- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The JPEG code in IPP fails on these JPEGS, all valid and loadable by other programs. We have fixed or rewritten the Intel code for the first two, but are stumped by the last one.
ftp://ftp.jriver.com/pub/downloads/music/tmp/ipp_fail_1.jpg
Intel code tries to seek to a negative byte offset -- I can provide the source
code fix for the Intel code if you like
ftp://ftp.jriver.com/pub/downloads/music/tmp/ipp_fail_2.jpg
Intel ippiYCbCrToBGR_JPEG_8u_P3C3R function does not work reliably on boundary
cases and creates an artifact in the lower right box -- we rewrote the color space converter with our own MMX assembly
ftp://ftp.jriver.com/pub/downloads/music/tmp/ipp_fail_3.jpg
This produces a colored box in the lower right corner. I think the DCT value is incorrect, but I can't find the problem. Any help would be greatly appreciated.
All of these fail with JPEGView and with our production code (statically linked, running on Windows).
Thank you for any support.
Link Copied
- « Previous
-
- 1
- 2
- Next »
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thank you very much.
This appears to fix both this issue (crash on invalid JPEG) and the 411 scaling bug here:
/en-us/forums/showthread.php?t=59769
Thanks again!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Douglas,
we are just fine, working hard on the next version of IPP product.
Regards,
Vladimir
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- « Previous
-
- 1
- 2
- Next »