[Logo] Jaikoz and SongKong Forums
  [Search] Search   [Recent Topics] Recent Topics   [Members]  Member Listing   [Groups] Back to home page 
[Register] Register / 
[Login] Login 
Messages posted by: paultaylor  XML
Profile for paultaylor -> Messages posted by paultaylor [7298] Go to Page: Previous  1, 2, 3 ... 235, 236, 237 ... 242, 243, 244 Next 
Author Message
As you say Recording Time is used as Release Time I think I may just change it to the Year (although still allow full dates) because this si what most users know it has.

I like this autocorrect idea, although the correction/inserted of dates on an album basis would have to be optional.

Ive now stopped anonymous postings.
Anonymous Postings no Longer allowed. This has been done to stop spam and to solve the problem where you forget to login, and end up posting anonymously by mistake, the main problem with this is you cannot edit your posts.

ozone wrote:
(I first want to point out an obvious bug; If I select Default Word Separator for the "Split the Filename Using..." in the correct tags from filename setting, it doesn't save this selection when I click "ok". When I re-open the correct tags from filename settings window, it just defaults back to "tags from filename word separator"...)
 

Noted, and added to list.

ozone wrote:
So is there some way that you could implement some AI to determine this and match this stuff up instead of Jaikoz skipping over it? I don't know if it is possible but that would be great.
 


The general logic is that tags (even ID3v1) are more likely to be correct than filenames, so I dont think it would normally make sense to overwrite tag data with filename data. You could force this to happen yourself by just deleting the tagdata and then running tag from filename.

Nocturnal wrote:
What I am wondering about is the error format:
Code:
 [Fatal Error] :-1:-1: Premature end of file.
 23/06/2007 01.06.15:WARNING: Unable to create Puid for:11:null
 

What programs generate these? The first one doesn't seem like your log format. [edit]I found out it's an error from your XML parser. So probably the output received from genpuid isn't up to par all the time...[/edit] The second one states the list number (so probably from your program) and then 'null'. Where do you print this message? What should follow the list number? Seems like that parameter isn't initialized...

I'll retry this with more debug output later...
I'm also going to try redoing the entire directory with genpuid, seeing if it also sometimes fails on different files. 


Thankyou for taking the time to look at this, it appears that the problem is not file related but related to the interaction between Jaikoz and Genpuid.

This lines comes from jaikoz:
23/06/2007 01.06.15:WARNING: Unable to create Puid for:11:null
The null is because there is an exception but it has no error message, it is not because something is not inilitized.

I dont know where
[Fatal Error] :-1:-1: Premature end of file.
is coming from but I guess it is coming from genpuid in some way. Ive just reread your message how do you its from the xmlparser, thats interesting.
Yes please try with more debugging, this may give the extra information I need to fix this.
There does seem t o be a problem with the `, I think this is because it is a special character in java Regular Expression lanaguage.

BTW It isnt a word in this context, a word has to be separated by word separators such as spaces

e.g Love Won ` t Wait Vinyl

should work but it doesn't, we need a Punctuation Replacer as well
Hi, yes I think that is already on the list

ozone wrote:
I want to be able to choose in the options menu whether I want the recording time displayed as... YEAR-MONTH-DAY (i.e., 2007-12-06) or YEAR (i.e., 2007).

If implementing this would take some time, then can you at least in the next patch standardize the current method? Currently, Jaikoz appears to do no analysis during auto correction on the recording time. So you could have 1 album with five files in YEAR-MONTH-DAY format, and six files in YEAR format. I wish it would tidy the recording time up automatically for me instead of leaving songs in the same album with a different format. 


Do you want it DISPLAYED as YEAR/ YEAR-MONTH_DAY .. or STORED as YEAR/YEAR_MONTH_DAY. I ask because just changing how it is displayed in jaikoz is not going to change how it is displayed in another application such as iTunes. The recordingtime value is taken from the value in the MusicBrainz database when a match is made and the value exists.

Assuming you mean STORED could add a Local Correcter for the 'Recording Time' column which lets you specify the format values should be held at could be an option. However if you specify you would like to store it as a full date YYYY-MM-DD and you only have the year value what should Jaikoz do. For exaxmple if the field has a value of 1984 storing it at 1984-01-01 would be misleading as it implies more accuracy then it actually has, whereas leaving it as 1984 would mean a mixture of formats.

So waiting for comments...
If Jaikoz calculates a Puid it will be stored as a Music IP Puid field, so as long as you save the file its Puid will be saved. The subsequent Tag From MusicBrainz task will not remove any Puids even if it does not exist in the MusicBrainz database.

However the message "cannot be analysed" indicates that a Puid could not be created for the track, so no puid will be created for this track.
These are the sketched out instructions for doing a language translation:

Let me know what language you would like to translate and I will provide an auto-translated for your language. This gives a reasonable starting point so it should take less of your time.

I will send you the latest default file and the file for your language, the example instructions below uses Dutch.

Replace jaikoz.properties in your installation with the one attached
Also copy jaikoz_nl.properties to the same location

Goto https://prbeditor.dev.java.net/
Download prbeditor-0.9.7_1.zip file and install by unzipping to your folder of choice.
Double clickon prbeditor-0.9.7_1.jar, it should start automatically
From the Dialog select 'Open Resource File'
Then select Jaikoz.properties from within the Jaikoz Installation Folder
And select only the Dutch Locale
This will load all the properties key with their (Default) english in one column and the Dutch value in the next column

You can now type in the values as required and correct the values, you don't have to supply every value to use the file.
One thing to be careful of is some of the values contain parameters, these are of the form {0}.{1},{2}.. You must ensure the same number of parameters are passed to the Russian language strings in order for it to be correct.

You can save changes with select 'File/Save' option
Now if you start Jaikoz and select 'Settings/Language Settings' you can select Dutch, and it will start using the values in the Dutch properties file, if a value doesnt exist it will use the default (English) value.

When you are happy with it you can send me the jaikoz_nl.properties file and I will put it into the next release of Jaikoz,

You will be credited in the acknowledgements section of Jaikoz. I am also considering the possibility of sharing the the revenues of Sales that come from your country, but this isnt definite yet.

thanks Paul

Anonymous wrote:
Okay. Let me try to explain what I mean one more time. The songs that Jaikoz didn't touch, because it couldn't --- the songs that didn't have PUID's and didn't have a close enough percentage match [the percentage you put in the MusicBrainz settings] to MusicBrainz to change --- should be separated at the end, so that THOSE songs could be manually looked up, since Jaikoz couldn't do anything with them. 


Yes, but if you run Tag from MusicBrainz on all your files the list of files that you want to put aside because MusicBrainz did not match them is IDENTICAL to the list of files you would get by selecting 'Action/Filter MusicBrainz UniqueIds/MusicBRainz Unique id Does not Exist', have you tried it ?

Anonymous wrote:
It'd be great to be able to customise which fields would appear in the details panel - just an idea... 


Thats a good idea in the long run. But one of the aims of adding this was to make Jaikoz simpler to use for casual users so I will release with predefined layouts to start with, using iTunes getInfo() as an inspiration.

Pitman6787 wrote:
Okay. If you are a fan of older music, then you might have a lot of compliation/best of/greatest hits albums. If you run Jaikoz, Jaikoz usually changes the album names back to the original studio albums in which the songs were on
 


This is not a decision made by Jaikoz, but a consequence of the makeup of the MusicBrainz database. If you have an Acoustic Id Jaikoz uses this to do a match using the record that is returned, if more than one record matches the Acoustic id then it selects the record with the best matching data. So if your album is the 'Queens Greatest Hits' compilation and one of the tracks returned by the Acoustic Id has an album of 'Queen Greatest Hits' this will be the one selected. If you don't have an acoustic id the matching is purely done on the values in your record, so if your record has an album name of 'Queens Greatest Hits' this is likely to receive a better score than the original album the track was from.

I expect what is happening in your case is the track you check with the Get Acoustic Ids is sonically identical on the greatest hits version and the original album version. But when the the Acoustic Id is checked against the MusicBrainz Database it only finds the track linked to the original album. because users of the MB database are more likely to tag origin albums than compilations. So Jaikoz updates the record to this.

If you don't want the album to be changed, you could change the settings of 'Action Settings/MusicBrainzSettings/Format' for Album to be 'Only Populate if Empty'

Ok, so this problem has happened on three occasions nows, and always with Linux, I haven't managed to replicate it myself.

