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

Song Tagging and Metadata Blog

Wednesday, 22 May 2019

SongKong 6.1 Elephant - Now with Artist Images released May 22nd 2019


We are pleased to announce the new release of SongKong, SongKong 6.1 Elephant named after the rocking fourth album by The White Stripes.




Artist Artwork


We have a major new feature in this release. SongKong can now find artist images for your music. And when we say artist images we don’t just mean the release artists, we can find also find images for the tracks artists, composer, lyricist, conductor, and orchestra! The images are added to your actual music files so that they are transferable.

 
We do special processing for the Album Artist, if the album artist is made up of multiple artists and we can find images for each album artist we try to stitch the images together to create a single image that represents the album artist.


Artist Image per Artist Folder


We can also save album artist artwork to each Artist folder if your songs are organized Album Artist/Album folder. This can be useful if using an UPnP server and like to browse music by folder. The album artist will be stored as a folder.jpg file within the each Artist folder and this means that your uPNP control point should use the artist image as the folder image.

You can see in this screenshot below (using Linn Kinsky Control Point being served by MinimServer) that artist folders with an artist image look so much nicer than the default folder image.


Artist Image Repository

SongKong can add all the artist artwork found to a single folder, named by artist name, this gives you all your artist images images in one folder making it easy to find them if you want to use them for other purposes.


Caching

When Songkong downloads an album from MusicBrainz it has previously always stored the  album details in the internal SongKong database. This means that if we need the album again we do not have to retrieve it from the Internet again. But databases introduce overheads such as locking that we do not really need since the album is read only, we have no need to change it, and we also have the issue that the database will continue to grow over time.

We have now moved to using a Cache instead called EhCache. This is faster than using a database and automatically prunes the data when it reaches a certain size, removing the albums that have not been accessed for the longest period of time.

This improves performance, and we have also added caches for Artist, Acoustid and AcousticBrainz lookups.


We also have a number of other fixes and improvements in this release.

Friday, 29 March 2019

SongKong 6.0 Fumes, now with Profiles - released 29th March 2019

SongKong 6.0 Fumes Released 29th March 2019

We are pleased to announce the release of SongKong 6.0 Fumes on 29 th March 2019. This new release is named after the album Fumes by american pop-folk duo Lily & Madeline, similar to First Aid Kit but less well known & better. We have cheated a little, favorite album is their eponymous debut  but that album does not match the self-imposed one word album name rule for SongKong releases.


Profiles

A long requested feature has been the ability to save multiple SongKong configurations, and then be able to switch between them to perform specific tasks. Now we have added them, they are called Profiles




You can modify an existing profile or create a new one with the Clone button based on the current profile. Any modifications made to your configuration will be made to the currently selected profile.

If you want to delete or rename a profile, this can be done from the Preferences page, we have added some predefined profiles.


Rename Masks

When we looked at profile masks it was clear that if you created a new filename mask it would make sense for that mask to be available to all profiles. So we have now moved the created and editing of filename masks into Preferences.



and within the Fix Songs task, there is now just the option to select the rename and compilation rename masks.





Update of Reports

We have now updated all reports so that they no longer use html framesets as these are no longer supported in Html 5. And we have updated the style so that they match the pages of the Remote User Interface.



Create Support Files

We had an issue with potential customers uploading their support files before they had a license, and therefore we had no idea who had sent the support files! So we have now added an email field, this will be auto populated for customers who already have a license.



Full details of all improvements can be found here

Tuesday, 5 March 2019

SongKong 5.15 Crocodiles Released 5th March 2019

We are pleased to announce the release of SongKong 5:15 Crocodiles on 5th March 2019. This new release is named after the debut post punk album by Echo & The Bunnymen from 1980. Great tunes, guitar, vocals and image on one of the best debuts ever.


