Mahler's Annotated Symphony No.1

[horndude77]

It's an ad to get you to buy their recordings, but the score is still pretty interesting: http://nyphil.org/buy/eStore/maazelMahl ... erset_0902

[Yagan Kiely]

That's wonderful!

Anyway to download it? The SWF I downloaded doesn't seem to work (doubted it would though).

[Leonard Vertighel]

Anyway to download it?

Should be possible, but not easy. Unfortunately I don't have the time to work on it at present.

The high-resolution images are split into small tiles in jpeg format which need to be downloaded individually and spliced together again in the correct order. Unfortunately, at first glance I haven't been able to figure out the URL structure. If anybody wants to give it a try, I suggest using HttpFox for Firefox to find the URLs of the individual tiles (be sure to get the highest resolution/largest magnification).

By the way, does anybody know if there is a way to losslessly splice together jpeg images? I know that lossless cropping is possible (with some limitations), but I don't know about the converse.

[varnis]

URL structure is the following, I can automate the download process, but again, i echo leonards comment, does anyone know how to splice jpegs losslessly?

http://barhama.com/NYPHILARCHIVE/MUSICS ... page{{PAGE NUMBER}}/TileGroup0/{{ZOOM LEVEL (5 highest)}}-{{X}}-{{Y}}.jpg

The images are presented as 19x24 zoomify splices, so theoretically, given that each piece is about 6 kb, the total score weighs in at 246mb, so it would need compressing, but this is definitely a worthwhile project.

[sbeckmesser]

Theoretically it is not possible to directly splice JPEGs losslessly, even if they are pieces cut from the same original image. There will always be some splicing artifacts where the borders of the smaller images meet and especially if the image sizes are not integral multiples of coding "blocks". Whether or not the artifacts would be visible depends on the image and how good the original JPEGs are.

The easiest solution in this case would be to turn all the JPEGs into a lossless format (like a bunch of TIFFs), splice those together and save the resulting larger montage also in a lossless format (i.e. without resizing the image or other processing). The image will therefore have only undergone one stage of quality degradation (the original JPEGs). While this is labor intensive, it is doable even with fairly common software packages (Photoshop etc.). Although this might produce a large image file, file size is not nearly the problem nowadays as it used to be, what with today's big disk capacities and broadband internet connections.

--Sixtus

[varnis]

Gorrecting myself, in a irritating coincidence, zoomify has split the image up into different farms, and it does not SEEM to have any rhyme or reason, for example on page 0, TileGroup0 holds the left-most 4 pieces for the first 14 rows, then the leftmost 3 for the rest, and TileGroup1 holds columns 4-16 on the first 14 rows, then 3-15 for the remainder. However on page 5, TileGroup1 switches at row 9.... Im going to write a grabber that downloads all tile groups where the filesize >200 bytes, which should exclude 404 pages....

[Leonard Vertighel]

Im going to write a grabber that downloads all tile groups where the filesize >200 bytes, which should exclude 404 pages....

This is not important, but I'm curious: why not directly evaluate the HTTP response code? After all, they are called 404 pages because, well... they return a 404 code...

[varnis]

This is true, truth be told, i havent spent much time web programming with * Shudders * vb, but your point is a good one