he photos below show how different
scanners perform when displaying scenes from Lightspeed Design Group's
award-winning show "Creation". The scenes were selected from among those that
look good on 30K-tuned scanners, but look worse on 50K-tuned scanners. To see enlarged views of
any photo button, just click on the button.
A separate page has photos
showing how these scanners perform when displaying test patterns. Click
here to see the test pattern photos.
||hese photos show rounding where
relatively large steps are occurring. The sample scene is an arena with
doors. On GS G-120 with TurboTrack 2, and CTI 6800 scanners tuned to ILDA 30K, the lower
left corners of the doors meet correctly, at a sharp
However, on the Catweazle at
ILDA 30K and the two faster-tuned scanners, the lower left corners are
rounded. You can see this better by clicking on the small buttons which
show close-ups of the doors.
This frame is one of many
in the Creation show where rounding occurs at 50K but not at 30K. To solve this, the
show would have to be played at a smaller scan angle, or would have to be
output at a slower point rate (fewer points per second).
limiting on abstracts
Notice that the TrueK 50 power limiting, in
both the abstract and raster photos, is symmetrical. In contrast, the
Cambridge 6210's power limiting is asymmetrical.
This is because the TrueK 50 approach is to power
limit the entire axis if part of the scan causes overheating. The
Cambridge 6210 approach is to power limit only the part of the scan which
Each approach has its merits, depending on the
particular image being displayed.
||ith the rounding problem,
audience members may not notice -- especially if they have not seen the
normal, non-rounded version of a show. But audience members will
definitely notice the next problem.
This is "power
limiting". It happens when showing an image with so many large jumps
that the scanner may overheat. The scanner amp limits the signal to
Power limiting is already
present on the Cambridge 6800 tuned to ILDA 30K. But it happens
significantly more often with the two faster-tuned scanners. Both of them
shrink this particular abstract, when it shows fine on ILDA 30K-tuned
6800s. In fact, they go beyond shrinking, because the resulting image (in
this case) looks like a scribble.
Such distortion is
unacceptable. The only solution is to reduce the size of either the
abstract within the show or the entire show.
limiting on raster images
the raster images
The tests were done with colors turned off
(all-green or all-red images). This is why the rasters shown here don't
have pictures, just light and dark areas.
Raster images are normally square. The ones
shown here are from a scene where a square raster "blows up"
like a balloon. This is why the rasters have curved lines instead of
||nother unacceptable distortion
can occur with raster images. These images require the scanner to slowly
scan a line from left to right, then to snap back to the left as fast as
The raster below is large
enough that power limiting occurs on the two faster-tuned scanners. Again,
the solution is to reduce the size of either the raster within the show or
the entire show.
This page last updated:
Tuesday, June 19, 2007 08:54:19 PM