home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!mips!odin!fido!mopar!hal
- From: hal@mopar.asd.sgi.com (John Hallesy)
- Subject: Re: Need help for VideoLab
- Message-ID: <ntl6a30@fido.asd.sgi.com>
- Sender: news@fido.asd.sgi.com (Usenet News Admin)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <29JUL92.22324396@twncu865.ncu.edu.tw>
- Date: Wed, 29 Jul 1992 18:27:31 GMT
- Lines: 46
-
- In <29JUL92.22324396@twncu865.ncu.edu.tw> guest@twncu865.ncu.edu.tw writes:
-
- >Hi, everyone:
- > Our lab is planning to order a VideoLab for a 320/VGX but we recently
- > heard that product can't support the function of capturing single frame
- > from input video. It is different from data sheet that we got from
- > local distributor and the NEWS RELEASE released by SGI company in which
- > "Video can digitized to the screen at real-time video rates or stored on
- > disk or in memory as SINGLE IMAGES or SETS OF IMAGES" was reported.
- > Could anyone tell me what is the difference or the truth? Compared to
- > VideoFrame and VideoCreator, what's the improved function of VideoLab ?
- > The last question is why the VideoLab don't provide the video format of
- > NTSC or PAL? Is there any method to convert NTSC or PAL to the format
- > provided by VideoLab when we must use it?
-
-
- > Any help will be much appreciated.
-
- > J. R. Liao engineer@amy.cc.nctu.edu.tw
-
- I don't know where you heard that VideoLab can't capture single frames from
- video input, but it most certainly does; we capture an incoming RGB, YUV
- (e.g. Betacam, MII, Smpte), or D1 video signal and place it in a dedicated
- frame buffer. From there it can be transferred to the graphics buffer or to
- host memory. You can transfer frames into the graphics buffer at live video
- rates, and can transfer frames into host memory at less than live video rates.
-
- Compared to VideoFramer, you get realtime video output from a scanned portion
- of the workstation's graphics buffer, in either 'pass-through' mode or in a
- VideoCreator style 'scan conversion' mode. The scan conversion on VideoLab
- is done digitally, so the video output has much higher quality than the
- VideoCreator. Unlike the VideoFramer, VideoLab allows you to read RGB data
- directly from a YUV or D1 input source with no software conversion overhead
- (so it's much faster), and the performance of the graphics bus to the MP bus
- is much better than the VME bus limited VideoFramer. (So, video input to host
- memory is much faster than VideoFramer).
-
- Lastly, NTSC & PAL video outputs were not provided becuase we couldn't fit a
- broadcast quality implementation onto the board; we felt that users who needed
- to go to these formats would keep their video processing in component from
- up until the last edit cycle to tape, at which point you can use an external
- encoder to convert from RGB/YUV to NTSC/PAL. Similarly, you can purchase an
- external decoder to convert from NTSC/PAL to RGB/YUV, at which point it can
- be fed into the VideoLab's video input.
-
- --hal@sgi.com
-