Decoder crash: DOREMI 2K4

[Pawel Lipinski]

Hi, and hello everyone!

Long time our server has no error at all with any content we give him to play, but just now after 11 years hi start crash when screening Stops playing and eject content on frame and always on the same one, first was FTR „Limbo” that after 00:05:45 just crash with 3 or 4 times flickering screen, then content was eject form player. When the same content I played but start form 00:05:47 movie played well till to 00:24:XX. We contacted with our service and downloaded new copy of content, this help. So we was think that is problem with the content. But today the same was happened with copy that I was create use DCP-o-matic, so in start to be worry that not only content but also hardware issue. Please if any of you get expirience wiht similar cause can tell what helps and is any way to check for sure what cause this problem. Thanks

[Harold Hallikainen]

Apparently there MAY be some issues with older JPEG decoders. ISDCF is looking into this. It would be interesting to know if others are seeing the same issue, at the same frame, and on what model equipment.

[Carsten Kurz]

Doremi Dolphin decoders are known to have some weakness concerning data rate. It may be that some DCP mastering/encoding software does not consider these specific weaknesses and creates J2K streams that crash the dolphin decoder. In one case I had created a documentary from 4k content with DCP-o-matic at 230MBit/s. It played without issues on various servers, but crashed a Doremi. Only when I recreated it at 210MBit/s, it caused no issues anymore. Frame rate may also play a role here. Something may crash at 25fps but not at 24fps.

[Marco Giustini]

There were Post Production software some time ago with some bugs when it came to maximum bitrate. Setting a max bitrate would work for all but a few frames which would end up larger than the allowed maximum, crashing some decoders.

It could be easily spotted by unwrapping the MXF and check the size of each individual frame. Particularly if encoded at constant bitrate, it's easy to spot one which is different.

[Pawel Lipinski]

Thanks for anweser,

Now i reading Diagnostic Package and found some of errors. I don't think so that could be to much bitrate. But still i not sure that is only content issue.

---

same error repeated 22 times


Line 7822:Oct 20 22:23:59 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 7888:Oct 20 22:23:59 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 7954:Oct 20 22:24:00 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8020:Oct 20 22:24:00 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8086:Oct 20 22:24:00 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8152:Oct 20 22:24:01 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8218:Oct 20 22:24:01 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8284:Oct 20 22:24:02 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8350:Oct 20 22:24:02 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

Line 8416:Oct 20 22:24:02 2023 dcp2000 kernel: Dolphin DEBUG: !!! CRITICAL !!! Decoders crashed at frame 24891

[Marek Dudzik]

Pawel, the same thing happened in at least two other cinemas in Warsaw with the same package (also on Showvault, not only 2K4). The funny part is I've done all tests I know, including easyDCP+ QC on this very DCP and it has found nothing wrong - the bitrate there is ca 180mbps (IIRC) and there isn't any steep jump in bitrate near the timecode you mention. The distributor and post house are both aware of the problem but at the moment they, to my best knowledge, have not identified the reason.

What specs you've used for the second package you mentioned is crashing?

[Leo Enticknap]

When the same content I played but start form 00:05:47 movie played well till to 00:24:XX. We contacted with our service and downloaded new copy of content, this help. So we was think that is problem with the content. But today the same was happened with copy that I was create use DCP-o-matic, so in start to be worry that not only content but also hardware issue...

The fact that the crashes are consistent and at the same point (as your log extract shows) is very useful information for troubleshooting, as it tells us that we can reproduce the fault consistently. This says to me a software (or firmware)/hardware bad interaction. It could well be the Dolphin bitrate issue that Carsten notes. But because deleting and reingesting changed the system's behavior, I'm also wondering about the system's response to attempting to read from a bad sector on a RAID drive. If deleting and then reingesting the same content changes the system's behavior, this would suggest to me a RAID issue.

Would it be possible for Pawel to download a "detailed report" log package, upload it to Dolby Log Analyzer, and then post the link to the log analysis here?