Genpuid advice

[erland]

I've determined that Archive Analysis within MusicIP GUI stores MUSICMAGIC DATA and MUSICMAGIC FINGERPRINT within mp3 files. Not sure what the purpose/difference is between those two tags.

Based on the names, it sounds like MUSICMAGIC FINGERPRINT is the fingerprint data needed to get a PUID and MUSICMAGIC DATA is the analysis data which MusicIP needs to be able to support smart mixes. I know MusicIP differs between the data needed for PUID detection and the data needed for music analysis. I've got the impression that the music analysis (which is required for smart mixes) is only possible through MusicIP and not through genpuid.

How does the MUSICIP DATA tag contents look like in comparison to MUSICMAGIC FINGERPRINT AND MUSICMAGIC DATA ?

Another problem seems to be where archive analysis has already occurred - I can't get PUIDs to be updated into the tags.

You mean that you don't have PUID tags and want them to be written ? Why ?

If you already have PUID tags I assume they don't need to be updated because the PUID should be the same as last time.

As far as I've understood, there are only two reasons to have the PUID:
- As an intermediate step get a musicbrainz identifier more automatically using picard or similar tagging tool.
- To get MusicIP analysis from http://www.musicdns.org to speed up MusicIP analysis if you need to do it again and haven't archived analysis to the tags earlier and lost the MusicIP library file.

If you want a unique identity, you want to use Musicbrainz id, the PUID is just an intermediate step to get it automatically. Musicbrainz identiers can be retrieved without a PUID but the lookup will then be based on the album/track titles and you will have to do a manual selection.

As I've understood, MusicIP doesn't require a PUID to do analysis, the PUID is just needed to speed up the analysis by retrieving it from www.musicdns.org instead of analyze the file again.

So if you already have Musicbrainz id tags and archived MusicIP analysis tags, you don't need the PUID.

GenPuid therefore must instead be writing the MUSICIP DATA tag - it doesn't detect when MUSICMAGIC DATA or FINGERPRINT tags already exist, and adds other tags (different tag names containing different values).

Therefore I don't know what benefit there is using genpuid to create MUSICIP DATA or PUID tags (if MusicIP uses MUSICMAGIC DATA tag instead).

I'm confused!

Maybe we are talking about the name change. Maybe MusicIP still uses the old MUSICMAGIC name while genpuid has been updated to use the new name MUSICIP ?
Are you using the latest version of MusicIP ?

I wonder what the Musicbrainz picard tagger uses when it retrieves Musicbrainz identifier, maybe it uses both ?
It has to use the stuff generated by genpuid because that's the reason Musicbrainz provides the genpuid binaries now when AmpliFIND no long does.

The main issue with all this is that MusicIP and genpuid is really no longer supported, the source code is closed, their forum is gone and AmpliFIND don't answer questions from end users, so it's probably going to be hard to get any answers on some of these questions.

What we really need is something that can replace MusicIP/genpuid which is officially supported by someone. It doesn't have to be open source, it's enough if someone supports it. I did some investigation a year ago in the following thread but I haven't really looked at it further since then:
http://forums.slimdevices.com/showthread.php?t=73504

For the automatic retrieval of Musicbrainz identities I know of at least one open source project started earlier this year which looks very promising. I'm not involved in it personally and I don't want to post a link to it here because it's in an early stage and I'm not sure if the developer wants it to be public yet.

[audiomuze]

>don't think puid's are used as unique identifiers, there's a separate
>musicbrainz tag for that. From memory there's a parameter you can pass
>genpuid to inhibit writing of puid's. Can look in morning for you
>
I thought PUID stood for Portable Unique IDentifier.

Ok, the parameter to stop PUID generation is "-nopuid ... bypass the PUID lookup. Useful for just analyzing files."

Correct re meaning of PUID, but it relates to "semantically identical" files - see http://musicbrainz.org/doc/HowPUIDsWork. So I'm guessing for every track you have that has the same PUID, it is in fact the same track, just a different version, length etc?

