Corrupted images after reprocess-media

alejoelconejo

Hi everyone,

I'm trying to reduce the size of digital objects on the server. I changed the quality of the original version in the media_processing.conf file.

If I run caUtils reprocess-media -v original everything works fine. But when I run the same process without specifying the version the images get corrupted (original + all derivatives).

Is there a safe way to reprocess all media?

I'm using CA version 1.7.17 with GMagick.

Thanks,
Alejo

alejoelconejo

The image processing rules are these ones:

            # ---------------------------------------------------------
            # Image rules
            # ---------------------------------------------------------
            rule_icon_image = {
                    SCALE = {
                            width = 72, height = 72, mode = fill_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_iconlarge_image = {
                    SCALE = {
                            width = 250, height = 250, mode = fill_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_tiny_image = {
                    SCALE = {
                            width = 72, height = 72, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_thumbnail_image = {
                    SCALE = {
                            width = 120, height = 120, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_widethumbnail_image = {
                    SCALE = {
                            width = 110, height = 75, mode = fill_box, crop_from = center, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_preview170_image = {
                    SCALE = {
                            width = 170, height = 170, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_widepreview_image = {
                    SCALE = {
                            width = 200, height = 120, mode = fill_box, crop_from = center, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_preview_image = {
                    SCALE = {
                            width = 180, height = 180, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_small_image = {
                    SCALE = {
                            width = 240, height = 240, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_medium_image = {
                    SCALE = {
                            width = 400, height = 400, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_mediumlarge_image = {
                    SCALE = {
                            width = 580, height = 450, mode = bounding_box, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_large_image = {
                    SCALE = {
                            width = 700, mode = width, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_page_image = {
                    SCALE = {
                            width = 1000, mode = width, antialiasing = 0.5
                    },
                    SET = {quality = 75, format = image/jpeg}
            },
            rule_to_ctm = {
                    SET = {format = application/ctm}
            },
            rule_to_stl = {
                    SET = {format = application/stl}
            },
            rule_to_pdf = {
                    SET = {format = application/pdf}
            },
            rule_original_image = {
                    SET = {quality = 60, format = image/jpeg}
            },
            rule_tilepic_image = {
                    SET = {quality = 40, tile_mimetype = image/jpeg, format = image/tilepic}
            },

monica

The quality is set lower on original than the derivatives. The derivatives are created from the original.

What type of files are the originals you’re trying to save space with?

alejoelconejo

monica

Thanks for your reply, Monica!

I’ve increased the quality to 80%, but the images still get corrupted when reprocessing all versions. In fact, even with the default media_processing.conf configuration (no transformation rules for the original version), they also get corrupted.

The files are JPEG images.

These are all the images generated for one object representation (using the default media_processing.conf): .

I’m looking to reduce the average size of all the generated images by half. The tilepic version is taking up most of the disk space.

monica

alejoelconejo unless you restored everything from a backup, then you are trying to reprocess the media on corrupted files, so it’s not going to magically repair them and create full quality images again. This is why originals usually aren’t touched. When you upload the original it saves a checksum of it and can compare that to ensure nothing has changed on it, a very basic preservation system.

If you want to save space, then you can adjust the document preview settings in app.conf so it’s not making an image per page. Also comment out the different derivatives you’re not using.

I believe annotations can be used on non tilepics. But tile pics enable high resolution images to be rendered quickly in the browser. If you are not using them, then comment them out. You’ll need to update media_display.conf to not use them anymore and pick either original or large to be the display item.

But do make sure you have backups before doing this, or the only way to get your media back if something goes wrong is to re upload them.

yelto

i would consider consider getting rid of the tilepic version altogether.

alejoelconejo

yelto yeah, that was my first approach. After testing it i figured out that the tilepic viewer is the only one that accepts marks and annotations in particular areas. Am I right?

alejoelconejo

monica yeah, I'm running the tests on a development environment. Will follow your advice and not modify original images.

Is there an easy way where i can see all the derivatives used on Providence default theme? I've started to do this job manually... but it's a difficult job to track all image versions in the whole site.

About disabling tilepic version on media_display.conf, i've already tried it (without success). If I pick original version instead, the TileViewer just shows a fragment of the image and the controls doesn't appear. I changed the viewer to Mirador or UniversalViewer and realize that these viewers are not working on my installation (1.7.17).