Manschk3rl 2 Posted June 23, 2022 Share Posted June 23, 2022 Hi, Is there a way to manually remove the "Intro Flag/Timestamp" so emby is forced to rescan it? Because if you update a existing TV-Show (because of a repack from the uploader or so), the skip intro button is not there. I guess because emby is not scanning for intros again on that episode because it already does. Or is there a way to manualy set the intro timestamp for this specific episode? Link to comment Share on other sites More sharing options...
Luke 37065 Posted June 23, 2022 Share Posted June 23, 2022 Hi, as long as the date modified timestamp of the file changes, then it will be rescanned for intros. Link to comment Share on other sites More sharing options...
Manschk3rl 2 Posted June 24, 2022 Author Share Posted June 24, 2022 (edited) Hi, sorry no its not. The original episode had the intro at the beginning at 00:00. The Repack have a "Previously on"-Scene at the beginning so the intro is now at 00:30. But the "Skip intro" still shows up at 00:00. Thats definitely not the intro : Â Edited June 24, 2022 by Manschk3rl Link to comment Share on other sites More sharing options...
rbjtech 4265 Posted June 24, 2022 Share Posted June 24, 2022 5 minutes ago, Manschk3rl said: Hi, sorry no its not. The original episode had the intro at the beginning at 00:00. The Repack have a "Previously on"-Scene at the beginning so the intro is now at 00:30. But the "Skip intro" still shows up at 00:00. Thats definitely not the intro : Â Refresh the metadata on that items and it will be re-scanned on the next schedule. Link to comment Share on other sites More sharing options...
Manschk3rl 2 Posted June 24, 2022 Author Share Posted June 24, 2022 (edited) 10 minutes ago, rbjtech said: Refresh the metadata on that items and it will be re-scanned on the next schedule. hi, unfortunately no. I refreshed the metadata and started the schedule manually but to no avail. the schedule only runs for 5 seconds. Edited June 24, 2022 by Manschk3rl Link to comment Share on other sites More sharing options...
rbjtech 4265 Posted June 24, 2022 Share Posted June 24, 2022 19 minutes ago, Manschk3rl said: hi, unfortunately no. I refreshed the metadata and started the schedule manually but to no avail. the schedule only runs for 5 seconds. Luke will likely want to see a log for the task. An editor for the Intro (like the Plugin below) is really needed to ensure accuracy or edit these episodes. I believe it's in the pipeline for the Core but I don't have any further details or timeline I'm afraid.   Link to comment Share on other sites More sharing options...
Manschk3rl 2 Posted June 24, 2022 Author Share Posted June 24, 2022 I just found out that when you click refresh metadata the intro isn't there anymore. As soon as the task for detecting the intro is running, the intro is there again. but again at the wrong time embyserver.txt Link to comment Share on other sites More sharing options...
rbjtech 4265 Posted June 24, 2022 Share Posted June 24, 2022 38 minutes ago, Manschk3rl said: I just found out that when you click refresh metadata the intro isn't there anymore. As soon as the task for detecting the intro is running, the intro is there again. but again at the wrong time embyserver.txt 384.41 kB · 1 download I'm not 100% sure how the core works vs the Plugin - but did you try a metadata refresh, then a library scan and then run the Intro Task again ? If the Intro really has moved, then the logic is to recreate the fingerprint files again - and then re-compare with the other episodes to determine the location of the Intro.  If the existing FP files are being used - then this is why it's just re-detecting the incorrect Intro timings as it's using the existing FP files. @Lukewill need to confirm how this works - it's possible a library scan may be required to remove the old FP files ? Link to comment Share on other sites More sharing options...