The TrackStat wiki page states: "If you have MusicBrainz tags in your music it will even survive a situation where you have moved or renamed your music files, without MusicBrainz tag only statistics for unmoved files with the same name as before will survive."

I don't know what actual tag this looks for - I think MusicBrainz uses PUIDs?

No idea which tag TrackStat references for its purposes, but I'm guessing it is MusicBrainz Track Id as opposed to the MusicIP PUID Identifier.

PUIDs are generated I think dependent on roughly the first 8 mins of music. Fingerprint analysis is from 2 mins of music. Neither seem to be guaranteed to be unique, although PUID's seem to suggest they are unique.

There's a clear explanation of the differences between analysis and fingerprinting here.

I seem to have ~5000 tracks that do not have unique PUIDs.

This means the MusicDNS server has not yet generated a PUID for the tracks in question. If you've run the full analysis via genpuid, the analysis result is sent to the server and utlimately a PUID is generated at the server side. Give it a few days, rerun genpuid against the tracks and you'll likely find that the PUID tag has been added. I've certainly seen this happen on my end.

[audiomuze]

I've got the impression that the music analysis (which is required for smart mixes) is only possible through MusicIP and not through genpuid.

I don't think this is correct. I've analysed tens of thousands of files using genpuid only and they mix properly in headless server or MiP GUI.

For the sake of clarity I've taken a single FLAC track, removed all tag data, generated an mp3 version of said track, and analyzed both using genpuid. What follows is the resulting tags added by genpuid.

genpuid's XML file:

<genpuid songs="2" xmlns:mip="http://musicip.com/ns/mip-1.0#">
  <track file="Dear Mr. President.mp3" puid="fa1fb49c-45c5-37bc-3be4-6c98b05798f4"></track>
  <track file="Dear Mr. President.flac" puid="fa1fb49c-45c5-37bc-3be4-6c98b05798f4"></track>
</genpuid>

As can be seen, the PUIDs are the same, i.e. semantically the files are identical, albeit the actual waveform differs.

FLAC file:

$ mutagen-inspect "10 - Dear Mr. President.flac"
-- 10 - Dear Mr. President.flac
- FLAC, 273.89 seconds, 44100 Hz (audio/x-flac)
musicip_data=TWlQAQACAAoAUAEBAQE+B4wNmZzsp4W7vX0vXsWNgpqNGfj7njg0lIol4xakk5SaELqpmI8Zt8+Z7IZsgAHe0LJc6dC+54Jz0QMDpRlm0kTHAYACgAEABAAVAQAAAAAAAAAEasLZ/wAAAAA=
analysis=MusicMagic DataAQEBAT4HjA2ZnOynhbu9fS9exY2Cmo0Z+PueODSUiiXjFqSTlJoQuqmYjxm3z5nshmyAAd7Qslzp0L7ngnPRAwOlGWbSRMcBgAKAAQ==
fingerprint=MusicMagic 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
musicip_puid=fa1fb49c-45c5-37bc-3be4-6c98b05798f4

MP3 file:

$ mutagen-inspect "10 - Dear Mr. President.mp3"
-- 10 - Dear Mr. President.mp3
- MPEG 1 layer 3, 186696 bps, 44100 Hz, 273.89 seconds (audio/mp3)
TXXX=MusicIP Data=TWlQAQABAAoAUAEBAQE+B4wNmZzsp4W7vX0vXsWNgpqNGfj7njg0lIol4xakk5SaELqpmI8Zt8+Z7IZsgAHe0LJc6dC+54Jz0QMDpRlm0kTHAYACgAE=
TXXX=MusicIP PUID=fa1fb49c-45c5-37bc-3be4-6c98b05798f4
TXXX=MusicMagic Data=AQEBAT4HjA2ZnOynhbu9fS9exY2Cmo0Z+PueODSUiiXjFqSTlJoQuqmYjxm3z5nshmyAAd7Qslzp0L7ngnPRAwOlGWbSRMcBgAKAAQ==
TLEN=274542
TXXX=MusicMagic Fingerprint=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

So, for FLAC the following tags are written: analysis, fingerprint, musicip_data, musicip_puid

for MP3 the following tags are written: audiolength, MusicIP Data, MusicIP PUID, MusicMagic Data, MusicMagic Fingerprint

[erland]

I don't think this is correct. I've analysed tens of thousands of files using genpuid only and they mix properly in headless server or MiP GUI.

