[sdiy] OK, one of you computer geeks, explain this..

[Matt Simpson]

It has a whoooole lot to do w/file compression & graphic rendering.
I *believe* PDF format utilizes some form of vector rendering, which
is similar to what Flash uses (and maintains small file sizes). 
Bitmaps are raw pixel data - each pixel is represented by a unique
value, and that value is the same bit-size each time.  PDFs take
advantage of tons of facts about the document in question (for
instance - are there wide swaths of a certain color?  if so that
significantly decreases file size - same thing MPEG compression does)
and adds its own algorithmic flair to compress the data


On Sun, 30 Jan 2005 18:36:53 -0500, J. Larry Hendry <jlarryh at iquest.net> wrote:
> While this pertains to scanning and file structure, it is not off topic
> because I am scanning synth schematics.  Due to the black and white scanning
> I have done in the past, I pretty much understand how to minimize PDF file
> size for a given resolution (through the BMP file).  However, I am now
> working on documents that require color and grayscale. While some of the
> document is color, about 2/3 is grayscale.  So, I thought I would save file
> size by scanning those in gray instead of color.
> 
> Here is what puzzles me. Let me throw some file sizes out there for one of
> my large 11x17 pages to see if any of this makes sense.
> 
> This page was scanned (paste a few together) in grayscale as a bitmap and
> results in 13.2 MB file size.  I normally scan everything to Bitmap and then
> convert to PDF as that has always given me the most efficient file size
> (past B&W experience).  So, I convert the file to a PDF and the resulting
> file is about 2.6 MB.  Well, I was hoping for better, but I thought that was
> it.  Then, I started noticing some of the color pages were smaller once
> converted to PDF.  I wondered why since the color BMP files were so much
> larger.  So, as an experiment, I took my 13.2 MB grayscale page and
> converted it to full 16 million colors file in my photo editor.  As
> expected, the file size jumper up to 39.7 MB.  I'm still understanding OK up
> to here.  But, now I take that color file and convert it to PDF.  The
> resulting file is only 1.15 MB.
> 
> So in summary,
> gray file started 13.2 MB bitmap
> converted to PDF @ 2.6 MB
> 
> same gray file to full color = 39.7 MB
> converted that color file to PDF @ 1.15 MB
> 
> Of course, since the original source file was grayscale, even the color PDF
> looks grayscale.  In fact I cannot tell the two images apart.  However, the
> one that was converted to the very large color file first ended up the MUCH
> smaller file size when converted to PDF.
> 
> If any of you who are considerable more geeked than I concerning computers
> can 'splain this to me and Lucy, that would be great.  Maybe I will learn
> something that will help me additionally control the file size on these
> large, highly detailed scans.
> 
> Larry Hendry
> computer geek NOT.
> :-)
> 
> 


-- 
~mt
sonic alchemist
eat ! or ! be ! eaten