Tight VNC isseus

Frank Evan Perdicaro frank "at" dsea.com
Fri, 03 Aug 2001 18:40:47 +0000

In addition to the other comments that have been posted, let
me note a few things.

I have been using Tight over a 28.8K dial-up PPP connection to a W98
box, pretty much since it first appeared.

The benefits of Tight VNC are many, but only apparent when 
   1) the information density of the screen is low
   2) the link is slow
   3) both ends have fast CPUs

In my application, Tight gives a real-world speedup of about 5x.  I
have a fast P3 at one end and a fast SPARC at the other.  Relative to
the speed of the communication line, the processor speed is infinite.
The application in use works in 8 bit color, changing large sections of
the screen.  I have altered the PC screen resolution to work well with
the given communication channel speed.

Based on some research, I know it is possible to provide a substantial
apparent speedup of VNC past what Tight does, but the required changes
to the code are non-trivial.  Check my "fade to color" note from about
6 months ago, any good book on information theory, and a good book on
human vision.

Tight is quite effective.  Until VNC is altered to implement the
human-vision-specific tricks I outlined, I do not expect to see any
better solution for low-speed links.

> I've compiled both versions on Solaris 7 Intel
> successfully.  The box is a P3/600mhz w/ 512mb RAM. 
> On a 10base-T network, I noticed that if I use the
> "tight" encoding, the screen display is slower than if
> I just use "hextile".  I'm assuming this is because it
> takes more time for the Unix box to compress the data
> stream it sends to the client and it takes additional
> time for the client to decompress the data stream. 
> The client is a Windows 98 laptop, P2/300mhz w/ 96mb
> RAM.  So the bottleneck is the processing power of
> both machines, not the network.  Is this assumption
> accurate?  Has anyone else had similar experience?  I
> have not tested TightVNC vs regular VNC on a dial-up connection.
To unsubscribe, send a message with the line: unsubscribe vnc-list
to majordomo "at" uk.research.att.com
See also: http://www.uk.research.att.com/vnc/intouch.html