In the new version of our apps we have switched from using our own custom search to Apple’s PDF search, which was previously only available in macOS. In our testing we found that Apple’s search is much better at finding results in scanned documents (even than Adobe’s own search). This is because when scanning, the OCR process identifies text where it visually appears on the document, so sentences, phrases, and even words can be broken apart if the text is skewed or misaligned. Apple’s search is able to find phrases in a document even if that phrase is separate blocks of text as recognized by the OCR process.
Unfortunately, since Apple’s search is more powerful it is slower, and consumes more memory. During testing we did find that a minority of documents, due to the way they are scanned, would consume an exceptional amount of memory during the search. If there are many of these documents inside a case, the new search can consume too much memory when searching multiple documents at a time. Unfortunately, searching one document at a time is not practical.
We tried to strike a balance between performance and reliability for our initial release, but we are working hard on optimizing this in a way that will allow our users to keep the accuracy of the new search along with speed and reliability. In the meantime there are a few things you can do that can help prevent the search from crashing the app:
- Organize your case into folders and search the folders to avoid searching more documents than you need to.
- Avoid scanning multiple documents into a single PDF. Scanned documents contain images of each page, and the more pages a document contains, the more memory it will use while it is being searched.
- Scanning at a higher DPI will cause higher memory usage when working with the document. This itself isn’t necessarily an issue, but when combined with #1 it definitely can be. (For example, a letter sized document scanned at 300dpi will have a 2550 x 3300 pixel image for each page.)
- For scanned documents, file size doesn’t matter. With everything else equal, a highly compressed 300dpi scan, once decompressed, will use just as much memory as the same image from a larger file. For scanned documents, pay the most attention to the number of pages per document and the page size x DPI.