OT: Best Option for Web Delivered Voice Overs

[Sivakatirswami]

Looking forward to future slideshows either distributed as revlets or as 
javascripted/jQuery enhanced web slideshows which are produced by a 
RunRev desktop client toolbox (Almost everything I do on our web sites 
is facilitated with some kind of in-house production widgets... as I am 
sure is the case with most old-time RunRevolutionaries)

I am looking any input from experienced sound producers in our field for 
the following. Some of this may be fairly obvious, but everytime I pose 
a query like this to this list I always get some nugget of useful info 
that I did not know or think of before:

1) MP3 Settings: for voice over narration: what would be the best MP3 
(lowest you think is reasonable) settings for files that woudl be played 
along with a slide.

2) Optimum format for Voice Over: Is MP3 the best format for voice 
over?  Will it play "out of the box" in a hidden QT player object on 
both Mac and Windows?

3) Embedded Clips: Has any one embedded sound clips into a revlet? AIFF 
and WAV seem horribly "expensive" in terms of bytes added to the stack 
vs streaming short mp3's from the web server. But perhas AU is small and 
OK for voice... comments?

4) Text to Speech: How is text to speech working these days on Windows? 
is this a viable option for revlets... some kinds of "kids stuff" will 
do just fine with a computer voice reading the English.

5) Production Work Flow Strategies: Any other insights will be 
appreciated. you can get a Samson USB mic these days really cheap on 
Amazon.dot and the quality is excellent. It would be marvelous to build 
a stack, show slides, record on the fly, save and deploy, but I don't 
think it's going to be that easy... Especially if MP3 is the optimum 
format, it means stepping out of revolution into Audacity.. Processing 
clips and going back to set up the QT player to the external file.  Or 
assuming you were really, really good with narration and required no 
edits:  Switch to batch process all the narrations at once after saving 
as aiff --> process out to .mp3 ... 15 short VO's at average 2-8 secs 
each will be fast.

Goal: production-work flow that gets from 15 photos with voice capture 
to a slideshow with VO on the web in under 45min. (assumes of course 
your narration is either ad lib or pre-scripted...)

Thanks for your patience with my "mine the list" queries

Sivakatirswami