Home User Forums SongKong Music Tagger Jaikoz Music Tagger Albunack Music Service

Wednesday 12 September 2018

SongKong 5.9:Improved Song Identification and Performance

SongKong 5.9:Improved Song Identification and Performance

We are pleased to announce SongKong 5.9 with improved song identification and performance.

Improvement to Albunack MusicBrainz Disc Ids 

In SongKong 5.7 we introduced Albunack Disc Ids, this provides a way to identify complete albums without requiring acoustic fingerprinting, both speeding up song identification and increasing the number of matches that can be found because it was able to find albums without known acoustic fingerprints or existing metadata.

We have now improved this algorithm to get more matches. Testing against a 1200 album test set with SongKong 5.7 could identify  501 albums just using Albunack Disc Ids, with SongKong 5.9 that figure goes up to 570 albums, an improvement of 13 percent. (After subsequent matching with AcoustIds and metadata we can identify 1108 albums.

Better Performance

We have made a number of further performance improvements. Many of these improvements reduce cpu usage so have most impact on slower/single cpu computers. Such computers should see a 30% increase in performance.

  • SongKongs internal database has been optimized to reduce the number of calls made to the database 
  • Scoring algorithm has been optimized   
  • AcousticBrainz data can now be looked up per album instead of having to do individual lookups for each song


Support Message now can be sent with Support Files

Now when you send us your support files you can send a support message at the same time. This makes the reason for sending the support files clear and avoids you having to send a separate email message.

Help can now be hidden

In Remote mode help text was displayed on screen making it easy to see what all the options do. But this does take up screen space so we have now added the option to toggle Help on/off with the Show Help/Hide Help button

Report Creation has own Progress Bar

Previously no progress bar was shown for the Report Creation that occurs after the task has completed. However the report creation itself can take some time if you have been working with a large collection. So now Report Creation has its own progress bar that increases as the report is created making it clearer when the task is going to complete. The Report Creation progress bar uses a 1-100 percentage scale.

User Log is now more useful

The user log, songkong_user0-0.log has now been improved. Previously user log only contained errors and a summary after a task had completed. We now add a summary of song groups matched as they are matched. 

10/09/2018 18.18.12:BST:SEVERE: G1164:Matched to MB/AcoustId:Ave Maria:9 of 12 to Ave Maria:cf1d4e8f-ae69-46cb-9c5c-c941951ded58 by Vyatcheslav Kagan-Paley
10/09/2018 18.18.12:BST:SEVERE: G1164:Linked to Discogs:Ave Maria:9 of 12 to Ave Maria:7135466 by Vyatcheslav Kagan-Paley
10/09/2018 18.18.13:BST:SEVERE: G8883:Matched to MB/AlbunackAcoustId:Picture This:12 of 12 to Picture This:423f245e-1c75-4948-9f48-cc4c5c8159f4 by Wet Wet Wet
10/09/2018 18.18.13:BST:SEVERE: G8883:Linked to Discogs:Picture This:12 of 12 to Picture This:6859962 by Wet Wet Wet

Each line is preceded by date, and then a Group Id of the form Gnnnn. The Group Id is simply a way for us to identify a group of songs (usually a folder). Then we have a message explaining how matched to MusicBrainz or Discogs

Whilst the user log is primarily a debugging aid for use by JThink you may find this additional information useful.

Many Bugs Fixed

We have been hunting down some of the more obscure bugs in a sustained effort to try and get bugs down to zero.In this release there are over twenty bug fixes

The full list of improvements and fixes can be found on the News Page

1 comment:

Coffee Recipes said...

Thiis was lovely to read

Jthink blog Jthink Facebook page google_plus Jthink YouTube channel Email Paul at Jthink Subscribe to Weekly Newsletter