You are correct, I think I was thinking of the libofa library and PUID retrieval which picard uses, as described in the link in your previous post this only works if someone has submitted the analysis using MusicIP or genpuid for the same track.

For the sake of clarity I've taken a single FLAC track, removed all tag data, generated an mp3 version of said track, and analyzed both using genpuid. What follows is the resulting tags added by genpuid.

...

So, for FLAC the following tags are written: analysis, fingerprint, musicip_data, musicip_puid

for MP3 the following tags are written: audiolength, MusicIP Data, MusicIP PUID, MusicMagic Data, MusicMagic Fingerprint

In my FLAC files, which have tags generated with MusicIP Archive Analysis, I only have ANALYSIS and FINGERPRINT. So this means that genpuid probably does the same thing as MusicIP archive analysis but on top of this it also writes the MusicIP Data and MusicIP PUID tags.

Do you know what libofa (picard) uses when it requests a PUID from musicdns.org ? Or maybe it always recalculates the fingerprint it uses when requesting a PUID ?

[Philip Meyer]

>TrackStat uses Musicbrainz Id tags, not PUID's. PUID's are just a way
>for the Musicbrainz tagging software, for example picard, to find the
>appropriate Musicbrainz Id more automatically. You need the Musicbrainz
>tag to get the extra functionality in TrackStat and I think neither
>MusicIP nor genpuid will write any Musicbrainz tags. To get Musicbrainz
>id's you need to use puddletag, picard or some other tool that support
>retrieval of Musicbrainz tags.
>
Thanks. That makes more sense now. Still seems a bit arbitrary to me, in that PUIDs cannot be unique ID's if they are created from n minutes of audio, and having processed my library, I have duplicate PUIDs. However, they are still useful - it has identified some actual duplicates that mean I can trim my music collection down a bit. As long as they will not be used for uniquely identifying content, I'm happy ;-)

>I believe PUID's are almost but not completely unique. Musicbrainz
>switched from TRM to PUID in their picard tagger because PUID's were
>more unique.
>

>> I seem to have ~5000 tracks that do not have unique PUIDs.
>>
>Sounds like a lot, I've got a feeling something is wrong if this really
>is different songs.
>
>Just out of interest, are they very similar or are we talking about
>completely different tracks with a lot of silence in the beginning or
>some other kind of similarity ?
>
That was about 2500 tracks that were duplicated to achieve ~5000 hits. Some of these really are duplicate music content, usually where I have ripped a CD single and also the album, and thus the title track is a duplicate, or where I have ripped compilation albums. This is useful for finding duplicate songs, but often the lengths are different - different mixes, maybe remastered versions, etc; eg. on some compilation albums, tracks are cross-faded into one another, so I can't eliminate the duplicates, and need to treat them as a different unique version of the song. I may even rate the song versions differently.

A lot of hits were very similar, except for a little extra silence padding on the end of tracks, etc.

>Are we talking about cuesheets ?
I don't use cuesheets very much, as songs defined by cuesheets are not mixable, and other software doesn't support cuesheets.

Phil

[Philip Meyer]

>There's a clear explanation of the differences between analysis and
>fingerprinting '*here*'
>(http://musicbrainz.org/doc/HowPUIDsW...Fingerprinting).
>
I found that last night and had a read. It's still not clear to me exactly what happens though.

i.e. I gather that Audio Fingerprinting is done on up to 2 mins of audio content, and from the fingerprint it can locate a PUID from the server. I see that PUIDs are created from performing Music Analysis on up to 10 mins of audio content, but this is apparently only done on the MusicDNS servers. The MusicDNS servers can't access the actual music data, so is that web page reporting incorrect facts?

