This documentation will be written very shortly. ETA by July 30th, 2010.
The following document are notes on the functionality and behaviour of Decapod 0.4. While there may be some useful information here, we recommend reading the Decapod 0.4 User Guide to start.
- Decapod 0.4 now uses gphoto v2.4.9 and libgphoto2 v2.4.9.
- Previously Decapod 0.3 used gphoto2 v2.4.7 and libgphoto2 v2.4.8.
- The Canon PowerShot G10 is the only camera supported in this release because gphoto2 can uniquely identify these cameras by serial numbers. Unique identification of cameras is required to preserve the ordering of left and right pages in the event cameras are disconnected and reconnected during a session.
- Previously Nikon D50, and Nikon D200 have been dropped in this release, and may be added back in the future.
- Remove memory cards from cameras before using Decapod for best results. Otherwise Ubuntu may try to mount the memory card file system and lock out Decapod.
- Set cameras to capture in JPG. Other capture formats not supported or tested.
- Recommend using identical pair of cameras with identical settings for capturing.
Camera Setup Interface
New in this release is the Camera Setup interface. This UI detects attached cameras, reports any problems or incompatibility, and then allows the user to test capture using the cameras.
If the cameras work,
- On first capture, camera detection is performed, there is a slight delay while cameras are detected. User is notified that detection is in process.
- If only one camera is attached, a user action dialog appears alerting the user to attach a pair of supported cameras.
- If only one camera is attached, the dialog does not list the camera currently attached. It should.
- If a pair of cameras are attached, and neither are supported, a dialog should appear to prompt the user to attach supported cameras.
- If a pair of cameras are attached, and only one is supported, both cameras are listed and indicated whether or not it is supported.
- If both cameras are supported, the user action dialog does not appear.
- Pressing the Try Again button does not appear to do anything, even though the user may have attempted to fix their camera setup. They must close the dialog, and press Take Picture again before detection is executed again. (http://issues.fluidproject.org/browse/DECA-34)
- Picture taking occurs one after the other. This should occur in parallel so both cameras take images at roughly the same time. (http://issues.fluidproject.org/browse/DECA-53)
Post Capture Processing
- Stitching and thumbnail processing executed immediately after capture. This creates a delay between capture time and the time it appears on the UI.
- Stitching should be run parallel and UI updated as stitching is done.(http://issues.fluidproject.org/browse/DECA-25)
- Currently stitching assumes a Book Liberator configuration, and rotates left and right images accordingly. (http://issues.fluidproject.org/browse/DECA-50)
- This means that if images are captured in a different orientation, then the stitched image will be rotated incorrectly.
- If left and right pages are out of order, then the rotation may end up wrong resulting in upside down pages.
- If source images are different sizes, then stitching will top-align the images when stitching. This results in white space around the smaller image.
- Stitching process will create a greyscale stitched image, but leaving the original files unmodified.
- Thumbnail generation is based on the greyscale stitched image.
- Export results in a greyscale image PDF with no OCR text or tokenization
- Exported PDF is created in the following path: ./decapod/components/server/pdf/DecapodExport.pdf
- Export process also generates a colour multipage TIFF located in ./decapod/components/server/pdf/multi-page.tiff
- Export to PDF with tokenization and OCR is not in this release.
- Export process can take a long time, and no feedback is given to the user that exporting is working or when it is complete. (http://issues.fluidproject.org/browse/DECA-57)
- The PDF generation will omit pages if it does not detect any characters. This may result in multiple pages captured by the user, but a subset of those appearing in the PDF. (http://issues.fluidproject.org/browse/DECA-58)
- Note: some pages may be included in PDF even if it does not have valid characters.