Reply To: Size / Format Discussion

HOME Forums Sierra History Size / Format Discussion Reply To: Size / Format Discussion

#20927
Unknown,Unknown
Participant

(re: re: re: re: re: re: re: re: Kinko’s scanning) For reference of this size/quality issue, I got a copy of the Totally MAD collection, which contains high-res scans of every single issue of the original MAD Magazine between 1952 and 1998 along with a lot of side-material like sound and video clips and various wacky things (if you’ve ever read MAD I think you get the idea). Everything is packaged within a nice and funny user interface. It is produced by Broderbund. Hey, that’s your old buddies Ken! 😉
Anyway, this collection is 7 CD’s big. The total number of scanned magazines on them are 509, plus a dozen or two special editions. If we omit for a moment the fact that there is lots of extra material on the CD’s, this makes an average of approximately 73 issues per CD. A very rough estimate would be that the average MAD issue contains about 40 pages. That’s almost 3000 pages per CD! So, the quality must be pretty bad you say. On the contrary! There is a feature to send a page to the clipboard, and you get a crisp 24-bit picture with the resolution 1110 x 1626 pixels. Many of the pages are in black and white though, as they were printed in the magazine. But even though this makes it possible to reduce their byte size by at least half (I don’t know if that’s how they’re stored since you still get a color image if you copy such a page to the clipboard) there is no doubt that it’s possible to get lots of scans on a CD at high quality. I didn’t manage to figure out exactly what format the scans are in (PDF?), although I know from analysis of them that the images are definately stored in the regular sequential format defined by CCITT for JPEG images, with 8 x 8 DCT-coded blocks. So there’s no secret super-effective image compression at work here or anything. Apart from the scans themselves, the viewing software allows you to search throughout the entire archive with great precision. If you know a word related to a specific article in some MAD Magazine you remember from somewhere, your chances of finding it immediately are great. This is my own experience anyway. The software has lots of nice features to use with the material, such as the option to interact with those famous fold-in back covers. 🙂
For copyright reasons, I cannot post you one of the scans for demonstration, but I did a little experiment (that you can easily repeat yourself), taking that 2.5 mb scan you posted Brad and reduced its size to the Totally MAD dimensions (actually, those scans have a black border where the paper ends around them so I got a slightly bigger page with your picture) and saved it in standard JPEG format, with 15% compression. The result was quite pleasing, with all images and text practically as good as on paper (and almost twice the size on my 19inch screen) with a file size of 326k. With 800 megabytes you would be able to fit 2512 such JPEG pictures. If every magazine had 40 pages (I’m just guessing here) you would fit at least 62 magazines on 800mb at that quality. For the convenience of having everything on one CD I think this quality is perfectly acceptable. I am all for extra care right from the beginning when doing a project like this so you guys could very well keep scans of much higher quality on your own, but these calculations made me confident that 1 CD is enough to get something that people will be happy with.
For whatever this is worth, the Totally MAD archival project was far more ambitious than this consiering the immense amount of material they had to go through, so it’s reasonable to assume that they did some proper research on what quality/size ratio was acceptable, and it translates very well into this project. Seriously, there is nothing to complain about regarding the quality of these MAD magazine scans and if my calculations holds, it would probably be possible to squeeze in even a bit more quality than they did and still make it all fit on one CD.