Jump to content

Show Intro Skip Option


Liquidfire88

Recommended Posts

STR8

I checked my outcome for star trek and it shows the intros to be true but the numbers are not the same... don't know if this means anything or is it fine, just sharing.

 Screenshot_1.thumb.jpg.1e905cc2874079d8135400f0fd40e423.jpg

  • Thanks 1
Link to comment
Share on other sites

kaj
16 minutes ago, rbjtech said:

Has the Detection Task been run ? - if you run it manually, does it skip through all the shows ?

If it has been run, then it looks as if this show's episode has not been identified for some reason.  Is it exactly the same Intro per episode ?  Are the other seasons the same ?  If it is, then it seems very odd that this has not been picked up for any episodes. 

Yes, I ran the detection task, took 4 minutes. Yes, the same incredibly annoying title music, and all other TV Shows are showing the same thing, no intro.

 

embyserver (20).txt

Edited by kaj
Link to comment
Share on other sites

Cheesegeezer
1 hour ago, kaj said:

I've gone to the plugin settings - Intros tab. Selected a TV Shows (Start Trek: Enterprise - Season 1, incredibly annoying title sequence) and this is what it says:

image.thumb.png.d95b8f5a251f48525c5756f1b0d1bcdc.png

Hi Kal,

What you can do now, is click the Reset Data For Season 1.  Then "REMOVE ALL"

Now we want to increase the detection accuracy, so go to the advanced settings tab and switch off Fast Detection. (leave the setting at 0.6) and run the FP & detection task again.

Hopefully you will see some good results.

Link to comment
Share on other sites

Cheesegeezer
43 minutes ago, STR8 said:

I checked my outcome for star trek and it shows the intros to be true but the numbers are not the same... don't know if this means anything or is it fine, just sharing.

 Screenshot_1.thumb.jpg.1e905cc2874079d8135400f0fd40e423.jpg

So the addition of the thumbnails is great because now i can see that all your intros are detected correctly just by looking at the images.  They are pretty much all identical, which would suggest that the title sequences have been identified correctly.

  • Like 2
Link to comment
Share on other sites

rbjtech

According to your log - this show has not had any Detection yet (this is why no Intro's have been found) ..

Let me dig a bit deeper ..

Capture.PNG.a1fd55867aa6e2f6f75b1fde3c3a4a49.PNG

Edited by rbjtech
Link to comment
Share on other sites

rbjtech

@kaj - The only shows that have any Intro Detections on them - are Wandavision S1 and Game of Thrones S5 - it then errors with a SQL error - meaning the reset of the shows are not being detected.

@Cheesegeezer @chef

2021-10-15 08:11:07.861 Warn Intro Skip: One or more errors occurred. (Exception of type 'SQLitePCL.pretty.SQLiteException' was thrown.) (Exception of type 'SQLitePCL.pretty.SQLiteException' was thrown.)

 

Edited by rbjtech
Link to comment
Share on other sites

chef
1 hour ago, rbjtech said:

@kaj - The only shows that have any Intro Detections on them - are Wandavision S1 and Game of Thrones S5 - it then errors with a SQL error - meaning the reset of the shows are not being detected.

@Cheesegeezer @chef

2021-10-15 08:11:07.861 Warn Intro Skip: One or more errors occurred. (Exception of type 'SQLitePCL.pretty.SQLiteException' was thrown.) (Exception of type 'SQLitePCL.pretty.SQLiteException' was thrown.)

 

I just don't know.

 

@kaj your library has some series, season and episodes I'm not sure about:

 

This emby library entry hasn't got a season number attached.

2021-10-15 05:16:56.933 Debug Intro Skip: TV - S: - E:3: Extracting chunks from binary chroma-print.
2021-10-15 05:16:56.935 Debug Intro Skip: TV - S: - E:3: .bin file removed.

 

That would throw intro skip for a loop because everything we do revolves around working within season.

Here is another item in your library that is interesting:

2021-10-15 05:14:11.449 Info Intro Skip: FINGERPRINT: Season 1 - S: - E:1 complete - 18 seconds.

Is the series called "Season1"? If so it also doesn't have a season number attached to it.

 

Also here in GOT, some of those episodes have no index number either.