>> I seem to have ~5000 tracks that do not have unique PUIDs.
>This means the MusicDNS server has not yet generated a PUID for the
>tracks in question.
>
In MusicIP GUI, I can see that the vast majority of songs in my music library have PUIDs. I have about 180 that don't have PUID's (30 unanalyzable, some analysed by name only, and the others analysed some time ago but still don't have a PUID).

Out of the ones that have got PUIDs, there were ~5000 that were listed via running the "List Duplicates" report (generates a playlist). i.e. different source files in the library that have the same PUID.

Phil

[Philip Meyer]

>Based on the names, it sounds like MUSICMAGIC FINGERPRINT is the
>fingerprint data needed to get a PUID and MUSICMAGIC DATA is the
>analysis data which MusicIP needs to be able to support smart mixes. I
>know MusicIP differs between the data needed for PUID detection and the
>data needed for music analysis. I've got the impression that the music
>analysis (which is required for smart mixes) is only possible through
>MusicIP and not through genpuid.
>
>How does the MUSICIP DATA tag contents look like in comparison to
>MUSICMAGIC FINGERPRINT AND MUSICMAGIC DATA ?
>
>From a song chosen at random from my library, I have id3 tags containing:

MUSICIP DATA = TWlQAQAGAAoAUAEBAGoGDp5swviNNoFauDIcjoABllCAAiIfmv szfoFJwBaAAYABBXWsM4wOxcCAAYLmgAG2OimdB929g4ABsigA egXjgAGqzoACgAEABAAVAQAAAAAAAAQAX0+1/wAAAAAAAgAhAQAAAAAAAAAAX2h2DQAAAAAAAAAAAAAAAAAAAAA ABwATAQADABIHEQATDgcARAWvAAYADwEAAgCqAOcArACHAAUAQ QEAAAAHAABzIwAAAH8AATSWAKMAAAABt3kA/wAAAATuEQB+AAAABd6SAFgAAAAGEAkA4gAAAAghAwC1AAA=

MUSICMAGIC DATA = AgEAagYOnmzC+I02gVq4MhyOgAGWUIACIh+a+zN+gUnAFoABgA EFdawzjA7FwIABguaAAbY6KZ0H3b2DgAGyKAB6BeOAAarOgAKA Aa8T1vKc2IF50iY6c739i5qAT/dmk3I6pIDzwBau0YIq1PSs7J2qu/SAZIQxgAGz3vAb6/a+8YABv+0BbOPbiyqjXIBmgAE=

i.e. one was created by MusicIP, and one created by genpuid. I assume that MusicIP will use only the one it generated, and won't use the one generated by genpuid.

MUSICMAGIC FINGERPRINT = AQKXHT5CISvTQU4iiiN9ELghVgK5A5kPrAUlCuoJeghLAYEPow FsBHILiAWMB00EXwGRALAAyQOwARADOQHqAXQAqgDaAFQBdQFf ASsBYABd/wjlw9jPydZcW+VR+q7oPB6P/8oA/wGuAOYD1AHi8x0ASg+uABT8Cv6oABT/jACX/9v/+f/6/2v/0/+W/0//7v/8//j/7f/aABb/ogAIAAEAsQx1BoIFpDJW9jkQgQMtk8D+r/6L8/z+ffnj+979V/8k3en/WgEu9zn8f/pK/E//7f/M/8MCdwA0AKoAkQD2AEwAfwAdAM4AtABhANEANP8p3uAdqy51FRT fDZz2C5r7cP9uAIMLG/dk9VoFgw7//78EFf+U/RMCaf/YAmv/ZP3i/2H/iPqY/c75p/nN/eb/Zf7u/1T9Wf66/Ij+0f+iATP7F6SqMkERsSpt/Y0zpv73AZYAQAcQ//8DhAhAE68AswDPAS4HSghjAc4GUAMDAFUAPgBVAMX/7gBT/0sAegA4AC//+QAJAIT/pABkACf/dAXPCqbZTQa1Zy3KNOhO8wH/2f8UACr9i/71/TTwL/+m+1P/evsg+Mr+SPvV/un+1v9t/7f+fv7K/I/8Zf9y/8v/5v+o/vv/2/3UACcAAADSHCLnBENfByX+LfwOsLUSqv8X/EfguP5X+hXzFdh+/oHwXP7S99bpb/v49Xf9Lv84/3n/bgCy/7z/7f/qAGgACQBG//sAWQAU/50ALQAYLCspHQ==