Compilation now means Compilation 

Originally the definition of a Compilation that was used by SongKong was the same as the one used by MusicBrainz. This means that both Various Artist Compilation albums and a Single Artist Compilation are considered to be compilations and have the Is Compilation field set. But most of us consider these two cases differently, and only really consider the Various Artist album as being a compilation.

So now only Various Artist compilations are marked as compilations, and we have added a new field called Is Greatest Hits  for a single artists Greatest Hits albums, this makes more sense.

Things are slightly different for Classical music  because it is difficult to determine the distinction we have for now continued with Is Compilation for all types of classical compilations.




File renaming for HD albums

We have added some new filename masks that will put HD albums in a separate top level folder called HD. They make use of a new addHD() javascript function, and we have also added similar addCompilation() and addGreatestHits() functions.


Improvements to Song Only Matching

SongKong always tries to match logical groupings of songs to one album, if this is not possible then it does song only matching, in such cases we only update song only fields such as Song Title and not album fields such as Track No. But often it is the case that all but one song was matched to the same album, and there is just some minor problem preventing all songs being matched. 

If subsequent Song Only Match matches the groups of songs to recordings that have links to the same MB Release then we also add some basic release information such as Album Name, Catalogue No and Cover Art using the most popular releases for the songs. So for each song we use the most popular release that the song is linked to.

And we know that Cover Artist a key piece of metadata customers like to have. And it is much easier to remove an incorrect piece of artwork then manually search and add artwork yourself, but we only add album data when songs are missing that data we don't overwrite exsting data.


And we only add album data for albums that are not compilations. The reasoning behind this is that we cannot be sure we have the correct album because we cannot match all songs to the album, but if it is an original album there is a good chance it is the correct one, or a variation of the correct one. But if it a compilation then because popular songs can find their way onto many compilations there is a good chance that this is actually the wrong one.



Improved Matching

We have extended Albunack DiscId matching so that it can now work for multi disc albums as along as each disc in your collection is stored in a separate folder under the parent folder. 

We have increased the fuzziness of Discogs Albunack DiscId matching to find more potential hits. And we have fixed a bug when using Albunack DiscId matching with Mp4 files.

Plenty more fixes, please see here




Thursday, 31 January 2019

Jaikoz 10 Released:Improved Matching

We are pleased to announce the first release of Jaikoz 10, in this major release we have concentrated on improving probably the most important feature of Jaikoz - matching your music collection to MusicBrainz.

There are three key new features:

Song Only Matching

Song Only Matching  ensures that Jaikoz can at least partially identify almost all your songs  even if it cannot match a group of songs to an album. 

We explain Song Only Matching in detail here
 

Albunack Disc Id Matching 

AlbunackDiscIds is an alternative to Acoustid matching that can be faster and lead to more matches for both MusicBrainz and Discogs albums.

We explain AlbunackDiscIds in more detail here

 

Match Directly to MusicBrainz

First of all Jaikoz looks in our dedicated Albunack server instead of directly contacting MusicBrainz. But if no match can be found we now do an additional search directly against the MusicBrainz database.

We explain Direct MusicBrainz matching in more detail here


Console Improvements


Also note the Jaikoz console now gives more information on how your songs were matched





Jaikoz, Match Directly to MusicBrainz Release

Jaikoz, Match Directly to MusicBrainz Release

Jaikoz does not do most of it searches directly from Musicbrainz or Discogs, instead it uses Albunack. Albunack is a single database containing both MusicBrainz and Discogs datasets. It is specially designed for tagging your music and hence allow matching much quicker and much more reliably than if Jaikoz had to use MusicBrainz and Discogs directly.

Albunack is updated regularly from MusicBrainz and Discogs with the latest data, but there can be a delay of up to one month. So if you are tagging the latest releases or have added a release yourself to either of these databases it will not be in Albunack immediately.

With Jaikoz 10  we have added an extra search to MusicBrainz itself when we are unable to find a match in Albunack. This ensures you always have access to latest releases as soon as they are added to MusicBrainz.

