Empirisoft Support

    Welcome to Empirisoft Support
Results 1 to 4 of 4

Thread: Compressed vs. uncompressed image files

  1. #1
    Join Date
    Aug 2008
    Posts
    36

    Compressed vs. uncompressed image files

    You have already discussed in the User's Guide the the advantage of uncompressed over compressed image files, but I would like to know more. Sure, it takes extra processing to decompress compressed files for display. But the smaller compressed files also take less time to read from the disk. In a case where disk I/O is the bottleneck (i.e., disk I/O is slow, processing is fast), won't the reduction in disk I/O more than make up for the increased processing load, making compressed image files like .jpg & .gif faster? Ultimately, real data from actual time tests would settle this, perhaps you have already done this and can report the answer.

    Thanks as always,
    -- David McFarlane, Professional Faultfinder

  2. #2
    Join Date
    Nov 2005
    Posts
    3,328
    Interesting. As you've suggested, it does get to be a complex issue. In accord, we decided recently that the user should have some control over this. In fact, in the new Vista/XP beta version, I believe you can use JPGs now (maybe even GIFs?). If you want to do any performance comparisons, I'd be thrilled to hear that compressed images can outperform non-compressed if the processor to disk-speed ratio is just so. I know many users would be thrilled at the idea of never having to convert another image to a BMP without a threat to speed.

  3. #3
    Join Date
    Aug 2008
    Posts
    36
    I am still thinking about this. As I understand from the manual (most of my knowledge of DirectRT comes from reading the manual rather than actual usage), all the stimuli for each trial are fully cached during the ITI before the trial. I.e., all the image files for the upcoming trial are read from disk during the ITI, and after that DirectRT simply flips screens during the trial. In that case, any delays from disk I/O or image decompression happen during the ITI and do not affect timing within trials. So any added delays from disk I/O or image decompression would only potentially add to the ITI, and if the ITI is already long enough then no one will notice the difference between .bmp, .jpg, .gif, or whatever, and researchers should use whatever format is most convenient for them.

    Also as I recall, if the specified ITI is too short then DirectRT will lengthen it as needed, and will record this time in the output file. That gives us a way to run the test of, say, .bmp vs. .jpg. First make two copies each of several image files in both formats (should really make several versions of each .jpg at different compression rates, etc.). Make a simple DirectRT input file that just shows one image per trial, and set the ITI to 0. I think this works even better if you use stimulus lists, then you can swap .bmps for .jpgs just by changing the name of the stimulus list files. Anyway, run the input file and look at the actual ITIs that result under the different conditions.

    I will not have the time to run this test myself, I just wanted to specify a testing methodology in case anyone wants to try.

    -- David McFarlane, Professional Faultfinder

  4. #4
    Join Date
    Nov 2005
    Posts
    3,328
    Yes on all counts (as far as I can tell), and I agree this would make a nice test. Would love to hear/see the results if someone could try it. Otherwise. I'll try it myself when I get the chance and will report back on the results.

Similar Threads

  1. Name of image appears instead of actual image
    By jarvis24 in forum DirectRT Older Versions: Troubleshooting
    Replies: 1
    Last Post: 10-10-2006, 02:42 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •