[detectads] Automatically bookmark ad breaks

Also worth a mention that you can run DetectAds automatically following Decryption by going to Settings for detectads package.
 
MymsMan said:
Detectads requires the file to be decrypted first and symptoms that you see are common if the file has not yet been decrypted,

It would be helpful if it checked for the Decrypt flag and gave a more meaningful message,

Do you have autodecrypt specified for the folder? Otherwise you will need to run decrypt manually first
Click to expand...

Thanks. You were right, I got confused over what was decrypted and what wasn't.
 
For the first time that I am aware of DetectAds failed to spot two ad breaks in this evenings Celebrity Chase program
Code: 
15/11/2014 20:50 - Processing item 244 (/media/My Video/The Chase_ Celebrity Special_20141115_1920.ts) from the queue
15/11/2014 20:50 -   DETECTADS: /media/My Video/The Chase_ Celebrity Special_20141115_1920.ts
15/11/2014 20:50 -   ad break found (0 - 15)
15/11/2014 21:00 -   ad break found (2585 - 2656)
15/11/2014 21:02 -   ad break found (3158 - 3257)
15/11/2014 21:02 - done...processed in 725.717s - 3 ad breaks bookmarked
15/11/2014 21:02 - Removing item 244 (/media/My Video/The Chase_ Celebrity Special_20141115_1920.ts) from the queue
The missed breaks were approx 18 and 29 minutes into the program
 
MymsMan said:
For the first time that I am aware of DetectAds failed to spot two ad breaks in this evenings Celebrity Chase program

The missed breaks were approx 18 and 29 minutes into the program
Click to expand...

I reran the analysis manually and silences were detected during the missed ad breaks but obviously not sufficient to be recognized as an ad break
Code: 
15/11/2014 23:33 - DETECTADS: /media/My Video/The Chase_ Celebrity Special_20141115_1920.ts
15/11/2014 23:33 - Initial bookmarks: 0 15 2585 2656 3158 3257
15/11/2014 23:33 - debug@Threshold=-70.0, MinQuiet=0.20, MinDetect=5.0, MinLength=60.0, MaxSep=90.0, Pad=0.00
15/11/2014 23:33 - debug@Threshold=679094, MinQuiet=5, MinDetect=5, MinLength=1500, MaxSep=2251, Pad=1
15/11/2014 23:33 - debug@Frame rate is 25.00, Detecting silences below 679094 that last for at least 5 frames
15/11/2014 23:33 - debug@Clusters are composed of a minimum of 5 silences closer than 2251 frames and must be
15/11/2014 23:33 - debug@longer than 1500 frames in total. Cuts will be padded by 1 frames
15/11/2014 23:33 - debug@< preroll, > postroll, - advert, ? too few silences, # too short, = comm flagged
15/11/2014 23:33 - info@ Start - End Start - End Duration Interval Level/Count
15/11/2014 23:33 - info@ frame - frame (mmm:ss-mmm:ss) frame (mm:ss.s) frame (mmm:ss)
15/11/2014 23:33 - debug@ Silence 1- 47 ( 0:00- 0:02), 47 ( 0:01.9), 0 ( 0:00), [ 154776]
15/11/2014 23:33 - debug@ Silence 349- 384 ( 0:14- 0:15), 36 ( 0:01.4), 303 ( 0:12), [ 152917]
15/11/2014 23:33 - info@< Cluster 1- 384 ( 0:00- 0:15), 384 ( 0:15.4), 0 ( 0:00), [ 3]
15/11/2014 23:33 - cut@= Cut 1- 383 ( 0:00- 0:15), 383 ( 0:15.3), 0 ( 0:00), [ 0]
15/11/2014 23:33 - ad break found (0 - 15)
15/11/2014 23:37 - debug@ Silence 27545- 27559 ( 18:22- 18:22), 15 ( 0:00.6), 27162 ( 18:06), [ 301465]
15/11/2014 23:37 - debug@ Silence 28534- 28559 ( 19:01- 19:02), 26 ( 0:01.0), 976 ( 0:39), [ 146195]
15/11/2014 23:37 - debug@ Silence 29273- 29297 ( 19:31- 19:32), 25 ( 0:01.0), 715 ( 0:29), [ 144179]
15/11/2014 23:37 - debug@ Silence 30036- 30084 ( 20:01- 20:03), 49 ( 0:02.0), 740 ( 0:30), [ 168521]
15/11/2014 23:37 - info@? Cluster 27545- 30084 ( 18:22- 20:03), 2540 ( 1:41.6), 27162 ( 18:06), [ 4]
15/11/2014 23:39 - debug@ Silence 45047- 45058 ( 30:02- 30:02), 12 ( 0:00.5), 14964 ( 9:59), [ 234837]
15/11/2014 23:39 - debug@ Silence 46035- 46046 ( 30:41- 30:42), 12 ( 0:00.5), 978 ( 0:39), [ 141994]
15/11/2014 23:39 - debug@ Silence 46786- 46796 ( 31:11- 31:12), 11 ( 0:00.4), 741 ( 0:30), [ 154903]
15/11/2014 23:39 - debug@ Silence 47515- 47583 ( 31:41- 31:43), 69 ( 0:02.8), 720 ( 0:29), [ 181411]
15/11/2014 23:40 - info@? Cluster 45047- 47583 ( 30:02- 31:43), 2537 ( 1:41.5), 14964 ( 9:59), [ 4]
15/11/2014 23:42 - debug@ Silence 64618- 64633 ( 43:05- 43:05), 16 ( 0:00.6), 17036 ( 11:21), [ 204799]
15/11/2014 23:42 - debug@ Silence 64861- 64872 ( 43:14- 43:15), 12 ( 0:00.5), 229 ( 0:09), [ 103765]
15/11/2014 23:42 - debug@ Silence 65361- 65371 ( 43:34- 43:35), 11 ( 0:00.4), 490 ( 0:20), [ 137029]
15/11/2014 23:42 - debug@ Silence 65860- 65875 ( 43:54- 43:55), 16 ( 0:00.6), 490 ( 0:20), [ 155648]
15/11/2014 23:42 - debug@ Silence 66351- 66397 ( 44:14- 44:16), 47 ( 0:01.9), 477 ( 0:19), [ 119916]
15/11/2014 23:42 - info@- Cluster 64618- 66397 ( 43:05- 44:16), 1780 ( 1:11.2), 17036 ( 11:21), [ 5]
15/11/2014 23:42 - cut@= Cut 64619- 66396 ( 43:05- 44:16), 1778 ( 1:11.1), 0 ( 0:00), [ 0]
15/11/2014 23:42 - ad break found (2585 - 2656)
15/11/2014 23:44 - debug@ Silence 78952- 79008 ( 52:38- 52:40), 57 ( 0:02.3), 12556 ( 8:22), [ 183960]
15/11/2014 23:44 - debug@ Silence 79122- 79169 ( 52:45- 52:47), 48 ( 0:01.9), 115 ( 0:05), [ 114687]
15/11/2014 23:44 - debug@ Silence 79896- 79919 ( 53:16- 53:17), 24 ( 0:01.0), 728 ( 0:29), [ 114687]
15/11/2014 23:44 - debug@ Silence 80646- 80669 ( 53:46- 53:47), 24 ( 0:01.0), 728 ( 0:29), [ 117418]
15/11/2014 23:44 - debug@ Silence 81397- 81437 ( 54:16- 54:17), 41 ( 0:01.6), 729 ( 0:29), [ 118284]
15/11/2014 23:44 - debug@> Silence 81437- 81437 ( 54:17- 54:17), 1 ( 0:00.0), 1 ( 0:00), [ 0]
15/11/2014 23:44 - info@> Cluster 78952- 81437 ( 52:38- 54:17), 2486 ( 1:39.4), 12556 ( 8:22), [ 6]
15/11/2014 23:44 - cut@= Cut 78953- 81437 ( 52:38- 54:17), 2485 ( 1:39.4), 0 ( 0:00), [ 0]
15/11/2014 23:44 - ad break found (3158 - 3257)
15/11/2014 23:44 - Final bookmarks: 0 15 2585 2656 3158 3257
15/11/2014 23:44 - done...processed in 683.111s - 3 ad breaks bookmarked
=== Refresh page to see bookmarks ===
 
Last edited:
if I could suggest a 'feature' - how about taking 5 seconds off the bookmark position at the end of ads (so bookmarks 3,5,7,9,etc.) then you'd know if any of the prog was missed due to a miscalculation.
[not sure if this is possible]
 
This seems to be working really well over all.

I would suggest that since me and others find that the selection needs inverting and we can not be sure which way around is best. A way to tell is to keep the larger bits and cut the smaller bits.
I don't think there is a channel with more ads than show....but who knows.

I would like to process a massive folder of a few hundred recordings, then invert and crop. Is there a way to automat this? Sorry if this has been mentioned and I've missed something.

Thanks again, great package.:)
 
Last edited:
Has there been any further development of this excellent package? In particular:
  1. Option of automatically running Nicesplice to cut out the ads that have been found
  2. Bypassing DetectAds on non-commercial channels
  3. Tweaking detection options - I have found some episodes of The Chase have short ad breaks that aren't recognized which means when I press the skip button I suddenly skip a long way through the program
 
You must log in or register to reply here.
Back
Top