Compression question / problem

[Alex Tweedly]

Richard Miller wrote:

> Looking for the simplest and quickest solution to a compression 
> problem. The goal is to take 50 jpeg images (highly compressed) 
> starting on a Mac, move them to our directory on a hosted web server, 
> then have them downloaded and displayed on a Windows XP machine. 
> Transmission time (over the Internet) is a very big factor here.

I already replied with a "detailed" suggestion - but I also want to give 
a personal opinion on the overall problem.

using Rev's compress (or any other standard lossless compression 
available today) is unlikely to give you very much benefit on JPG files 
- it's rare to see as much as 4 or 5 % decrease in file size, and you're 
unlikely to get better than 3% average on a large number of files.

I'd focus my time and energy on making good use of either the "load URL" 
/ cachedURLs commands or perhaps libURL. Getting a suitable number of 
transfers happening in parallel is your best hope of achieving results 
you like. Exactly what "good use"  means will vary depending on the 
context (number of files, what's being done with them at the receiving 
end, etc.). If they are photos being displayed in a gallery / slide 
show, then you probably want to start with just one transfer (to get the 
first one to display asap), but then build up to 4 or 6 in parallel; in 
other cases you should simply start up N immediately - where N would 
vary depending on photo sizes / connection speeds / wind direction (?) / 
etc.. 
I'm sure that various people on this list would be willing to suggest 
things if you can describe that part of the problem.

If "simple" is important as well as "best" result - set off all the  
"load URL"s in parallel and use the resulting URLs as they become 
available; let Rev's library deal with how many it emits in parallel. 
[This depends on there being only 50 files to retrieve - if you have 
hundreds, you perhaps shouldn't do this, certainly not without testing 
it .... - but I have done upwards of 50 without problems; I didn't test 
to see how many actually happened in parallel].

-- Alex.


-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.300 / Virus Database: 265.6.10 - Release Date: 10/01/2005