Completed Show Intro Skip Option

[chef 3810]

15 hours ago, TeamB said:

I find this all a bit sad

What's sad teamB?

[Cheesegeezer 3104]

3 hours ago, crusher11 said:

And what about scans/edits I've already done? Is all my intro point data going to get deleted?

I'm sure we can take info from the existing titleSequence database and migrate it, but i'm not sure it's worth spending the time on it, because there was such a low number of testers, that a rescan will probably just be the solution.

However, it was always suggested to try this on a small library, while it was in dev, alpha and beta testing.

Cheers

 

[TeamB 2438]

8 minutes ago, chef said:

What's sad teamB?

The work we did in the beginning to bring in some of the work from other projects on the audio fingerprinting, to work out how to get the correct fp data from ffmpeg, to then use that data to find similarities in the episode audio streams. It was some interesting work and i am sorry to see it go closed source.

Edited April 18, 2022 by TeamB

[crusher11 1101]

3 minutes ago, Cheesegeezer said:

I'm sure we can take info from the existing titleSequence database and migrate it, but i'm not sure it's worth spending the time on it, because there was such a low number of testers, that a rescan will probably just be the solution.

However, it was always suggested to try this on a small library, while it was in dev, alpha and beta testing.

Cheers

 

I haven't done much, but you or @chef said in one of the more recent updates that it was in its final form and the database would be retainable thereafter, at which point I started to work on it a bit more seriously in preparation for the integrated version.

[chef 3810]

49 minutes ago, crusher11 said:

I haven't done much, but you or @chef said in one of the more recent updates that it was in its final form and the database would be retainable thereafter, at which point I started to work on it a bit more seriously in preparation for the integrated version.

Right, but the task is pretty quick.

When the newest version has all the bells and whistles, we can do something.

Right now we are seeing advancements in core integration, which (in my opinion) is what we have been striving for.

We have to be patient. Emby has a lot of moving parts and, really, only two lead developers handling it all. 

It's exciting! It's going to be awesome!

[crusher11 1101]

That's as may be, but right now it's frustrating because I'm not sure how much I can do with it, given the risk of everything being wiped.

[chef 3810]

1 minute ago, crusher11 said:

That's as may be, but right now it's frustrating because I'm not sure how much I can do with it, given the risk of everything being wiped.

The plugin is sperate from what is happening now. 

You can run both, but turn off chapter extraction in the plugin.

Let the new version handle chapters.

The plugin database will not be touched. That data is yours, and will be available, untouched.

[crusher11 1101]

Well, reportedly we need to re-scan right now, which means losing existing edits.

Not sure what you mean by "turn of chapter extraction".

[chef 3810]

1 hour ago, TeamB said:

The work we did in the beginning to bring in some of the work from other projects on the audio fingerprinting, to work out how to get the correct fp data from ffmpeg, to then use that data to find similarities in the episode audio streams. It was some interesting work and i am sorry to see it go closed source.

I can add you to the original plugin GitHub repo if you want to see how it all worked originally.

The only reason I kept the plugin repo private was because I was hoping to see emby get the ability before ... the other guys. (Although the other guys have decent programmers too... And I'm sure they'd figure it out if they wanted too).

Emby has the ability. I'm not going to make the original plugin repo public, but for inquiring minds and programmers like you teamB

[chef 3810]

8 minutes ago, crusher11 said:

Well, reportedly we need to re-scan right now, which means losing existing edits.

Not sure what you mean by "turn of chapter extraction".

Your existing edits are safe.

The new version implements intro skip by adding the data to the chapter objects in emby.

We suggested small library samples to test edits, and detection.

 

[crusher11 1101]

How are they safe if I need to re-scan? This is what confuses me.

I don't have a great deal of edits to maintain, but I'd rather keep them than start over if that's possible.

[Cheesegeezer 3104]

1 hour ago, crusher11 said:

How are they safe if I need to re-scan? This is what confuses me.

I don't have a great deal of edits to maintain, but I'd rather keep them than start over if that's possible.

Hi Crusher,  what we are saying is that all your edits will not be destroyed by moving over to the core version.

the core version has no way to edit these points anyway, but you might actually find that the core is more accurate and pick up all your intros correctly anyway, therefore you wont need your edits.  We would love to have testers for the core integration and also feedback, this will speed up the move from dev to beta.  then eventually release.

we are waiting for beta or final release from emby core before we update our plugin to use the data from the core.  This will bring us back to where we were with editing, etc.  

If you don't want to try the core version or you want to use your current edited points.  Stick with the plugin until there is a more final version from the core.

Not rocket science my friend

[fillidill 175]

I am so happy this is moving along and I appreciate all the effort people involved have contributed with and will. You guys rock!

People who don't understand what beta means should keep away until there is something stable.

