Open genfhk opened 4 years ago
Testing attachment was a proof of concept about what we can do, but it seems it adds a bug here.
I tried to reproduce the issue, but I can't.
My test policy:
<?xml version="1.0"?>
<policy type="and" name="MyPolicy">
<rule name="FFV1 uses slice crcs?" value="extra/ErrorDetectionType" tracktype="Video" occurrence="*" operator="=">Per slice</rule>
</policy>
CLI:
mediaconch 6587862.mkv -p policy.xml
pass! 6587862.mkv
GUI:
(so "N/A" for the attachment, no fail)
Which version of MediaConch do you use?
In any case, we should at least have an option for disabling tests on attachments.
I just tested with another policy and got the same error. Can you send me your policy file so I can check? I'm using version 18.03.2. But, yeah an option to disable would be helpful.
Can you send me your policy file so I can check?
It is the one listed in https://github.com/MediaArea/MediaConch_SourceCode/issues/671#issuecomment-629148892
Also in MyPolicy.zip
We'll add the option (I think it will be off by default, and people will have to add the option if they want to test attachments), better for all.
@genfhk actually it is already disabled by default, but there is a bug when there are more than 1 attachment. This bug is fixed in development snapshots, please try a development snapshot.
Successful test with this snapshot. Passed without issue.
We need to do a new public release at some point...
I have recently had this same problem with MediaConch failures for the attachments in Version 21.09, encoding 4K with 4 attachments (including reversibility data). If it's any help @genfhk my work around for automation is to run the policy but then pipe the response to a grep searching for the 'pass!' and the DPX sequence path, like this:
mediaconch -p ${path_to_policy} ${DPX_sequence_path} | grep "pass! ${DPX_sequence_path}"
This ignores all attachment failures, and only returns a string with the pass for video stream 0. If video stream 0 fails it returns nothing.
I'm trying to create a policy for our new rawcooked MKVs per these specs, and 1) mediaconch is checking reversibility data, which I'd prefer it not do, and 2) the data is failing just the "FFV1 uses slice crcs?" rule (copied from the example public policy), which is also confusing...
Thanks!