Manschk3rl 2 Posted June 24, 2022 Author Share Posted June 24, 2022 22 minutes ago, rbjtech said: I'm not 100% sure how the core works vs the Plugin - but did you try a metadata refresh, then a library scan and then run the Intro Task again ? If the Intro really has moved, then the logic is to recreate the fingerprint files again - and then re-compare with the other episodes to determine the location of the Intro.  If the existing FP files are being used - then this is why it's just re-detecting the incorrect Intro timings as it's using the existing FP files. @Lukewill need to confirm how this works - it's possible a library scan may be required to remove the old FP files ? Hi, i just tried it with refresh, rescan and run the task but no I think i just have to wait for a editor. Link to comment Share on other sites More sharing options...
MBSki 1016 Posted June 24, 2022 Share Posted June 24, 2022 51 minutes ago, rbjtech said: ... and then re-compare with the other episodes to determine the location of the Intro. This part is interesting. I've had mixed results with the core intro skip detection so far. Maybe this is why as I my episodes come from many different sources. DVR recordings, various streaming services, etc. The timing of the intro for episodes from different sources is going to be different. Are you saying though that the Core (and I assume the plug-in as well) use the start time of the intro in all the episodes of a season? And then what? Average out the start time? I'm not understanding why it's comparing to other episodes. Link to comment Share on other sites More sharing options...
ebr 14913 Posted June 24, 2022 Share Posted June 24, 2022 8 minutes ago, mbarylski said: I'm not understanding why it's comparing to other episodes. The way the detection works is by looking at all the episodes and finding a common piece of audio (the intro theme song, presumably). Then, once that common audio is identified, each individual episode is scanned for the occurrence of that specific audio "signature". Where ever that is, is assumed to be the intro for that episode. So, you need multiple episodes in order to determine the "common" audio (theme song). Does that make sense? 2 1 Link to comment Share on other sites More sharing options...
rbjtech 4265 Posted June 24, 2022 Share Posted June 24, 2022 (edited) 34 minutes ago, mbarylski said: This part is interesting. I've had mixed results with the core intro skip detection so far. Maybe this is why as I my episodes come from many different sources. DVR recordings, various streaming services, etc. The timing of the intro for episodes from different sources is going to be different. Are you saying though that the Core (and I assume the plug-in as well) use the start time of the intro in all the episodes of a season? And then what? Average out the start time? I'm not understanding why it's comparing to other episodes. To add to what Eric has said - the source of the file and location of the Intro within the total fingerprint length (default of 10 minutes for the core) should not matter - as the audio is actually turned into a numbered pattern and it is this which is matched. However, it is not an exact science - so if one source is lets say a 30fps NTSC TV capture using MP3 and another is a 23.976 Blu-Ray capture using HD-Audio - then there is a real possibility that the algorithm used would see the AUDIO as different - maybe the pitch or the tone (the video stream is not used at all) - this is highly likely to be inaudible difference to the human ear, but as computer 'patterns' - they are different 'enough' to be considered not the same. For the best detection results - try and keep the (audio) source the same - but it doesn't need to be a HQ source - consistency is key.  Edited June 24, 2022 by rbjtech 1 Link to comment Share on other sites More sharing options...
MBSki 1016 Posted June 24, 2022 Share Posted June 24, 2022 12 minutes ago, rbjtech said: To add to what Eric has said - the source of the file and location of the Intro within the total fingerprint length (default of 10 minutes for the core) should not matter - as the audio is actually turned into a numbered pattern and it is this which is matched. However, it is not an exact science - so if one source is lets say a 30fps NTSC TV capture using MP3 and another is a 23.976 Blu-Ray capture using HD-Audio - then there is a real possibility that the algorithm used would see the AUDIO as different - maybe the pitch or the tone (the video stream is not used at all) - this is highly likely to be inaudible difference to the human ear, but as computer 'patterns' - they are different 'enough' to be considered not the same. For the best detection results - try and keep the (audio) source the same - but it doesn't need to be a HQ source - consistency is key. Ah, well, that's interesting. I can't always keep sources the same since I have to find missing episodes wherever I can. I'll pay closer attention to which file is having issues though and compare to the others in a season to see if the sources are different. Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now