Closed storytracer closed 4 years ago
makes very much sense
will be available in release 1.4.3
The issue persists. Although there is a dropdown menu available in the GUI now, selecting the "Original" option does not have any effect for me. Example of one document: my binarised input TIFF files have a combined size of 1.42 MB, but the PDF export (it does not matter whether server or client export) is 25.8 MB big.
Problem is that internally the image is written to a ByteArrayOutputStream as 'jpeg'. Solution: we directly access the image (either 'original' or 'viewing')
latest Snapshot contains a solution for this problem.
The PDF export function (client + server alike) produces very large file sizes, since all images are converted into JPEG by default, even if they have been optimized beforehand.
For example, I mainly use binarized, monochrome TIFF files at 600 dpi, which average ~ 40 KB. But the PDF export converts them to JPEGs, which average ~ 550 KB.
For documents with many pages, this results in PDFs with sizes of several hundred MB, instead of a few MB. Furthermore, the quality of the images is reduced due to compression.
Would it be possible to introduce an option "Image type: Original" for the PDF export, as it currently exists in the METS export options?
I think it would not only make the use of PDF for users more practical, but also reduce traffic and CPU load for the Transkribus servers.