Jaikoz Albunack Disc Id Matching

Albunack Disc Id Matching 

When you have a complete album stored in one folder with no missing or additional tracks, and the track order can be safely determined from either existing Track No metadata or the filename we can generate an AlbunackDiscId for your album. We can then look for a MusicBrainz match in the Albunack database, if we find matches we double check by looking at any existing metadata and/or generating AcoustIds for the first track.

This has three advantages over fingerprinting all songs and then finding matches based on fingerprints/acoustIds:

1. It is faster

Generating Acoustid fingerprints is a cpu intensive task, even on a fast machine it can take about a second per track. This might not seem much but it can become a bottleneck, especially on a machine with slow cpus, or only one cpu.


2. It finds more matches

Sometimes your album may be in MusicBrainz but there are no acoustic fingerprints linking to tracks on the MusicBrainz release. If you have no existing metadata then we cannot find a match based on fingerprints/metadata but we may be able to find a match based on AlbunackDiscIds !


3. It can match Discogs releases

The AlbunackDiscId is not just for MusicBrainz releases, it can also be used to match Discogs releases. This is very helpful, since we cannot directly use AcoustIds to match to Discogs.

Do I need AcoustIds for All My Songs ?

If you get an Albunack match then an acoustId fingerprint only needs to be created for one track. However there are occasions when you may want to get acoustIds for each tracks regardless.

AcoustIds are one way of finding Duplicate files so if you are using one of these options you may want to add AcoustIds for all songs.




Also the Filters menu lets you filter by variations of AcoustIds.

You can ensure AcoustIds have been added for every song before running Delete Duplicates by using the Retrieve AcoustIds Task 



Jaikoz Song Only Matching

Previous Method

Jaikoz always attempts to group song that belong to one album to one album. If it cannot achieve this the the songs remain unchanged, you can override this by disabling the Only Allow match if all songs in grouping match to one album
option.


But then this can cause a grouping of songs that do actually represent one album to be split up if the album simply does not exist in MusicBrainz yet.

Song Only Matching


Song Only Matching works differently. If Jaikoz fails in its attempt to match a group of songs to an album it then just attempts to match the songs via AcoustId. Having found a match it add song details such as Title, Artist and MB Recording Id but not album details such as Track No and MB Release Id .


We can add some album information

But its a bit cleverer then that. Firstly when an AcoustId is linked to some MusicBrainz Recordings, we look at the MB Releases that these recordings are linked to. And if the groups of songs have recordings that have links to the same MB Release then we also add some basic release information such as Album Name, Catalogue No and Cover Art using the most popular releases for the songs. So for each song we use the most popular release that the song is linked to.

And we know that Cover Artist a key piece of metadata customers like to have. And it is much easier to remove an incorrect piece of artwork then manually search and add artwork yourself.


We do not add MusicBrainz Release Ids

Because we are not sure if this is the correct album we do not add any MusicBrainz Ids that affect MusicBrainz released based tasks, so we don't add MusicBrainz ReleaseGroup Id, MusicBrainz Release Id or MusicBrainz Track Id
 

We do not replace existing album data

But we only add this additional information if these fields are currently blank for your songs, so we don't overwrite your existing data as we cannot be sure its 100% correct. 

We do not use Compilation album data

And we only add album data for albums that are not compilations. The reasoning behind this is that we cannot be sure we have the correct album because we cannot match all songs to the album, but if it is an original album there is a good chance it is the correct one, or a variation of the correct one. But if it a compilation then because popular songs can find their way onto many compilations there is a good chance that this is actually the wrong one.

We can use User Submitted Metadata 

There are many tracks in AcoustId database that we cannot find in MusicBrainz, there are currently 45 million unique recordings in Acoustid compared to 20 million in MusicBrainz.

 

In these cases we can at least make use of the basic user submitted metadata provided by Acoustid. This allows us to add artist, song title and sometimes album for these songs. For album we try use metadata from original albums that rather than compilations.