Can you clarify for me is this happening consistently for the same files. For example in the debug log where it says Unable to create Puid for:132, the 132 refers to the 133nd file (because records are labelled from zero) in the list when sorted by record number. So if this problem occurs again after identifying the record you could right click on it s record header and just select 'Remote Correct/Correct Acoustic Ids' to retry this one file.

If it does reoccur could you send me the file and/or try against Windows or OSX to see if the problem persists.

Another thing that might be useful would be if you increased the logging output but this will slow down jaikoz so you only really want to do it for when running a test , not all the time. You can increase the logging by using the jaikoz.sh shell script and editing the value -l2 in jaikoz.sh to -l7.

Pitman6787 wrote:
Because if it is, it shouldn't. The album name shouldn't be relevent, since one song can be on more than one album, and, if someone has a song, trying to match partly based on the current album name might make it harder to find the song on THAT SPECIFIC ALBUM, and it could alter the percentage of similarity, which could cause it not to change. 


Yes it does match on the album if a value has been provided. Yes songs can be on many albums but by default surely it makes sense to return the top match the one that matches on title/artist/trackno and album rather than only on title and artist. If you have a MusicIP Puid then that takes precedence and the album is only one of a number fields that help pick the best match for a PUID for the cases where one PUID matches more than one track.

If you are not getting any matches at all you can always reduce the matching percentage required, or use the Manual Match to see the top ten choices for each track.

Nocturnal wrote:

Take for instance the Radiohead album "OK Computer". It has a song titled "Exit Music (For a Film)". If you remove the parenthesis from the title, but the parenthesis ARE present in the Musicbrainz DB, wouldn't the matching be made more difficult?
 


Yes, I think you are right, the Punctuation Remover should be disabled by default. This was implemented sometime ago and has not been looked at for some time. Ill disable it for the next release.
Good work !

I'll think about the consequences of this wrt jaikoz. (I cant use recursive mode because the analysis has to be done on the records as selected and in the order they appear on screen which might not relate to the filesystem)
For Mac customers I am trying hard to make Jaikoz match the Apple Human Interface Guidelines as accurately as I can but I'm not sure what the main issues are. Please help us out by telling me what Jaikoz does that could be improved from an Apple perspective.
I am adding a detail panel to Jaikoz, thats shows details of the currently selected record. The first panel is a read only summary , but other panels are editable allowing you to change the details, and the changes get reflected in the datasheet immediately. The panels will contain information broadly similar to what you would get by selecting on 'Get Info' on a record within iTunes but will also contain additional MusicBrainz Info.

Below is a screenshot of the mac version:



The idea is that you can get access to all information without having to scroll right as you sometimes have to do using the datasheet, and can quickly edit that information as well. The information is also presented in an audioformat independent way rather than ID3 specific as is the case with the datasheet view.

I would be interested in any suggestions you have for this.

Acoustic analysis is the slowest task but you only ever have to do it once per track, 25seconds seems a bit slow.

Ive found that acoustic analysis doe use alot of cpu, but I think some songs take significantly more resources than others. Ive actually implemented multi-threading for acoustic analysis but in my own tests I didn't get much improvement in throughput because the cpu was already under heavy usage with just one thread. Ill take another look when I have time, I know you have a PC but what operating system are you running on (Windows/Linux)?

Hi I don't know if this is related to your /ext2 filesystem because I dont seem to have this problem, please clarify.

On OSX enter U2\u01ff into artist field ( adds a character that requires 2 byte unicode encoding)
Run Correct Filename from Tag, U2? shown in fieldname.
Save file, U2? shown in filename on hard-disk.
I'll investigate, for users less technical than yourself it would be better if the filtering was automatic so that filenames are always valid for the filesystem being used, but this may be difficult to do with so many possible variations.
Hi what OS are you using

On OSX if there are ? in the field they are not removed from the filename when you run 'Correct Filenames from Tags' , but on OSX I it is ok to create files with '?'s in the filename, on Windows they are filtered out.

'*'s are automatically removed whether they are inputted by running 'Correct Filenames from Tags' or by typing asterisks directly into the filename or subfolder fields.

So I cant see the problem
Hi, Ive never seen this behaviour. Can you send tell me how and when it goes down, and send me your jaikozuser0-0.log and jaikozdebug0-0.log

Anonymous wrote:
I can't find a task in the Autocorrecter settings for Retrieve AcousticID. 

