Philip Meyer;542129 Wrote: > >this is very interesting to me for many reasons. is what you say > above > >true? i hope so, b/c its a step in the right direction. > > > I hope not, because it would be a step in the wrong direction; > unnecessary extra logic at scan time, that also enourages bad practice. > No other apps that I am aware of have this logic so it's introducing > yet another inconsistency; tag metadata sources are not consistent, so > some albums could be tagged with "Various", "Various Artists", "VA", > etc.
thats why i said it should NOT be hardcoded, that the user could exactly specify the criteria/string. a user should be able to specify any/all strings that would invoke the behavior. just leave it blank if you want. besides, fixing the occassional bad tagging from a tag source is expected in ALL tag fields. and it has nothing to do with the actual scanning process; it would be a post scan analysis, just as "merge various artists" is now, but if it were left blank, as you would leave it, it wouldn't happen (be invoked) at all! its up to you! SBS uses the simplistic 'any TPE1 mismatch is a comp' behavior, (and itunes appears to as well, although maybe it gets that info from gracenote?) but only itunes is a "major app" of those two. you are all for that behavior though, so really saying "no app does this or that" isn't in and of itself a valid argument. and besides, itunes get a CD like Sinatra's Duets wrong, it gives it a comp tag. its not a comp, its clearly a Sinatra CD which simply has guests, and should be in the Sinatra section with all his other CDs which should be done via an AA tag. winamp btw, gets it right. Philip Meyer;542129 Wrote: > > >first, its important to stipulate that all major marketshare apps and > >tagging sources (except freedb) will fill in TPE2 (ripped or > >autotagged) as meaning AlbumArtist info, comp or not. that means > >almost all audio file users will come to slim with that data. (they > >also do it for other formats) > > > It's not important to stipulate, especially when you are wrong. > > I think it's fairly safe to say that iTunes is one of the major > marketshare apps. I just stuck a regular CD in my drive, fetched track > details from gracenote, and the Album Artist was blank. The same > happens with my ripper of choice, EAC, which is quite popular. first of all, itunes uses TPE2 for albumartist. thats a fact. has since ver 7. http://www.macosxhints.com/article.php?story=200609140956019 http://www.id3.org/iTunes i don't argue over whether it should or shouldn't, i just say it does. the user can enter that info, or have it recognized by itunes when present and imported from other original sources than itunes. (btw, idk what apple store tracks have, i've never downloaded a song from apple) now, i should add that 'ihate' itunes and don't use it much, so i'm no expert on it, so i can't say that i know for certain how it implements (or not) the data gracenote has. but what i can say, is that it is MOOT as to whether itunes fills in TPE2 info or not on its own rips and autotags, for the simple reason that it already uses explicit COMP tags! i have never said that SBS should not respect a COMP=1 tag, (at least insofar as simple classification of comp or not of a given album to SBS). in other words, it doesn't matter if itunes doesn't use the gracenote albumartist info on its own rips/autotags (if thats the case), when it 1. respects the TPE2 field as meaning albumartist, and 2. uses the explicit comp tag anyway, as that supercedes everything else. itunes even goes so far as to suggest that albums of differing names without comp tags can be grouped together by using the TPE2 field alone! again, i don't know exactly what they mean by that, but its in their docs. regardless, the main point is that most users will come to SBS with EITHER TPE2 info, (major apps like WMP, winamp, etc) OR an explicit comp tag (itunes). and in the case of itunes, the user can add AA tags, and have itunes respect them. but either way, my argument stands. why? b/c the amount of users coming BOTH WITHOUT AA tags AND/OR an explicit comp tag, is going to be the vast minority of the marketplace, as i just demonstrated. Philip Meyer;542129 Wrote: > Interestingly, dbPowerAmp does (sometimes?) retrieve the > irrelevant/unnecessary Album Artist name for regular albums (with all > songs by the same artist), but doesn't for compilations, where perhaps > it would be more useful! I stuck in a compilation disk, and tried > dbPowerAmp - this identified the album as a compilation (sets the > compilation tag), and Album Artist was blank. i don't use dbpoweramp either, but i have used winamp and WMP, and both of them will fill in TPE2 info if gracenote/AMG (respectively) have it, on rips and autotags. they do it automatically, just as itunes automatically sets a comp tag. that means most users coming to SBS will have one or the other. and btw, EAC is by no means "popular" in terms of marketshare, even though it is what i personally