Breaking Eggs And Making Omelettes

Topics On Multimedia Technology and Reverse Engineering


Samples RSS And Flashback Samples

December 21st, 2011 by Multimedia Mike

I made good on my claim that I would create an RSS feed for the samples repository.

Here is the link to the samples RSS feed [ ]. Also, here is the Python source code I threw together for the task.

I just want to check: I’m not the only person who still relies on RSS these days, right? The tech press has been cheerfully proclaiming its demise for some time now. But then, they have been proclaiming the same for Adobe Flash as well.

I’m no expert in RSS. If you have any suggestions for how to improve the features presented in the feed, please let me know. And, of course, keep the samples coming. This script should help provide more visibility for a broader audience.

Mario and Flashback Samples
Thanks to LuigiBlood who sent in some samples that allowed me to test out my new script for automatically syncing the repositories and updating the samples RSS feed. First, there are CPC multimedia files from the Japanese 3DO port of Flashback: The Quest for Identity. Then, there is an Interplay MVE file on the CD version of Mario Teaches Typing in which the video doesn’t decode correctly.

LuigiBlood also sent in another file from the latter game. It’s big and has the extension .AV. It could be a multimedia file as it appears to have a palette and PCM audio inside. But there’s no header and I’m a bit unsure about how to catalog it.

Posted in Game Hacking, Python | 14 Comments »

14 Responses

  1. aviadr1 Says:

    after the libav fork, there was a prolonged period that uploading new samples for the purpose of bug reports wasn’t really possible.
    do we now have a public place for submitting samples?
    what is the recommendation for uploading bug report samples?

  2. Multimedia Mike Says:

    Before went online, they managed to put about a gigabyte more of samples on it. Many of them were bug issue-related samples.

    For uploading bug reports, we now have

  3. Michael Mol Says:

    I still heavily depend on RSS…though Google may be bordering on killing it for a lot of people as they neuter Google Reader and try to shift it into Google+.

  4. Mans Says:

    I wouldn’t mind if RSS rotted in the deepest level of hell, as long as I could have Usenet back.

  5. Jim Leonard Says:

    If RSS is supposed to be going away, what is the proposed replacement? What is the “modern” workflow for retrieving regular news updates?

  6. clone2727 Says:

    The MVE files from Star Trek: Judgment Rites don’t play correctly either — I always just thought support in FFmpeg/libav was just “unfinished”.

    Do you want samples from that too?

  7. Multimedia Mike Says:

    @clone2727: If they don’t already exist at then go ahead and upload them. Use the new anonFTP server.

    @Jim: I think the argument is that the subscription model would just go away as software stops supporting it. I remember this article from earlier this year bemoaning this idea. Chrome doesn’t show RSS feeds; Mozilla announced they would remove the RSS URL bar button. I know that in order to “discover” an RSS feed while browsing in Chrome, I always have to view source and search for “rss”.

  8. LuigiBlood Says:

    I hope at some point i can convert those just fine. ^^

  9. Peter Says:

    @Jim: twitter(tm)

  10. Reimar Says:

    Well, the flashback samples seem almost certain to be uncompressed (gzip reduces size by 1/3, there are patterns in the data that seem to indicate quite clearly not even basic RLE is used – at least for the first frame).

  11. Multimedia Mike Says:

    @Reimar: I forgot to mention that the Flashback samples seem to have some relation to the FILM/CPK format which means they probably have Cinepak. I probably should have stored them in the same directory as the rest of the samples.

  12. Anonymous Says:

    @Mike: yes, after some closer look that seems likely. It seems like an I-/P-frame pattern, with the I frames containing more additional data than just a palette. However the pattern of the final image is clearly visible still for the I-frames. This seems close to cinepak.
    The dimensions do not seem to be encoded anywhere in the file.
    It looks like width 132 (though 4 of those pixels seem to contain only some border value), height something around 64.

  13. Paul Says:

    Interplayvideo decoder and ipmove demuxer clearly have few opcodes which are not implemented at all, they need first to be RE-ed properly.

  14. clone2727 Says:

    @Multimedia Mike: Uploaded the sample to the incoming folder — the “stjr_interplay_mve_bad”. Accidentally sent just the CAPT.MVE file the first time, sorry!