- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I have one question regarding decoding MJPEG. I use IJL library to decode JPEG images and it works fine for me. My intention is to decode a continuous flow of JPEG images (IP cameras mostly produce this kind of stream). Using the typical combinaton of READPARAMS and READWHOLEIMAGE for every image does not seem to be the most effective way of solving this problem. So I guessI could increase performance by saving persistent jpeg tables. I suppose that MJPEG or multipart JPEG streams make use of persistent jpeg tables (is it?). I had a look at the PDF dealing with FlashPix files, but the sample does not work for me (I still cannot get rid of some errors when compiling). But even if it worked some questions would remain open for me. Every JPEG(extracted out of the stream) contains the jpeg tables, so how can I enforce the usage of persistent tables when calling ijlRead? Anyone having an idea about the MJPEG? Thanks in advance.
Alex
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
In the next version of IPP, which is coming soon,we will provide Motion JPEG decoding sample, which work with some of MJPEG AVI files. You should know, there is no international standard for Motion JPEG, so some implementation can use only one set of JPEG tables for the every frame whereas other don't do it.
Regarding IJL, you can copy JPEG tables to some persistent storage and copy them back to JPEG_CORE_PROPERTIES structure before decoding each frame if you know for sure that they will work.
Regards,
Vladimir
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page