Scanned JPEGs are large and slow to load - can they be optimized losslessly?

Question

I have hundreds of JPEG photographs which were scanned about 5 years ago from negative using a Konica Minolta DiMAGE Scan Dual IV. The dimensions are ~4500x3000, and the filesize is around 12Mb, compared to shots from a DSLR with dimensions of 3000x2300 and filesize of 2-4Mb (actually, these are the output from a RAW convertor). The filesize is obviously quite a big difference, but the issue that's bothering me is that the (perceived) loading time is at least 10 times slower.

Is this size/speed discrepancy likely to be because the scanner software saved the JPEGs inefficiently / using an old compression format, or is it simply that the scanned negatives contain much more "detail" (in the form of grain/noise) than the digital images? If the former, is there a way to losslessly optimize them? I've tried re-exporting the scanned files to full size JPEG from my RAW software but the filesize is pretty much the same.

Both files will have been saved at 100 quality.

Answer 1

4500 x 3000 would be a 3000 dpi scan for 35 mm film, which is a pretty good scanning resolution which will retain all the film grain. I get similar file sizes when I do this and save at the maximum JPEG quality setting.

If you want to optimise them for speed of display, then reducing the image dimensions will make much more difference than compressing them more (which cannot be done losslessly with JPEG).

If I were you I'd keep these original scans filed away and run copies through a batch resizer to get them down to (say) 2250 x 1500 which will display quicker in slideshows etc.

Answer 2

Unfortunately there is not conversion which will preserve all of the information. There will indeed be loss if you wish to compress the image. From a human perspective however, you can limit the depth of the color to gain some modest savings in file size. This is also loss, however we are quite well suited to deal with the reduced depth in color.

Scanners typically scan at high resolution as you want as much detail as possible if you are going to be working with the scanner output. A poorly scanned item can ruin an otherwise stellar effort. Most scanners also have a means to dial back the scanning resolution to minimize this sort of huge image output.

Answer 3

What software are you using to view the images? If you just care about viewing them on a monitor, using a program that creates a cached preview of the jpg is a good way to go.

I can recommend Picasa (free) and Adobe Lightroom (not free) for this purpose, as I've used them both a lot.

Answer 4

JPEG is lossy by default, but JPEG2000 has a lossless compression scheme too. The thing here is that you scans are probably at JPEG's maximum quality. 100% or level 12 in Photoshop.

The modern camera throws much higher compression on the JPEG's with more loss, maybe 80% quality or level 9-10 in Photoshop.

So if you want to save room on your old slides, open them, and save them again in a lower quality, like 80% or 10 in Photoshop. Programs like Photoshop, ACDsee or Graphicconverter can do this in batch.