2021-10-15 08:07:57.428 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E: 4 - detection took 6892 milliseconds.
2021-10-15 08:08:03.708 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E: 5 - detection took 6279 milliseconds.
2021-10-15 08:08:09.979 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 6271 milliseconds.
2021-10-15 08:08:15.977 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5997 milliseconds.
2021-10-15 08:08:21.946 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5969 milliseconds.
2021-10-15 08:08:27.922 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5976 milliseconds.
2021-10-15 08:08:39.794 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5965 milliseconds.
2021-10-15 08:08:39.872 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E: 1 - detection took 5959 milliseconds.
2021-10-15 08:08:45.785 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5990 milliseconds.
2021-10-15 08:08:45.850 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 5977 milliseconds.
2021-10-15 08:08:51.790 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 6004 milliseconds.
2021-10-15 08:08:51.856 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 6005 milliseconds.
2021-10-15 08:08:57.761 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5971 milliseconds.
2021-10-15 08:08:57.828 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 5972 milliseconds.
2021-10-15 08:09:09.777 Debug Intro Skip: Game of Thrones - Season 5 - E: 2 matched E:  - detection took 5983 milliseconds.

 

I'm gonna to have to think about what is happening here, I'll be back soon.

  • Like 1
Link to comment
Share on other sites

Bradum
5 hours ago, chef said:

😬 Currently, it would be hard to get the recaps. Because we wouldn't really know where they are. Some times it could be story line prior to the intro, sometimes they could be recaps.

But outros, yes that is on the Todo list.

Infact the actual extraction code has be written and is already there, just not implement yet,  using fingerprinting, and I know there was talk of other ways to pinpoint the outro.

Thanks for the feedback!

Oh for sure, recaps would be impossible to do automatically. I'm just the kind of OCD that would be willing to enter the timestamps in for each video manually.

  • Like 1
Link to comment
Share on other sites

kaj
2 hours ago, rbjtech said:

@kaj - The only shows that have any Intro Detections on them - are Wandavision S1 and Game of Thrones S5 - it then errors with a SQL error - meaning the reset of the shows are not being detected.

 

This is what it shows for WandaVision, which doesn't look encouraging 

image.thumb.png.115ee3fd45c61ed059469976a9c101fb.png

1 hour ago, chef said:

I just don't know.

 

@kaj your library has some series, season and episodes I'm not sure about:

 

This emby library entry hasn't got a season number attached.


2021-10-15 05:16:56.933 Debug Intro Skip: TV - S: - E:3: Extracting chunks from binary chroma-print.
2021-10-15 05:16:56.935 Debug Intro Skip: TV - S: - E:3: .bin file removed.

 

That would throw intro skip for a loop because everything we do revolves around working within season.

Here is another item in your library that is interesting:


2021-10-15 05:14:11.449 Info Intro Skip: FINGERPRINT: Season 1 - S: - E:1 complete - 18 seconds.

Is the series called "Season1"? If so it also doesn't have a season number attached to it.

 

Also here in GOT, some of those episodes have no index number either.


2021-10-15 08:07:57.428 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E: 4 - detection took 6892 milliseconds.
2021-10-15 08:08:03.708 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E: 5 - detection took 6279 milliseconds.
2021-10-15 08:08:09.979 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 6271 milliseconds.
2021-10-15 08:08:15.977 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5997 milliseconds.
2021-10-15 08:08:21.946 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5969 milliseconds.
2021-10-15 08:08:27.922 Debug Intro Skip: Game of Thrones - Season 5 - E: 9 matched E:  - detection took 5976 milliseconds.
2021-10-15 08:08:39.794 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5965 milliseconds.
2021-10-15 08:08:39.872 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E: 1 - detection took 5959 milliseconds.
2021-10-15 08:08:45.785 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5990 milliseconds.
2021-10-15 08:08:45.850 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 5977 milliseconds.
2021-10-15 08:08:51.790 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 6004 milliseconds.
2021-10-15 08:08:51.856 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 6005 milliseconds.
2021-10-15 08:08:57.761 Debug Intro Skip: Game of Thrones - Season 5 - E: 1 matched E:  - detection took 5971 milliseconds.
2021-10-15 08:08:57.828 Debug Intro Skip: Game of Thrones - Season 5 - E:  matched E:  - detection took 5972 milliseconds.
2021-10-15 08:09:09.777 Debug Intro Skip: Game of Thrones - Season 5 - E: 2 matched E:  - detection took 5983 milliseconds.

