Scanning Slides and Negatives: JPEG vs. TIFF - what do I really give up?

Question

I am about to scan about 1,200 slides and color negatives and am trying to determine whether I should produce JPEGs or TIFFs. I did a test run and compared the quality of 4,000 dpi JPEG vs. TIFF and the image quality appeared similar - even when comparing print enlargements using a lupe. The file sizes differ dramatically -- so I am biased to produce JPEG, all other things equal. I understand that one major difference is that each time a JPEG is saved, some image degradation occurs since the compression is not 'lossless'.
A few questions regarding the drawback of using JPEGs as my 'masters' in Aperture:
- Since I am using Aperture, and a feature of Aperture is always maintaining the 'master' without change, do I not have some protection against image quality loss (i.e., beyond that which occurred during the initial scan)?
- Is there any adjustment that cannot be made within Aperture to a JPEG, that can be accomplished with a TIFF?
- Any other drawbacks that I should consider when using a JPEG as the 'master' for my images within Aperture (vs. TIFFs)?
Note that I am not concerned about drawbacks associated with functionality in other programs (Photoshop, etc.), just drawbacks associated with functionality in Aperture.
Thanks!

Answer 1

Since I am using Aperture, and a feature of Aperture is always maintaining the 'master' without change, do I not have some protection against image quality loss (i.e., beyond that which occurred during the initial scan)? Is there any adjustment that cannot be made within Aperture to a JPEG, that can be accomplished with a TIFF? Any other drawbacks that I should consider when using a JPEG as the 'master' for my images within Aperture (vs. TIFFs)?
Along the lines of Live Picture, which was the first colour correction software for ColorSync, and LinoColor which was the first colour capture software for ColorSync, Aperture constructs ICC colour matching sessions to show the colours different colour devices are capable of forming, and does not change the source pixels in the colour correction process.
Nonetheless, JPEG is a lossy model by definition. It works the way compression works in the CIELa*b colour model or in Kodak's colour model for PhotoCD, that is, by destroying information in the chroma channels and preserving information in the lightness channel. Using JPEG as archival master is not a good idea imho, at least not if changes are to be made to the archival master.
For archival masters, select a colour correction ICC colour space (a 'working space' in Adobe terminology) that is large enough to hold any colours you are capable of capturing, and large enough to hold any colours the innumerable colour devices you will be matching to are capable of forming. You can use the Apple ColorSync Utility to check gamuts.
You cannot reproduce a colour that you do not have in your source colour space, even if the colour device you are matching to is capable of forming that colour. For instance, don't use sRGB that is too small to form the colour of pure cyan on bright art paper in ISO 12647 offset printing. And you may want to use TIFF that can be compressed losslessly. Always embed the ICC profile for the colour space in which you capture/corrected the image. Otherwise, there is no definition of the colours your colourants should form on colour devices that need to have the colourants changed in order to form the colours you intended.
/hh