For a flac file:
ANALYSIS = MusicMagic DataAgEAAFhYtcWbic4+hkSh+Fx5gAGBdYpk62ChZkPTg9D6Yo ABgAH6RLWsjTOvFY0UgFSAAQEgjW8UgL9vh7CtRe159LXQZM5v gAKAAaKZno+sN4JI+bM+boUEg1KHWOy8n9RAZYvp9sSUdYp9+c i0LLDtzdmgPIOJiAHr4JglBHm6cYVKt/T0AepPr3TBwIDwgEY=

FINGERPRINT = MusicMagic FingerprintAQdaKMZBMD/nHyYU8xJHEBwSgheVE08TrAzmDr8L+wuaDMcL3AeqCNkQ4xR2E UgMew1BCOIGHATeBHoEIwP2A1ACqQJzAc8B0AHNAe8BQwFG/7zlR2CPuZESOAU095cAoQT86i/3egaJAJ7+AP11+qP5XALH/Un5JPsd+WH4av5c+wH8bP1B/n/+Iv5q/p7+rf8s/r//Kv9P/xf/Kf+N/44HJid51xuvXB5wCwUH1hLiD4USbhr/FQ4SWxDkEAcGiAsVB6wG3gtSFoUfWQ5hDEUHEwiNBRoEqANGA+ MEHwOuAqoCRQHNAYYBxwFcANsBGfNpmlX4awarAuICZgL++ZUE 6hHQBAQDTw7eDPgH2BDkD8sN8Qi6GNcWWRpvJEILYRdHDs0Elg K9AwYC3gMbAqUB8gEwAKAA6ACyAgkBEQDWAnUEW+4TCRsLzRRU CUT6JyEgoo/fIxCv/NAQIOuhEBAnfBgx/JnpKgAQB28J5gn4DTsDCQSDAPYCQQGCAUYBsACx/78AfQDPAH0AsQAkACz+UuFb5fUOKlAOJLUTzBFXFo0H+/5ZBtEHKAP9ARTnhdeEDsUA1vav8x3eBOMp9Cfs3fPy+X79NP1j/pD9fv5P/0j+8P9m/xj/D/7d/6n/qQDYCCH/zAOFQdzV0NMK2U4D+BLL72bJUAQKAxj6nQ4UKegFpf79+YD4sh WZ8S7z1/5I+n4DnP/c/lgAFwEUAjEAuAD8AHcAxwBpAS//7gAhNx8dGA==

