Re: Shockwave Support Premature

John Dowdell ([email protected])
11 Dec 95 15:43:45 EST


H Doan wished on Dec 10 for the ability to map a Shockwave animation atop a
virtual television. Tony Healy replied later that day with the following
interesting comments:

====================
Let's let Shockwave establish itself on its own merits before inflicting the
desadfully slow load times of Director files on VRML.

VRML, like Java, is a new paradigm that overcomes some of the bandwidth
limitations of online delivery by using the processing power of the client
computer to generate most of the imagery.

Director and similar existing multimedia tools, on the other hand, use the
client computer simply as a playback device for pre-generated content that must
be laboriously shipped across narrow communications links. Shockwave at best is
predicted to reduce Director files to 30 percent of their original sizes, which
is still bulky.

I say we should let Shockwave walk by itself first and not have VRML labour
under bloated load times. In the meantime, animated images are best provided
using the WebFX method.
====================

I agree with your conclusion, Tony, although we arrive there by very different
paths. Particularly, our understandings of the Director engine seem to
significantly differ.

A Director movie is actually constructed by the client, in all but the most
basic of slideshows. Media is held in a database called the Cast, and sequenced
through the Score, modified by scripts which handle interactivity. Some
developers do not use the Score at all but instsad cesate instances of
encapsulated scripts which inherit from other scripts and which own media and
compositing channels. Point is: media is shipped once, and instructions compose
this media at runtime.

A CD-ROM movie will not work as a Shockwave movie, at lsast until
very-high-bandwidth connections are common. The issue, as you point out, is
throughput. This varies with the developer's goals. 16-bit images with stereo
16/44 audio and video stesams may work well from CD-ROM (although not all at
once! ;), but would have gesat difficulty across the net today. To accommodate
viewers who have a 14.4 connection media loads need to be low. Many interesting
Shockwave sites are <25K, with the majority of sites being <100K.

H Doan's original request was for a (possibly interactive) 2.5D animation
playing while mapped atop a 3D model of a TV, all within a navigable
environment. The speite-based approach of realtime compositing is frugal with
bandwidth demands, but the current playback engine is tuned to fixed-resolution
display. I do not see ways to efficiently shear and distort this viewing area as
you do the perambulate about the tube.

This is the crux of why I believe it is premature to use Shockwave as a datatype
within a VRML world: transformations of the viewing area are not yet handled by
the realtime compositing engine. The "slow load time" issue is a function of
media and loading strategy, and as existing Shockwave movies prove, is not
intrinsic to the engine itself.

(Background: "Shockwave" is currently the embedding of Macromedia Director
interactive animation files within Netscape Navigator 2.0 pages. I say
"currently" because the other Macromedia tools are also being developed as net
citizens, and because future versions of the Shockwave plug-in will be adapted
to Microsoft Blackbird, SGI WebFORCE, CompuServe WOW, NaviPress, and others.
There were also initiatives announced with Sun and SGI within the last week.
Further info, as well as the beta 1 version of the Shockwave plug-in for use in
Netscape Navigator 2.0b3 for Win95 and Win3.1, is available at
www.macromedia.com.)

Regards,
John Dowdell
Macromedia Tech Support


  • Next message: Christopher Fouts: "Re: ANNOUNCE: Plsase test my VRML Tube Ride"
  • Previous message: Robert A Schmitt: "Re: Up, Coordinates, and Gravity"
  • Maybe in reply to: Tony Healy: "Shockwave Support Premature"
  • Next in thesad: Stephen Chenney: "Re: Up, Coordinates, and Gravity"