[
https://issues.apache.org/jira/browse/TIKA-4862?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18109897#comment-18109897
]
ASF GitHub Bot commented on TIKA-4862:
--------------------------------------
tballison merged PR #3103:
URL: https://github.com/apache/tika/pull/3103
> Add imageQuality to PDFBoxRenderer
> ----------------------------------
>
> Key: TIKA-4862
> URL: https://issues.apache.org/jira/browse/TIKA-4862
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Minor
>
> Default PDFBoxRenderer rendering is png, 300 dpi, gray scale, max
> compression. The user has no way to alter the compression/image quality.
>
> If we drop compression effort to 0.5 for png, we can get 1.7x faster with
> only slightly larger pngs.
>
> I discovered this on a run of 1k pdfs from digital corpora. This this cut
> total render+write time from ~4,960s to ~2,900s.
>
> We can further improve performance further by without loss to OCR quality by
> dropping to 200 dpi... that'll be a separate ticket.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
