Virus TI/TI2 firmware RE — progress update, and a request for board photos
For those who are contributing or giving constructive feedback: Thank you! I mean it! It's been a slog.
For those who are nitpicking that the original post was too long AFTER they chose to read all 5 pages: I guess I should start a Patreon so you can help fund the project? I mean, at least then you might be entitled to something. ;)
USB
Both vendor-specific USB interfaces carry standard USB-MIDI packets. The device just never declares the descriptors that would say so. There's no proprietary protocol in there.
That makes a replacement driver userspace work (libusb, WinUSB, IOKit) rather than a kernel driver.
No signing, nothing to break on the next Windows update. The shipped VirusUSB.sys is a kernel
driver dated 2014, which is why Win11 turns it away, and it's a Ploytec OEM driver rather than
Access's own code.
FreeRangeEngineer pointed out in the last thread that the Linux quirk only switches on MIDI, not USB audio. He's right. The rest of it looks like it's controlled by ordinary Global parameters that Access's own plugin writes over plain SysEx at connect. One of them is called "Input and Output Configuration" and it decides whether USB1/2/3 appear as outputs at all. If that holds up it's reachable over MIDI today with no driver involved. I haven't tested it yet and I'm not claiming it until I have.
No promises on a working driver or a date.
What I'm actually asking for: photos of a TI2 board
Nearly everything I know about this hardware physically came from other people's teardown photos, and every one of them is a Virus TI. The host MCU, both DSPs, the SRAM, the codec, the USB controller and a 4MB flash chip were all read off other people's pictures.
I have never found one clear photo of a TI2 board.
If yours is already open for a repair or an LCD swap:
- both sides of the main board, in focus
- close-ups where the chip markings are readable
- the silkscreen text near the top-left corner (that's the board revision)
- any unpopulated header pads, especially near the DSPs
Say which model it is in your own words. A TI2's front panel also just reads "VIRUS TI", and the red-versus-blue display tell fails on any unit with a replacement LCD, which is most of the ones that get photographed. A TI board is still useful. I just need to know which one I'm looking at.
Two quick answers
The supersaw change at OS 3.34. I diffed it. 3.3.302 to 3.3.400 is three bytes: the version string and a build counter. The DSP code is identical. So the version number is off somewhere, not the memory. Name a different one and I'll diff that instead.
Unlocking polyphony. There's no host-side voice limit. No constant, no check, nothing in Access's parameter database. It's a DSP CPU budget, so it's probably a real limit rather than an artificial one. I wouldn't get hopes up.
If you want to check my work rather than take my word for it
The parameter names and indices above come out of REMOTEDEVICE.XML, which is sitting inside the
Virus Control installer you already have. Pull it out and read it yourself. The USB findings are
checkable against sound/usb/quirks.c and quirks-table.h in any kernel tree. Neither one needs you
to trust me, or an LLM.
If you think something in here is wrong, say which part. Vague doubt I can't do anything with; a specific claim I can go and check.
And on opening my own unit: I will if it comes to that. It's last on the list, not off it. If a board that's already apart on someone's bench answers the question this week, that beats dismantling a working synth to find out something another person can already see. Patches get tested in an emulator long before they go anywhere near hardware, which is rather the point of having one.