Completed Show Intro Skip Option

[BaukeZwart 94]

1 hour ago, Cheesegeezer said:

please read the instructions here. Let us know if you are still having issues.  But you need to run the Thumbnail Extract Image task afterwards or enable IntroSkip to automatically do this after the chapter insert task has run.

Earlier today I ran Audio Fingerprinting and when that was done Title Sequence Detection with 2.0.2.4.
Later today updated the plugin to 2.0.3.5, enabled Show Title Sequence in Chapters and ran that.
Then enabled  Auto Image Extraction, ran for 1 sec, but nog chapter image to be found.

Log is filled with these messages, noting about image creation.
 

2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 5: The competition
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for The competition
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 6: Episode 6
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Episode 6
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 7: Episode 7
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Episode 7
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 8: Monopoly
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Monopoly
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 1: Pivot
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Pivot

 

[BaukeZwart 94]

Also this isn't looking correct.
Double entries and sequence start is way off. I don't see the double entries for other shows, but I haen't found a single episode with correct sequence start time.
 

Edited September 17, 2021 by BaukeZwart

[Cheesegeezer 3087]

1 minute ago, BaukeZwart said:

Earlier today I ran Audio Fingerprinting and when that was done Title Sequence Detection with 2.0.2.4.
Later today updated the plugin to 2.0.3.5, enabled Show Title Sequence in Chapters and ran that.
Then enabled  Auto Image Extraction, ran for 1 sec, but nog chapter image to be found.

Log is filled with these messages, noting about image creation.
 

2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 5: The competition
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for The competition
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 6: Episode 6
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Episode 6
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 7: Episode 7
2021-09-17 19:48:31.228 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Episode 7
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 8: Monopoly
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Monopoly
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: TV Show: Grow - Season 1
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Getting Chapter Info for 1: Pivot
2021-09-17 19:48:31.229 Info Intro Skip: CHAPTER INSERT: Title Sequence Chapter already Added for Pivot

 

So images have already been added you will need to do a RefreshMetadata on your TV Show library.

Wait for it to finish

Run the Chapter insert task again 

 

[Cheesegeezer 3087]

4 minutes ago, BaukeZwart said:

Also this isn't looking correct.

Double entries and sequence start is way off.
 

I think you need to start a fresh.  shut down emby and delete the titlesequence.db.

Run the fingerprint and detection process again.  

[BaukeZwart 94]

2 minutes ago, Cheesegeezer said:

I think you need to start a fresh.  shut down emby and delete the titlesequence.db.

Run the fingerprint and detection process again.  

Will do that, no idea what is going on

 

[Cheesegeezer 3087]

2 minutes ago, BaukeZwart said:

Will do that, no idea what is going on

 

You certainly shouldn't have 2 instances of the same episode in the list.

Remember this is still in development testing phase, there will be some bugs but we appreciate you testing so you can provide feedback

[samuelqwe 313]

50 minutes ago, rbjtech said:

Don't rename it - that will restart the FP task again - just copy it !

edit - ah if you mean rename the copy - then sure .. 

Yeah, I meant just the copy. Sorry for the confusion 

[samuelqwe 313]

32 minutes ago, BaukeZwart said:

Earlier today I ran Audio Fingerprinting and when that was done Title Sequence Detection with 2.0.2.4.

If you did the detection in fingerprinting with that version, then starting fresh is probably a good idea since there may have been changes to the way the data was stored since then.

Starting fresh would likely solve your duplicates problem.

[rbjtech 4274]

2 hours ago, Sammy said:

To get this straight.

• Let the FP task finish
• Shutdown Emby Server
• COPY titlesequence.db to any_name_you_like.db
• Restart Emby Server
• Start the Detect task

Update above .

You can removed all the *_old?.db files - they will serve no purpose now.

[Sammy 737]

1 hour ago, rbjtech said:

Update above .

You can removed all the *_old?.db files - they will serve no purpose now.

So the FP file is a one and done and serves no purpose after the initial or subsequent scans? If so why not trigger a cleanup to get rid of it?

[samuelqwe 313]

3 minutes ago, Sammy said:

So the FP file is a one and done and serves no purpose after the initial or subsequent scans? If so why not trigger a cleanup to get rid of it?

That .db file contains all the data the plugin uses: fingerprints and found intro points.

Once we have calculated the intro points, the majority of the fingerprint data already gets cleaned up from the db file since we don’t really need it anymore.

For this reason, we recommend backing it up before running the detection task simply to save time if you ever have to run the detection from scratch. We can’t re-detect an intro without that fingerprint data.

At the moment, we’re basically only making changes to the detection task, which is why you can simply re-use a copy of the db which contains only fingerprints if you want to test detection improvements.

[Sammy 737]

That's why I was thinking I could stop the FP scan and restore my last saved *_old8.db file to *.db and move to the detection scan right away.

[rbjtech 4274]

Just now, Sammy said:

That's why I was thinking I could stop the FP scan and restore my last saved *_old8.db file to *.db and move to the detection scan right away.

If you have run the Detect task on any of those db files (likely), then the FP data is gone - thus you need to ensure you copy it before you start the Detect task.

[chef 3746]

7 hours ago, BaukeZwart said:

Will do that, no idea what is going on

 

Hi, when you finish the second attempt. I'd really like to know how it went.  

If you still experience duplicate entries, I'd also love to see a log file, if you please

 

[BaukeZwart 94]

4 hours ago, chef said:

Hi, when you finish the second attempt. I'd really like to know how it went.  

If you still experience duplicate entries, I'd also love to see a log file, if you please

 

I probably try it later today. I will let you know.

