Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Mar 1995 18:46:22 -0700 (MST)
From:      Jeremy Chatfield <jdc@crab.xinside.com>
To:        jkh@freefall.cdrom.com (Jordan K. Hubbard)
Cc:        hackers@freefall.cdrom.com
Subject:   Re: Video stuff...
Message-ID:  <199503250146.SAA05172@crab.xinside.com>
In-Reply-To: <23629.796085369@freefall.cdrom.com> from "Jordan K. Hubbard" at Mar 24, 95 02:49:29 pm

next in thread | previous in thread | raw e-mail | index | archive | help
Jordan K. Hubbard writes:
> 
> > When Accelerated-X 1.2 is finally released, we'll have support for
> > the Matrox Comet and Matrox Marvel II with the XVideo extension.
> > Other boards will be supported in the course of the year, but we
> > wanted limited objectives (one board in two model types, with the
> > existing X Consortium X Video extension as the API).  Later releases
> > will offer more function and feature enhancements.
> 
> Can you perhaps say a little more about these boards?  I'd like to
> have something good for 3D rendering (ala Matrox Impression) too, if
> possible, and I had thought up to now that I'd have to have 2 cards -
> the Impression and this Diamond board (or something else).  If the
> Comet/Marvel can do BOTH, then my checkbook is ready and waiting -
> money no object! :-)

The Comet is a cut-down Marvel II.  I'll describe the Marvel II and
then describe what's left out of the Comet.

	Video Input via RGB, S-Video or Composite to a 9-pin D type
	connector.

	Philips Video chips & Tseng ET4000/V33

	Display via Tseng ET4000/W32p with STG1702 RAMDAC, 2MB RAM,

	MPEG by C Cubed 

	Audio by TI (TMXE320AV)

The Comet doesn't have the Audio and MPEG sections.  The boards are
identical, differing only in component count.  *LONG* PCI boards may
interfere with fan or cooling fins on the Pentium.

We can present a normal speed ET4000/W32p server (last time I
benchmarked these chips, we hit between 150,000 and 200,000 xStones,
I think), up to  1152x900x16bpp or 800x600x24bpp.

We can select among the input channels (3), scale the image and
permit a freeze and XvGetImage.  There are controls for automatic
gain control enable/disable, gain, clamping, hue, saturation,
brightness, sharpness.  On the Marvel II, we also control the audio
signal, input can be seected from left, right or both channels, with
standard audio controls, e.g volume, etc.

The ET4000/W32p is fine for our software accelerated PEX, but is
obviously going to be much slower than dedicated 3d hardware.
However, we don't have 3d hardware acceleration (yet), so this issue
is irrelevant for a while longer.

> > MPEG hardware initially.  However, you will be able to take a live 
> > video feed from VCRs or cameras using RGB, S-video and composite 
> > inputs, size them using the on-board scaler and display them in an X 
> > Server at up to 800x600x24bpp or 1152x900x16bpp.  The "best" 
> > resolution is probably 1024x768x16bpp with 60Hz refresh.
> 
> That's essentially all I want anyway.  In fact, if I can get this
> combo to work then you guys may very well have a _very_ strategic
> partnership come out of it.  I'm not saying any more, except that I'm
> playing "point man" for a much larger (and richer) concern on this at
> the moment and am interested in any and all solutions.  Maybe you
> and I should have another telephone conversation soon! :)

Call me...  I'm babysitting tonight, so sometime after 8pm Mountain
Time would be best!

> > You can also grab images, using XvGetImage, but you can't feed the
> > video stream to disk (you'd need around 60MB/s sustained, and the
> > best that we can see on a PC is about 20MB/s).  Of course, by the
> 
> That's also an acceptable constraint, at least for now.  We just need
> to figure out a solution that will allow us to:
> 
> 1. Show scalable video in a window.

Can do.

> 2. Animate 3D objects at high speed using Reality Labs(tm) or something
>    similar.

Doug?  What's an ET4000/W32p like for you folks?  Our software
accelerated PEX is really machine limited, rather than display
bandwidth limited, so an Imagine 128 currently runs insignificantly 
faster than the ET4000/W32p.  I'd expect the same of Reality Labs'
product, until 3D h/w acceleration is in place.

> If you've any suggestions, like I said, I'm more than willing to invest
> in a research platform.  I've sort of backed off for the moment out of
> confusion over all the different video boards available! :-)

We picked the Matrox as it was the first board that Thomas Roell
thought was worth the effort.  The other boards have headed for the
anonymity of the back of the closet ;-)  We've not taken a serious
look at the Diamond boards yet, but the 9130 video chip leads us to
suspect that the Matrox still has the edge in display technology.
I suppose that we'll all find out later this year, eh?

There is, as usual, new stuff coming along, about which we are
foresworn to secrecy... but watch for new players and for current
players to have some surprising turns of speed and a new focus on 3D...

Cheers, JeremyC.
-- 
Jeremy Chatfield, +1(303)470-5302, FAX:+1(303)470-5513, email:jdc@xinside.com
        X Inside Inc, P O Box 10774, Golden, CO 80401-0610, USA.
   Commercial X Server - for more information please try these services
http://www.xinside.com            info@xinside.com            ftp.xinside.com



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199503250146.SAA05172>