 

I'm gonna to have to think about what is happening here, I'll be back soon.

I've checked and don't have any show called Season1. 

The GOT are old files. I'm happy to delete the entire show, to see if that helps.

Link to comment
Share on other sites

chef
8 minutes ago, kaj said:

This is what it shows for WandaVision, which doesn't look encouraging 

image.thumb.png.115ee3fd45c61ed059469976a9c101fb.png

I've checked and don't have any show called Season1. 

The GOT are old files. I'm happy to delete the entire show, to see if that helps.

Actually, WandaVision didn't get scanned. We broke intro Skip at the end of GOT, before we could scan Wanda vision.

So that wouldn't have intro data.

It's okay! I'm gonna PM you, and we're going to figure out where exactly we are tripping up the episodes, and then You'll have awesome intro data 😆👍

 

Edited by chef
  • Thanks 1
Link to comment
Share on other sites

neik

My fingerprinting is about to finish by today, so I just had a look on the detection options available.
First thing I noticed was the poster being out of place overview UI:

image.thumb.png.7e1f9a0010827993f904387ad391ffbe.png

Second, a question regarding the detection:
If I want to have the best results possible, does it even make sense to start with the fast detection?
Wouldn't that cause an enormous verification effort?

Link to comment
Share on other sites

Cheesegeezer

There will be a new dll up shortly.  For all to test.  No requirement to rescan the databases, however, you will need to clear your browser cache.

 

Link to comment
Share on other sites

chef
8 hours ago, neik said:

My fingerprinting is about to finish by today, so I just had a look on the detection options available.
First thing I noticed was the poster being out of place overview UI:

image.thumb.png.7e1f9a0010827993f904387ad391ffbe.png

Second, a question regarding the detection:
If I want to have the best results possible, does it even make sense to start with the fast detection?
Wouldn't that cause an enormous verification effort?

It is really up to you.

The fast detect will grab the first comparison, and move on.

 

But if it is turned off, then the plugin will gather all the detection results for every episode and weigh them out using statistical weights based on common durations in the season.

This takes longer. It still isn't perfect, but it will shift the sequences based on their position in the stream.

If it's close to the start, then it will calculate the most common duration from the start, and also choose the best comparison for the episode.

You can watch it all happen in the logs when debug is turned on.

I have turned off fast detect, and the results are a little more fine tuned. 

 

Would love to get feed back, which ever way you decide to go.

 

  • Like 1
Link to comment
Share on other sites

neik
4 hours ago, chef said:

I have turned off fast detect, and the results are a little more fine tuned. 

 

I definitely will do so. 😉

The poster issue I described above is - although a bit different - still present: on .15:

image.thumb.png.a3187c5bedf656c23c1d376bd87ee084.png

Link to comment
Share on other sites

CBers
2 minutes ago, neik said:

I definitely will do so. 😉

The poster issue I described above is - although a bit different - still present: on .15:

image.thumb.png.a3187c5bedf656c23c1d376bd87ee084.png

I had that earlier in testing and had to clear browser cache. 

Or, it could be screen resolution. 

 

Link to comment
Share on other sites

neik
25 minutes ago, CBers said:

I had that earlier in testing and had to clear browser cache. 

Or, it could be screen resolution. 

 

Using Edge with a cleared browser cache results in the same.

Screen resolution is 3000x2000 with 150% scaling (I think).

Edit:
Trying to access the plugin UI while detection is running all I see is the "circle of death".

Edited by neik
  • Thanks 1
Link to comment
Share on other sites

Cheesegeezer
27 minutes ago, neik said:

Using Edge with a cleared browser cache results in the same.

Screen resolution is 3000x2000 with 150% scaling (I think).

Edit:
Trying to access the plugin UI while detection is running all I see is the "circle of death".

Seems an odd resolution. And the scaling maybe causing issues with the layout. I will have to test tomorrow.

that circle could be because it’s waiting for a season to actually load. There is no order to which shows are processed.
 

You should wait until detection has completed and then report back any issues please. 

do you think we need to lock the UI until detection is completed?

Link to comment
Share on other sites

chef

Yes, the page with table should maybe have some kind of message that says something simple like "Detection task running..."

But users will want to access the advanced settings (hopefully before running the task but who knows..).

A quick call to the scheduled task manager should do well before loading introskipconfiguration.js fully.

👍

Perhaps we should use  a pixel count on the width instead of a percentage, that sould fix the poster sizing.

