nathan@jacobi (Nathan F. Janette) (05/09/91)
I am the unix systems manager for a prof that works with x-ray crystallography.
We use Polygen's Quanta package on a SGI 4D/220GTX machine to display structure
pictures. We would like to record some video presentations of the work we are
doing on the SGI. Our local sales rep has presented the following options.
1. VideoFramer
- high res capture
- 1 frame / 15 seconds
- $7,500 list
2. VideoCreator
- medium res capture ("less than ntsc res")
- 1 frame / 4 seconds
- $12,500 list
3. Genlock with NTSC
- low res capture
- ? frames / second
- ? list price
It seems that it would take an awful long time to make a ten minute high res
video capture with a reasonable number of frames! What gives? The sales rep
said the slow frame capture time has to do with rgb to ntsc conversion.
Are these the only/best options? Is there no real-time option?
Thanks in advance.
--
Nathan Janette
nathan@jacobi.biology.yale.eduwes@uh.msc.umn.edu (Wes Barris) (05/10/91)
In article <1991May9.164242.18284@cs.yale.edu>, nathan@jacobi (Nathan F. Janette) writes: > > 1. VideoFramer > - high res capture > - 1 frame / 15 seconds > - $7,500 list > > stuff deleted... > > It seems that it would take an awful long time to make a ten minute high res > video capture with a reasonable number of frames! What gives? The sales rep > said the slow frame capture time has to do with rgb to ntsc conversion. > We use the Video Framer. You have to supply it an IRIS "RGB" file of the proper size. This size varies depending on your selected mode. For example, I use the D1 output whose corresponding size is 720x486. Sending image data out the D1 (or any other) port on the Video Framer takes about five seconds. I send this image to an Abekas A60. If you are going directly to tape, you will have to do this frame-by-frame (preroll/edit/postroll, etc.) This process generally takes about ten seconds per frame. Making a 1000 frame movie could take around four hours just to record. > Are these the only/best options? Is there no real-time option? > I have heard of some up-coming SGI product that will provide a "real-time" scan converter and multiple-video outputs. However, this wont help if you can't display your images in real time. o o o o o o o . . . ________________________________ _____=======_T___ o _____ ||Wes Barris | | wes@msc.edu | .][__n_n_|DD[ ====_____ |Minnesota Supercomputer Center| |(612) 626-8090 | >(________|__|_[_________]_|University of Minnesota_______|_|_FAX: 626-1596_|_ _/oo OOOOO oo` ooo ooo 'o^o^o o^o^o` 'o^o o^o` -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- "You may get straight A's in school, but you've got a lot to learn about Poker!" -- Geordi LaForge
rapo@cs.cornell.edu (Andy Rapo) (05/10/91)
If you can display the stuff in real time then you can record it in real time. The only reason I can think of to use a frame grabber is to do your recording in the background, or to store images that take longer than 1/30 sec. to draw. We have a 4D220GTX and the VideoCreator. It does a real-time conversion to NTSC/S-VHS which we record directly onto tape. As long as our animations can play in real time, we can record them in real time. So far, we've been very happy with the VC. The videoCreator will also control your VTR decks allowing us to completely automate the recording process. I've just finished an Allegro Lisp interface to the VideoCreator's functions. This makes using the VC very easy. Andrew Rapo Cornell University Simulation/Robotics