All these features of Song Only Matching should ensure that Jaikoz can at least partially identify almost all your songs without requiring modification to the default settings.

 

Thursday, 17 January 2019

Hint and Tips:How to install SongKong database, logs and reports in a different location on a PC

SongKong stores a cache of the songs loaded and an audit of the changes made in its own database. It also creates comprehensive reports and detailed logs. With al this details there is no denying that over time these can eat into your disk space.

On Windows they are stored within

 C:\Users\username\AppData\Roaming\SongKong

where username is your Windows user.


Recovering Disk Space


All can easily be removed and the space recovered by using the correct menu item:

Database: Empty Database
Logs :Empty Log Files
Report: Delete Reports 

but what about if you want to keep them, but just store the files in a different location in the first place ?


Windows Junction Points

With Unix and MacOS this can be easily done using symbolic links but these are not available with Windows. 

However Windows has Junction Points. these can be applied to make a folder link to a different location. For example here 

MKLINK /J C:\Users\Paul\AppData\Roaming\SongKong E:\SongKong

I have pointed 

C:\Users\Paul\AppData\Roaming\SongKong 

to actually point to

E:\SongKong
 
e.g



Full Procedure

  • Shutdown SongKong
  • Create new SongKong file location (e.g E:\SongKong)
  • Copy the contents of old location to new location
  • Start command prompt (search for cmd) in toolbar search
  • Create Junction Point (mklink /J JunctionPoint  Folder)
  • Restart SongKong

and that should be it.

Monday, 14 January 2019

SongKong MusicTagger 5:14 Sister, Improved Matching and Metadata Release

We are pleased to announce the release of SongKong 5:14 Sister on 14th January 2019. This new release is named after the album by Sonic Youth from 1987. Sonic Youth have a great sound but can sometimes be a bit self-indulgent for my liking, but this album is full of sharp angry tunes plus the classic Schizophrenia


 

Improved Albunack DiscId Matching

Six months ago we introduced our own additional album matching algorithm that we call AlbunackDiscIds. When you have a complete album stored in one folder with no missing or additional tracks, and the track order can be safely determined from either existing Track No metadata or filename we can generate an AlbunackDiscId for your album. We can then look for a match in the Albunack database, if we find matches we double check by looking at any existing metadata and/or generating AcoustIds for the first track.

This has two advantages over fingerprinting all songs and then finding matches based on fingerprints/acoustids:


1. It is faster

Generating Acoustid fingerprints is a cpu intensive task, even on a fast machine it can take about a second per track. This might not seem much but it can become a bottleneck, especially on a machine with slow cpus, or only one cpu.


2. It finds more matches

Sometimes your album maybe in MusicBrainz but there are no acoustic fingerprints linking to tracks on the MusicBrainz release. If you have no existing metadata then we cannot find a match based on fingerprints/metadata but we may be able to find a match based on AlbunackDiscIds.

In this release we have increased the fuzziness of the track time matching, and in testing this has increased the number of AlbunackDiscId matches by 10%, giving a reduction in the time taken to match your music and an increase in the amount of music matched.

Improved AcoustId Matching

Conversely, there are many tracks in AcoustId database that we cannot find in MusicBrainz, in these cases we can at least make use of the basic metadata provided by Acoustid. We have now improved how we use the user submitted metadata provided by AcoustId to add more metadata to your songs, and we have also made modifications so that SongKong is more likely to use metadata from original albums that compilations.


Edit Metadata

Previously Copy Row 1 could be used to copy the metadata for row 1 to the other rows in the album, but what if you only want to copy metadata for a particular field.

In The Remote Browser interface we now have Copy Row 1 button for each column to do just this, and the old Copy Row 1 button has been renamed Copy Full Row 1


With the Desktop interface, the button is availble for each column when you right-click on a column header.

 

Fixes

We also have a number of fixes in this release. As always, full details on the News page.

 


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