use as well. and a lot of people hack it to use gracenote which has the TPE2 info available on i'd say about 90% of the tracks in my exp. Philip Meyer;542129 Wrote: > And, as I mentioned above, metadata content is inconsistent across > metadata sources, and even within the same metadata source engine > there's a lot of inconsistencies. so? no one says errant tags shouldn't be fixed. thats not the argument i'm making. Philip Meyer;542129 Wrote: > The only way to achieve a good music library, is to tweak tags. You > can't assume that an app will have all these special logic checks to > cope with various deficiencies in metadata sources. you mean like TPE1 mismatch checking? Philip Meyer;542129 Wrote: > > >should not just blanketly design the software on the > >assumption that the mere presence of AA info inherently means > something > >beyond what it is meant to do. > > > In what way is AA being treated as something beyond what it is meant to > do? > It's defining the artist that the album belongs to, grouping songs > together onto a single album. b/c, up until 7.6 apparently, SBS has taken the mere presence of ANY AA tag to mean that such and such is NOT a comp. likewise, it has taken the absence of an AA tag, combined with a TPE1 mismatch, to mean this IS a comp. neither one of those assumptions is valid! and moreover, since most users will come to SBS with either an explicit Comp tag (itunes) OR an AA tag (WMP, winamp), its fighting the tide to begin with. it just isn't smart. Philip Meyer;542129 Wrote: > In my opinion, the purpose of an artist tag is only to define the artist > name. It should not assume that certain names means that an album is a > compilation or not. ok, so leave my proposed setting blank! u get your way, i get mine, hooray for options! and can you not see the obvious irony here? you say "the purpose of an AA tag is just to specify an artist name, it should not assume that certain names mean comp or not." but yet you have no problem, and in fact endorse the notion that by doing so, by defining that artist name, SBS should ALSO take that to mean something is NOT a comp. come on. Philip Meyer;542129 Wrote: > > >http://bugs.slimdevices.com/show_bug.cgi?id=8324 > > > It's not a bug, and to turn off such logic is dangerous. indeed. i envision a chernobyl like exp. danger will robinson, danger! /sarcasm. sorry about that, but honestly, can we please not overstate such things? its not dangerous. Philip Meyer;542129 Wrote: > > >b/c for many people, being able to turn off VA artist mismatch > >detection logic would help speed up the scanning process. > > > Potentially it could speed up scanning process, but for many users they > could end up with a proliferation of albums, because songs would not be > grouped correctly, which would slow down the app when browsing. For > many users, gathering sub-optimal tagged songs is a life-saver, as the > majority of albums with differing artists *are* compilations. fine, then turn it on! leave it on! thats why its an OPTION. leave it on by default, fine by me. but don't begrudge me and others the CHOICE to NOT use it. every little bit helps on gear like the non-intel infrant units, which yes, i use at the station. and whatever SBS does with it off, may in fact BE DESIRED. and if not, it can BE CORRECTED. but we won't know until we can try it. Philip Meyer;542129 Wrote: > For a smaller number of cases where there are guest artists (many > metadata sources don't return guest artists anyway), they appear as > compilations, but that's better than the songs not appearing within the > album (eg. ending up with several albums with sub-sets of the tracks on > each album). gracenote very often does list guest artists in TPE1, AMG might, i don't use it near as much. and again i make the same points to you: 1. the unknown behavior (proliferation?) you don't like might be desired by others, and could also be further refined 2. most users will have AA tags or comp tags 3. those that don't can simply leave the VA detect logic on if they want to i just can not understand the beef you have with it given all that. Philip Meyer;542129 Wrote: > > >what would or should happen if its turned off for those users without > >AA info? well, implement it, and lets see what happens, and then go > >from there. > I see you have this all figured out. > > Phil am i supposed to know what SBS would do? do you? does anyone? there's a lot of implemented code people don't fully grasp. don't be snide for no reason. -- MrSinatra www.lion-radio.org using: sb2 & sbc (my home) / sbr (parent's home) - sbs 7.5b - win xp pro sp3 ie8 - p4(ht) 3.2ghz / 2gig ram - 1tb wd usb2 raid1 - d-link dir-655 - 40k+ mp3 ------------------------------------------------------------------------ MrSinatra's Profile: http://forums.slimdevices.com/member.php?userid=2336 View this thread: http://forums.slimdevices.com/showthread.php?t=77928 _______________________________________________ beta mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/beta
