Everyone, some interesting source code has come to my attention, where it is reported the existence of a supposed "ST-VE" machine, and according to some details, it seems to be an ST-V with an add-on "graphics enhancer" board, which adds to the B bus both a second VDP1 and two additional DSPs. Ever heard of this before? I haven't... Jollyroger
Quite possible, the STV had several addon boards, none of which are well known or documented. Only two I've seen pictures of are this: http://www.system16.com/boards/stvfishing.jpg Which is apparently a CDROM controller, and this: http://www.system16.com/boards/stv_video.jpg Which I've seen confusing and conflicting info on. Entirely plausible there are other expansion boards out there.
I've never heard of this board, although there were lots of rumours about the Saturn getting some kind of add-on for graphics. Do you have any info on what kind of DSPs would it add? That one is simple, it's a stock Saturn SH1 sub-board (all launch models had the SH1 on a sub board, so did all Sophias), and it has a stock Video CD card connected to it. I've never seen what it connects to, presumably a standard Saturn drive, but I've heard somewhere that it uses a SCSI one...? (which doesn't sound right)
This one looks pretty interesting. Stickered Altera Max7000 FPGAs(or CPLDs actually) have presumably something to do with graphics generating/processing. 27 MHz quarts is for PAL/NTSC encoding/decoding. Zoran chip is either video compressor (zoran did jpeg chips, but not back in 1996 afaik) or some simpler video processor. Video encoder is there to process ITU-R BT656 or whatever from zoran/CPLD into NTSC or PAL and onto RCA connectors. But what bugs me the most is video decoder. It converts PAL/NTSC into digital video stream (that is futher processed/compressed by zoran and moved to CPLDs/bus). That means there was camera (or other video source) connected to input RCA. VDP2 indeed have a digital video input through cartridge slot on stock saturns, but alas, no cart on saturn uses it. And it seems to me that it might be used on this board. And yeah, engrish sticker is cute ^_^
So, a camera input digitizer then? Perhaps used by this: http://www.system16.com/hardware.php?id=711&page=5#1889
The Zoran chip is a color space converter - based on the board layout, I suspect it's taking the YUV output from the Brooktree decoder and converting it to RGB. I have a vague memory that the ST-V was used in some sticker machine - I wonder if it was for that? Edit: Druid beat me to it
Yes, and also an RCA video output. I wonder why they needed an additional video out. Maybe for some preview screen? Probably. I have also found some "Movie Club" in that listing. What's this? Does it somehow capture movies? Or does it just sports some movie stars to make a photo with?
Unfortunately not, there doesn't seem to be directly accessible information about the DSPs, having said that the sources contain the binary (converted to a C data structure) for the two DSPs, so it should be possible to analyze/decompile it using various DSP manufacturers ISA and see which one makes sense... Anyway, the interesting detail is that in the code this is an expansion board, but the whole machine name is "ST-VE", which sounds like this would have been a system upgrade rather than an extension... Just thought it was interesting.
I'm pretty sure it's wrong - the protocol used on that ribbon cable is nothing like SCSI, and it would take a lot of hardware to convert it - to the point where it would make more sense to replace the controller with one that spoke SCSI natively than it would to convert it.
Do you have any accurate timestamps on the files, or a rough idea of when they originate from? It'd be interesting to know what timeframe this talk of an "ST-VE" system was taking place in. Also, any chance you could share the file, or at least the DSP code from the file? As you said, it should be possible to match up the architecture. Sega used a number of different DSP's in that mid 90's period (5 different DSP's from 5 different manufacturers come to mind off the top of my head!), and I'd be interested in where this project fit in with timeframe and architecture. The 90's were a very interesting time for DSP's.
Out of interest do you know who made the DSP inside the SCU? The SEGA documentation is apparently wrong/incomplete on that. I've went over the dev kit tools for it, but I haven't been able to find any reference to any other company, or it being called anything other than SCU DSP.
No idea, sorry. Around this mid 90's period embedded DSP's got quite popular, and it was common for these embedded DSP's to have a "modular" kind of design, where the capabilities and instruction set were able to be easily modified by the company licensing it. I doubt that Sega built this DSP from scratch internally, but with so little information on the chip provided in the docs, and with the possibility that it's customized to meet the needs of the Saturn itself, it could be hard to match up to anything from the info we have on hand. Decapping the chip may reveal something useful, otherwise it's probably a matter of trawling through journal and trade publications of the 90's looking for a hit. A lot of DSP designs were short lived and had small uptake though, and there's a good chance the DSP they used was developed in Japan, with all the documentation in Japanese only too.
The files are dated August 1995, but there is no trace of a timestamp inside the files themselves. I can share the DSP binary code with you, I would love to have a conversation about the different DSPs Sega has used... Jolly