I've had to do this with directories full of TIFFs a few times over the 
years and finally wrote a script. I found that working with ImageMagick 
alone really taxed the machine (may it's pulling all of the images into 
memory or something?) and so it was more efficient to make a PDF of each 
page and then use pdftk to string the PDFs together at the end.

https://gist.github.com/jpstroop/2956093

I'm not saying I'm proud of this, but it works, and could easily be 
modified to work with JPEGs.

-Jon

On 11/08/2013 02:09 PM, Ethan Gruber wrote:
> I've done something like this in imagemagick, and it worked quite well, so
> I can vouch for this workflow.  But just to clarify, I presume you will be
> creating static PDF files to place in the filesystem--not generate a PDF
> dynamically through Omeka when a user clicks to download a PDF (as in,
> Omeka files off an imagemagick process).
>
> Ethan
> On Nov 8, 2013 2:00 PM, "Kyle Banerjee" <[log in to unmask]> wrote:
>
>> We are in the process of migrating our digital collections from CONTENTdm
>> to Omeka and are trying to figure out what to do about the compound objects
>> -- the vast majority of which are digitized books.
>>
>> The source files are actually hi res tiffs but since ginormous objects
>> broken into hundreds of pieces (each of which can be well over 100MB in
>> size) aren't exactly friendly to use, we'd like to stitch them into
>> individual pdf's that can be viewed more conveniently
>>
>> My game plan is to simply have a script pull the files down as jpegs which
>> can be fed to imagemagick which can theoretically do everything I need.
>> However, I've never actually done anything like this before, so I wanted to
>> see if there's a method that people have used for combining lots of images
>> into pdfs that works particularly well. Thanks,
>>
>> kyle
>>