Path: utzoo!utgpu!water!watmath!onfcanim!dave
From: dave@onfcanim.UUCP (Dave Martindale)
Newsgroups: comp.graphics
Subject: Re: High-quality slides from graphics terminals
Message-ID: <15530@onfcanim.UUCP>
Date: 10 Jan 88 18:00:02 GMT
References: <3286@soma.bcm.tmc.edu>
Reply-To: dave@onfcanim.UUCP (Dave Martindale)
Organization: National Film Board / Office national du film, Montreal
Lines: 67

In article <3286@soma.bcm.tmc.edu> rick@soma.bcm.tmc.edu (Rick Gray) writes:
>
>2) How do other folks make high-quality color slides or illustrations of 
>   data collected with their Masscomps (or any other graphics workstation)
>   for presentation?

A very flexible method, but one that requires writing some software, is this:

Display whatever image you want on your workstation.  Then run a program that
grabs the contents of the screen frame buffer and writes it into a file.
You now have an image file that is the same resolution as your screen,
but in digital form, so you no longer care about the video sweep rates
used by your workstation.

Now, pick an image recording device.  If you have a digital film recorder
(i.e. one that is fed with a stream of numbers, not a video signal) which
directly supports the resolution of the file you have, that's all you need.

If it supports a different resolution (some digital film recorders have
fixed 2048 or 4096 X-resolution) then you need to resize your image to fit
the recorder.  For best results, you'll probably want to use linear or
cubic spline interpolation, though if the recorder's resolution is an integral
multiple of the workstation's simple pixel and scanline replication will
also work.

If you don't have a digital film recorder, but do have a video-fed film
recorder attached to some other workstation or frame buffer (you mentioned
a Polaroid Freeze-Frame), then resize the image to fit this workstation
or frame buffer and record it from there.  Frame buffers supporting NTSC
scan rates are typically 512 or 640 pixels in X by 480 or 485 in Y;
you may have to deal with pixels that are not square during the resizing.
This also allows recording directly to single-frame video equipment if
you have it.

The above discussion assumes that the original image was RGB.  If, instead,
it made use of the colour lookup table to set colours, you'll have to
deal with that.  Sometimes you can just pass the lookup table to the
output device and have everything work, but generally you'll have to
convert to RGB (particularly if you use interpolation when resizing,
since that will produce colours that weren't in the original image).


Although this method is slower than just buying the right video-fed film
recorder for your workstation, it will work with any film or video output
device and any number of different workstations, provided that:

1) You can grab the image from the workstation screen in raster form

2) You can record an arbitrary raster image on your recording device provided
   it is the correct resolution


During the past few weeks, we transferred 60 seconds of animation (at 24 fps)
from an IRIS workstation to video via the following process:

1) display each frame on the IRIS, grab screen contents and write to file

2) record each frame on film using a Celco digital film recorder
   (the Celco was set up for 1024 x 768 so no resizing was necessary)

3) process the film and get a print

4) transfer the film to video.

All of the repetitive processes were controlled by scripts or file lists
of some sort, so most of this proceeded unattended.  (You can't expect
a human operator to do anything correctly 1500 times in a row anyway.)