The task in the lhs panel is Retrieve Acoustic Ids,(its not enabled by default so you need to copy it from the left panel to the right panel)

smika wrote:

1. Tag OGG files (I know someone else made a topic about it)
 

Yes, I am working towards that

smika wrote:

2. When I upgrade to a newer version, all my setting I made with an older version will not be overwritten. 

Yes, this is on the list as well.
In Jaikoz 1.10 (released today) this is now easier to do from the MusicBrainz Detail tab.
Hi, are you trying to install it from a console (non-window) environment, possibly remotely. The errors indicates it cant start the Xwindowing environment

Anonymous wrote:
Hey is there a batch or command prompt, which will let say take file, analyze it to get audioID, correct tag comparing to MusicID, change filename and saves it. Running Jaikoz under java in kde, seems to take lots of resources...it doesnt matter how long it will take for me, if it will go one by one and you can see the result
 

There is no commandline program to do this, however you can do everything except the save automatically using the Autocorrecter.
Configure the Autocorrecter
Run it, leave it alone
When it s finished do a quick check and then Save changes

Anonymous wrote:

you wrote that best approach is to do 5000 songs per session, how do you load a limited amount of songs to jaikoz and how you will exclude those already repaired?
 

Well you can select folders to add using the Open Folder and the Add Folder options, there is no need to load all records - in fact I wouldnt reommend it. That said I am going to provide a file tree browser in the future to make it easier to select what you want. All the tasks can be performed against all records, or by right clicking on the row header only selected records.

Get AcousticId will skip tracks that already have an audio id
If you go to 'Action Settings/MusicBrainz Settings/Match' and enable 'Do not match online if already have a Musicbrainz Match' Tag from MusicBrainz will skip tracks that already have a MusicBrainz Id.

Most of the task that are performed locally work by comparing records with each other so the taks will always be peformaed against all/selected records.
Ok, basically MusicIP Puid is not stored by Music IP mixer because it has been added fairly recently and they havent decided how to store it yet. However it is used by Jaikoz to retrieve Puids so that tracks that have already been analysed by Music IP Mixer will get processed quicker than files that have not already been processed by MusicIP Mixer.
More details here:
http://forums.musicip.com/index.php?showtopic=2865

newname wrote:

1. Is the acoustic analysis the same in both programs? If so, why is the analysis in Jaikoz much quicker than MusicIP? Is MusicIP more accurate since it's slower?
 

MusicIP will add records to the MusicIP if they dont currently exist whereas
Jaikoz only trys to match records that already exist in the database. Creating a new id is much much slower than matching an existing id and there is no possibility of it immediately matching a record in MusicBrainz because it has only just been created which is why Jaikoz does not do this, I will add it as an extra task/option at some point. I dont know if this accounts for all the differences, MusicMixer may do some other analysis that isnt required for identification purposes.

newname wrote:

2. What exactly is "MusicMagic Data" and "MusicMagic Fingerprint"?
 

These are only saved to your file if you are using Music Mixer Premium and select 'Archive Analysis' . They are the data that is generated from the
file and sent to MusicIP, in order for MusicIP to find a track and return a MusicIP Puid -Im not sure of the difference between the two. I dont know why Music IP Mixer doesnt store the Music IP Puid as well, but I am looking into it.

newname wrote:

3. Why won't Jaikoz just accept the analysis tags from MusicIP?
 

Jaikoz needs to get a Puid in order to lookup the track in Musicbrainz but unfortunately this is not stored by Music Ip Mixer.

newname wrote:

4 (and last). For all the songs that Jaikoz/MusicBrainz doesn't recognize the songs (PUIDs, tags, etc) and doesn't supply a "Unique File ID", will submitting the info to the server (MusicBrainz/PUID pair) over time get these songs on the database?
 

Do you mean submitting them from Jaikoz, no tracks can only be submitted to musicbrainz as a uniqueid/puid pair. But records added to MusicIP do overtime get added to MusicBrainz
 
Profile for paultaylor -> Messages posted by paultaylor [7298] Go to Page: Previous  1, 2, 3 ... 235, 236, 237 ... 242, 243, 244 Next 
Go to:   
Powered by JForum 2.1.6 © JForum Team