[BillOatman 596]

3 hours ago, TeamB said:

The work we did in the beginning to bring in some of the work from other projects on the audio fingerprinting, to work out how to get the correct fp data from ffmpeg, to then use that data to find similarities in the episode audio streams. It was some interesting work and i am sorry to see it go closed source.

The plugin code is also closed source at the moment.

[Cheesegeezer 3104]

1 minute ago, BillOatman said:

The plugin code is also closed source at the moment.

And will remain that way for sure

[crusher11 1101]

1 hour ago, Cheesegeezer said:

the core version has no way to edit these points anyway, but you might actually find that the core is more accurate and pick up all your intros correctly anyway, therefore you wont need your edits.

The majority are just one-off different intros or whatever, nothing that would be picked up on either way.

[TeamB 2438]

4 hours ago, Cheesegeezer said:

And will remain that way for sure

Is this because of jellyfin?

[rbjtech 5284]

Unless you are desperate for the manual 'Skip Intro' button, then personally I think it is far too early to abandon the Plugin if it is delivering the functionality you have come to expect.

Remember, currently, in Core the following are missing or have no plans to be implemented by the Core :

• no individual episode level editing (intro+credit)
• no show/season level editing
• no credit detection
• no auto-skip (intro and/or credit)
• no user selectable options
• no tagging / recalculation of chapters
• no performance tuning options
• no show exclusion options
• no options for showing e01 of a season
• no statistics / reports

Also to remember -

Each MANUAL edit (or Bulk Manual Change) that was performed is currently tagged in the Introskip database (Confirmed=1) - thus these entries are prime candidates for IMPORTING into the Emby Core Introskip.  I'll take this offline with the Introskip team, but this does not seem like a big issue - we have the episode id and all the necessary timings - so writing/overwriting this to the 'chapters' table in the Core dB, seems trivial.    There are a few questions however, on how to KEEP the edits - as we are not sure if they would get wiped with a metadata refresh for example.

As @chef explained above - My understanding is some/all of the above functions will go into the Core, others will remain as a Plugin - we are just trying to understand which is which at the moment.

So in summary, if you are happy with the Introskip Plugin (and there is no reason not to be..) then for the moment, I would personally keep using it - any manual edited data is safe in it's own dB with options to import it when ready.

Edited April 19, 2022 by rbjtech

[gillmacca01 211]

If I have run the introskip task for the core version only, is the introskip info compatible with the plugin version, and would it be able to use some of the options in the plugin to edit them?

[Cheesegeezer 3104]

33 minutes ago, gillmacca01 said:

If I have run the introskip task for the core version only, is the introskip info compatible with the plugin version, and would it be able to use some of the options in the plugin to edit them?

unfortunately not at the minute. as RBJ stated we are still figuring out what's in the core and what's not gonna be in there.  Until we know what's what, it's pointless marrying the two to exchange data.

Please remember this is very much in the dev phase for Emby Core and still in Beta for Plugin version.

[ebr 16169]

4 hours ago, rbjtech said:

Unless you are desperate for the manual 'Skip Intro' button, then personally I think it is far too early to abandon the Plugin

No one is asking anyone to abandon the plug-in if that is what they want to use right now.  Remember the core feature is a hidden option in a beta server release.  I would hope you would not have expectations of that being complete.

[rbjtech 5284]

2 hours ago, ebr said:

No one is asking anyone to abandon the plug-in if that is what they want to use right now.  Remember the core feature is a hidden option in a beta server release.  I would hope you would not have expectations of that being complete.

Agreed - but peoples expectations are that it matches or betters the Introskip Plugin - so I'm highlighting the current differences - the choice of what they use is up to them .. 

Edited April 19, 2022 by rbjtech

[Weevillo 30]

21 hours ago, TeamB said:

Is this because of jellyfin?

Well they're already working on a intro skip feature of some sort, and eyeing audio fingerprinting which is how Emby's plugin works. So from a competitive standpoint being close sourced for now make sense.

Also @chef was wondering about a thing: You stated that the plugin would be free-of-charge in the thread a little while back, but now (from my understanding) it seems like it won't be. We're (well I'm not at least, it's probably the biggest buff Emby's had in a while) angry if that's the case, but could you guys be transparent about it if so would be the case?

[Weevillo 30]

I don't know if this is a know an issue (or even an issue for that matter), but wouldn't it be a good idea to code the intro skip to limit it's detection to the detected intro length? Like here it says that my season's intro is 1:30 min:

But it's not uncommon (perhaps 1-3 false detections per show/season) for these types of detections:

These might've been resolved in the core version, I haven't tried it. It might be a bug as well? Since the episode is 24 minutes in total and you've mentioned that it detects a certain percentage of the first few minutes depending on episode length, but you wouldn't expect an intro over 50% of the total runtime into the episode.

[maxiniedermeier 17]

Would it be possible to use this feature as an OutroSkip