[TBMike 7]

Hi everyone, I've been following this topic with great interest. Brilliant work so far by everyone. Just thought I'd share my findings with you.

My server runs Ubuntu 20.04 on an Intel Core i5-7600 CPU with 24GB RAM. My library has 173 shows with 25k episodes.

I just completed the FP task this morning with 2.0.3.4 which took just over 24 hours. Processing 2 series at once my CPU load sat at around 90%. I found that other scheduled tasks (from other plugins I think) kept stopping the FP task so I deleted all other scheduled tasks to let it run in peace for the first time. I have copied the db file and now I'm running the detect task using 2.0.3.5 and will report back when it finishes.

Edited September 18, 2021 by TBMike

[TBMike 7]

Interestingly I can set it to 10 series and my load runs between 70-90% for the detect task.

[Cheesegeezer 3087]

4 minutes ago, TBMike said:

Interestingly I can set it to 10 series and my load runs between 70-90% for the detect task.

Thanks for feedback, does it process them much quicker if you have it set to 10%

[TBMike 7]

2 minutes ago, Cheesegeezer said:

Thanks for feedback, does it process them much quicker if you have it set to 10%

The bar seems to move quicker but not exactly lightning fast, I can run it again when it's finished at 2 and see what the difference is.

[Markadmin 0]

On 9/16/2021 at 5:29 PM, samuelqwe said:

The Docker builds should have chromaprint included since Emby 4.6. However, it’s possible that using a third-party Docker container like the ones from LinuxServer and Binhex instead of the official Emby one, could not have the same FFmpeg build. I can’t confirm if that’s the case since I have only ever used the Docker official container.
 

 

@Markadmin I am on unRAID as well and I am using the official Emby container. In my case, it works perfectly.

If you can open a console for your Emby container (click on the icon of the container in unRAID, then Console), then you can run the following command:

/bin/ffmpeg -formats

Then, see if you can find a line (they’re in alphabetical order so look in the C’s) that says:

E chromaprint     Chromaprint

Because that’s what this plugin requires to function. If you do have the line, then we can try looking further at why it’s not working.

Hi thanks - using binhex Chromaprint was not available in ffmpeg so to make like easier i swapped over to the official docker and Chroma appeared. Unfortunately I still get the same error

 

2021-09-18 11:31:32.295 Info Intro Skip: Chroma-print database contains 0 items.
2021-09-18 11:31:32.320 Debug Intro Skip: -  aa S:1 - E:1: Extracting chunks from binary chroma-print.
2021-09-18 11:31:32.320 Debug Intro Skip: New aa(2018) - S:3 - E:1: Extracting chunks from binary chroma-print.
2021-09-18 11:31:32.320 Warn Intro Skip: aa- S:1 - E:1 .bin file doesn't exist.

Im not sure what .bin its looking for - any help would be appreciated - or i can just wait until the official release !  

[chef 3746]

2 hours ago, TBMike said:

Interestingly I can set it to 10 series and my load runs between 70-90% for the detect task.

Yeah, TPL (Task Parallel Library) is an interesting thing in the code. You may find it will push the CPU, or it might not. It's like a crazy driver on the highway. It will weave in and out threads. It might use 1 thread of the CPU, or it might decide it can more at once, if they are available.  

Let us know how you get on with the task. 

Thanks for taking the time!

 

[chef 3746]

1 hour ago, Markadmin said:

Hi thanks - using binhex Chromaprint was not available in ffmpeg so to make like easier i swapped over to the official docker and Chroma appeared. Unfortunately I still get the same error

 

2021-09-18 11:31:32.295 Info Intro Skip: Chroma-print database contains 0 items.
2021-09-18 11:31:32.320 Debug Intro Skip: -  aa S:1 - E:1: Extracting chunks from binary chroma-print.
2021-09-18 11:31:32.320 Debug Intro Skip: New aa(2018) - S:3 - E:1: Extracting chunks from binary chroma-print.
2021-09-18 11:31:32.320 Warn Intro Skip: aa- S:1 - E:1 .bin file doesn't exist.

Im not sure what .bin its looking for - any help would be appreciated - or i can just wait until the official release !  

Let me take a look at that error, because that particular log entry is from an older release of the plugin.

It could be something to do with folder permissions?

But, the plugin should have permission to save the .bin file where we are trying to place it.

I'll be back, we'll sort this out post haste!

 

@Cheesegeezer , @samuelqwe it would seem like the fingerprinting task is having an issue here saving the .bin. 

The Binary Reader is trying to read the .bin, but it isn't where we expect it to be. 

I would not have expected this. 

[TBMike 7]

Can I just check that the process to reset the detect task is to refresh the metadata on the library, then shutdown the server and copy the db file back? The metadata refresh is taking forever for me, just wanted to check that it is to be expected.

[BaukeZwart 94]

11 hours ago, chef said:

Hi, when you finish the second attempt. I'd really like to know how it went.  

If you still experience duplicate entries, I'd also love to see a log file, if you please

 

A fresh start did fix the double entry, but no chapter thumbs are generated. 

Step 3 is done in 1 sec when running it for the 1st after step. 

I will check the log when Im at my pc

Edited September 18, 2021 by BaukeZwart

[Cheesegeezer 3087]

9 minutes ago, BaukeZwart said:

A fresh start did fix the double entry, but no chapter thumbs are generated. 

Step 3 is done in 1 sec when running it for the 1st after step. 

I will check the log when Im at my pc

 

I scanned about 4000 episodes and it took 8 seconds for me

how long did the Thumbnail extraction take?  That is usually a good indicator to see if your chapters worked off the bad.

please send me the log.

 

Edited September 18, 2021 by Cheesegeezer