA replica of OPL3 to be put on "mono", that is, it will modify on the fly any commands to make a channel play only on left or right outputs and alway make those play on both channels. Doom Soundtrack track 03 is a good example of a song that played bad and that now plays ok, in mono, but properly.
- All devices except SMX-HB Frankysnd: VAUS Paddle and MSX Paddle emulation over a PS/2 mouse. SETSMART -8E enables VAUS (Arkanoid and Arkanoid II, use good msx versions that don't have paddle routines replaced by joystick routines), SETSMART -8F enables standard MSX paddle (remember that standard MSX paddle is not compatible with turbo speeds, if you use it with turbo speeds paddle range is lower and resolution is weird, turn off turbo when using that mode) so you can try GALAGA on Paddle (patch available at http://frs.badcoffee.info/patches.html), on MSX BASIC, HID TEST, etc, and SETSMART -8D will disable paddle emulation and return the joystick port to regular functions (or power cycling the device, as it is the default behavior).
- All devices Frankysnd build: (probably Frankysnd will be renamed to Extrasnd in the next build) SN76489 working like a Franky. As Franky uses ports 48 and 49 that are on switched i/o space and OCM IPL device let switched I/O selected, it won't work initially. VGMPLAY is quite nice on turning off switched I/O before trying to detect the SN76489, so nothing is needed to have it working with VGMPLAY. But to work with COL4MMM (COM\Franky versions) or SG1000 (using /F at the end of command line) you need to either disable switched i/o (i.e.: run VGMPLAY and play any SN76489 song) or use a SETSMART command to reserve I/O 48 and 49 to Franky even when switched I/O is enabled, command is SETSMART -8C to do that, SETSMART -8B will restore normal behavior of 48 and 49 when switched I/O is enabled (or power cycling the device, as it is the default behavior).
- All devices Frankysnd build: (probably Frankysnd will be renamed to Extrasnd in the next build) second PSG working on the known 10-13 I/O address range. Works quite nice with VGMPLAY and all patches, like the brand new Maze of Galious Patch for MSX2 that sounds best when double PSG and SCC is available. ;)
The regular versions exists just in case you want paddle emulation and OPL3 fixes and do not want the extra sound chips (i.e.: for SMX-HB that is the only option to have paddle emulation due to the limited resources of its FPGA, or, for the other devices, people that have MMM, real Franky, etc might prefer to use the real thing ;) )
If you are a SX2 or SM-X Mini owner and test any of those 3.9b versions, please let me know the results so I can update those builds as good for anyone without USB Blaster to try or work on having it working for those devices. ;)