MUSICIP_DATA = TWlQAQAGAAoAUAEBAABYWLXFm4nOPoZEofhceYABgXWKZOtgoW ZD04PQ+mKAAYAB+kS1rI0zrxWNFIBUgAEBII1vFIC/b4ewrUXtefS10GTOb4ACgAEABAAVAQAAAAAAAAACRenp/wAAAAAAAgAhAQAAAAAAAAAARWzt+QAAAAAAAAAAAAAAAAAAAAA ABwAbAQAFAAQa3AAHAjsADQCfABYFIwA5AvEABgAPAQACAKoAb QCsAHsABQTZAQAAAJoAAAANAAARAAAAACW/AAAAAAAALgAAAB8AAAA3oLUaAAAAAGkAAAAsAAAAewB+AAAAAA EjAAAAHwAAASwAfgAAAAABgwB+AAAAAAGIbwAAAAAAAaJBAAAA AAABsQAAABkAAAHMSQAAAAAAAexpAAAAAAACQgAAAEYAAAKQPw AAAAAAAu6BfgAAAAADDkcAAAAAAAMZALUAAAAAAz8AABEAAAAD UVIAAAAAAANUWwAjAAAAA1wAAABSAAADzl4AABkAAAPPAAARAA AABBZQAAAAAAAETbUAGgAAAARxAAAAZgAABHdpAAAAAAAEe2wA EQAAAATjAAAAGQAABTpKAAAAAAAFS08AAAAAAAWMAAAALAAABf 0AtQAAAAAF/1AAAB8AAAYd3gAaAAAABm/AtSMAAAAGqgAAEQAAAAbURQAAAAAABuAAAAAzAAAG4UIAAAAAA AdCAAARAAAAB2FTAAAmAAAH60EAAAAAAAgDAAAAGQAACG0AAAA ZAAAI1ZYAGiYAAAlxngAAAAAAClmkAAAAAAAKcQAAEQAAAApyQ gAAAAAACvpgAAAAAAALHwAAABkAAAsn+7U0MwAACzUAxv+SAAA LZQAAABkAAAuMrn4AOQAAC5xhAAAAAAALnkEAAAAAAAuqQQAAA AAAC8aHAAAAAAAMGv8AAIUAAAxkAAAALAAADQpSAAAAAAANWEo AAAAAAA2AQQAaAAAADZI
AAAAsAAAN2EEAAAAAAA6qdAAAAAAADs1uABEAAAAO0gAAACYAA A7XTQAAAAAADysAfgAAAAAPVKQAETMAAA+jAIoR/wAAD7VEAAAfAAAP00QAAAAAABBU+AAAZgAAEfhKAAAAAAAUHwA AGgAAABQgAAAjAAAAFHVNAAAAAAAVpAAAAPIAACdxAOIrAAAAK GMAAIMAAAAp9gAAEQAAAC02AAAAJgAALbgAngAAAAAwMQDGe5k AADItAH4AAAAAOScAngAAAAA74wAAAB8AAECtAH4AAAAAQSsAn u0AAABNTwDiAAAAAGLpAH4AAAAAZMYAfgAAAABohwAAADMAAHb 1AAAAJgAAiMMAxgAAAACOTwCeAAAAAJQXAJ4AAAAAnvAAngAAA ACyiwB+AAAAALYBAAAaAAAAv8EAigAAAADB2gCeAAAAAM56AH4 AAAAA2DoAngAAAADbrgAAgwAAARFAAAAAMwABGB4A4gAAAAFYz wCeAAAAAXjnAAAaAAABkNMAAAAzAAGTBQDiAAAAAavGAABXAAA BrHUAABEAAAG6eQDiAAAAAcT4AJ4AAAAB3m8AAEYAAAKGMgAAI wAAAv3tAAARAAADzawAngAAAAPkEQDGAAAABDggAOIAAAAEsuY AAAAmAATIRACeAAAABQRZAJ4AAAAFDfQAngAAAAUWtQAAEQAAB f8GAJ4AAAAGI3AAABEAAAZI4wAAADMACbAuAAAAmQAJxlIAABo AAAsisgAAVwAAC/GNAIoAAAAZmwAAngAAAFD5cwCeAAAAWErtAJ4AAABamIwAngAA AFyUJgAAAB8AXKfJAAAAMwBcrvcAigAAAFy6BAAAADMAXNOWAA AAJgBc5Z8AAAAfAGX8cAAAAB8AaeEOAAAAOQBr39UAAAAfAHdt 1QAAACYAeI/wAAAAMw==

MUSICIP_PUID = 0076d706-fac3-bd55-073d-b9153a1421eb


Observations from above:
1. MUSICMAGIC DATA tags appear to be much shorter that MUSICIP DATA tags.
2. Flac tag values include "MusicMagic Fingerprint" or "MusicMagic Data" at the start of the value.