 

 

Edited by chef
Link to comment
Share on other sites

neik
8 hours ago, Cheesegeezer said:

do you think we need to lock the UI until detection is completed?

 

6 hours ago, chef said:

Yes, the page with table should maybe have some kind of message that says something simple like "Detection task running..."

 

I think something like this would be good so that the user knows the UI isn't (fully) available while detection is running.

Another point: I just had a look at the confidences that were calculated over the night and they either seem to be 0,67, 1,00 or 1,33:
image.png.1f901fcb98d0245ec9b6656a152fc68f.png

Seems a bit more like a pattern than really calculated values?

  • Thanks 1
Link to comment
Share on other sites

Cheesegeezer
8 hours ago, chef said:

Yes, the page with table should maybe have some kind of message that says something simple like "Detection task running..."

But users will want to access the advanced settings (hopefully before running the task but who knows..).

true

8 hours ago, chef said:

A quick call to the scheduled task manager should do well before loading introskipconfiguration.js fully.

Yes that what i meant 

8 hours ago, chef said:

👍

Perhaps we should use  a pixel count on the width instead of a percentage, that sould fix the poster sizing.

 

 

From top and right. Easy enough to do, or use horizontal layout for top section and anchor options left and poster center. 

  • Like 1
Link to comment
Share on other sites

chef
5 hours ago, neik said:

 

I think something like this would be good so that the user knows the UI isn't (fully) available while detection is running.

Another point: I just had a look at the confidences that were calculated over the night and they either seem to be 0,67, 1,00 or 1,33:
image.png.1f901fcb98d0245ec9b6656a152fc68f.png

Seems a bit more like a pattern than really calculated values?

That's interesting.

I also see a lot of values with 0.67.

It seems that when we divide the result start/duration/end times into the common start/duration/end times and calculate an average weight, we get some reoccurring numbers.

I wonder what that means? 🤔

Perhaps, it is because we are losing decimal places because it isn't a true double number value.

Im going to try and cast the result values to a double before attempting the division,  and see if we get some more specific outcomes.

 

 

Edited by chef
  • Like 1
Link to comment
Share on other sites

BaukeZwart

I installed v.0.0.1.15 yesterday and today I noticed quite a few shows in the stats tagged with "needs attention".
I checked a couple with MKVToolNix, none of the problem episodes have chapters.

I als tried to reset data for a few seasons and run the skipintro routine again but that didn't change anything.
What can I do to improve this.

Log:
https://0bin.net/paste/DxwayRK4#HK+QKrxynSo3DED1+JwaTWEidELz3h6UJqGzkksdA17

chrome_fLl0txUydT.thumb.png.42559dc60133af2bac664fb00b9a69bc.png

Edited by BaukeZwart
Link to comment
Share on other sites

Cheesegeezer
21 minutes ago, BaukeZwart said:

I installed v.0.0.1.15 yesterday and today I noticed quite a few shows in the stats tagged with "needs attention".
I checked a couple with MKVToolNix, none of the problem episodes have chapters.

I als tried to reset data for a few seasons and run the skipintro routine again but that didn't change anything.
What can I do to improve this.

Log:
https://0bin.net/paste/DxwayRK4#HK+QKrxynSo3DED1+JwaTWEidELz3h6UJqGzkksdA17

chrome_fLl0txUydT.thumb.png.42559dc60133af2bac664fb00b9a69bc.png

This is purely for detection.  This has nothing to do with Chapters.

This is to show you how many of your seasons may need to get rescanned with Fast detect off.

So lets take American Crime Story Season 3

You have 7 Episodes in the Season.  Averaging 14sec long.  only 43%(3 episodes) were detected  as having an intro that matched.  So you have 4 episodes requiring attention.

So you either go into the Intros tab and fix the problem for start and end times or.... maybe those episodes actually don't have a intro or the intro is less than 10secs long.

EDIT:

image.png.882dd88d060929699fa04b56314d67fe.png

Edited by Cheesegeezer
Link to comment
Share on other sites

BaukeZwart

I started fresh with fast detect off. Curious to see the outcome of that. 

EDIT: outcome with fast detect off (Detection confidence: 0.6) is almost identical, same episodes turn up in stats as needing attention.

Edited by BaukeZwart
  • Thanks 1
Link to comment
Share on other sites

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 account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...