I found two different source files sharing the same PUID, that were the same song but different versions. "Good Times Bad Times" by Led Zeppelin. One song was an mp3 rip from Mothership (for my iPod), and another rip was FLAC from "Box Set 2". They have the same length, but different mastering. eg. the Mothership version has a track gain of -10.04, whereas the Box Set 2 track gain is -5.69 (sounds better to me). The DATA tags content are the same in both files, and the FINGERPRINT tag value content is the same.

>You mean that you don't have PUID tags and want them to be written ?
>Why ?
>
I have PUIDs within my m3lib file (I have added the PUID column to the MusicIP GUI display) - most songs have a PUID (99% of my library). However, these PUIDs haven't always been written to the music source files (archive analysis done before PUIDs were discovered from MusicDNS servers, or PUIDs created by genpuid, but weren't written to source files). It now appears that I can't run archive analysis again to force the PUIDs to be stored in the files.

The purpose is that if I move files around, MusicIP has to reperform analysis to determine that it is the same song. If it has the PUID in the source file, it should be a lot quicker.

>As I've understood, MusicIP doesn't require a PUID to do analysis, the
>PUID is just needed to speed up the analysis by retrieving it from
>www.musicdns.org instead of analyze the file again.
>
Ah! FINGERPRINT -> PUID -> DATA. So because I have a MUSICMAGIC (or MUSICIP) DATA tag, it doesn't need the PUID. The PUID could be discovered again from the FINGERPRINT (provided musicdns.org servers remain running).

Or, if the fingerprint is used to get a PUID and stored in a tag, then the MUSICMAGIC FINGERPRINT tag is then redundant?

Also, maybe the PUID is actually also within the DATA content?

>Maybe we are talking about the name change. Maybe MusicIP still uses
>the old MUSICMAGIC name while genpuid has been updated to use the new
>name MUSICIP ?
>Are you using the latest version of MusicIP ?
>
Yes, 1.9 beta 6, unregistered. Would have registered, but the company disappeared just when I got interested in the product!

So I'm unsure where

[Philip Meyer]

>So, for FLAC the following tags are written: analysis, fingerprint,
>musicip_data, musicip_puid
>
>for MP3 the following tags are written: audiolength, MusicIP Data,
>MusicIP PUID, MusicMagic Data, MusicMagic Fingerprint

>From your example, it looks similar to mine, in that you have two different analysis tags in a single file (analysis and musicip_data), with different content.

It intruiges me why there are two data/analysis tags with different content in the same file. Perhaps genpuid calculates the fingerprint, it looks up the PUID, and stores these tags, and then when they are scanned into the m3lib file via the MusicIP headless server, it also adds the musicip_data tag?

[Philip Meyer]

>In my FLAC files, which have tags generated with MusicIP Archive
>Analysis, I only have ANALYSIS and FINGERPRINT.
Do you have PUIDs in the m3lib file (can be seen within the MusicIP GUI if you add the column)?

I guess these are not being stored when running Archvie Analysis. I still don't know what tags are actually required and if others are redundant, but currently I have inconsistencies in what tags have been stored.

Some of these tags contain quite a bit of data, so storing per-file could add a lot of extra storage requirement and fragment the HD.

[audiomuze]

I gather that Audio Fingerprinting is done on up to 2 mins of audio content, and from the fingerprint it can locate a PUID from the server. I see that PUIDs are created from performing Music Analysis on up to 10 mins of audio content, but this is apparently only done on the MusicDNS servers. The MusicDNS servers can't access the actual music data, so is that web page reporting incorrect facts?

I think you've misunderstood. The analysis is done on the client (your) side and the PUID is generated on the server side. Analysis takes up to ten mins per track, not ten mins of audio content.

Out of the ones that have got PUIDs, there were ~5000 that were listed via running the "List Duplicates" report (generates a playlist). i.e. different source files in the library that have the same PUID.

Are they different versions of the same track?