I posted this months ago and in the meantime I have not been able to come up with a workable solution for keeping the songs from my mp3 collection updated on the Rio player.

Mostly the problems revolve around the following:

* Rio throws away the filenames on upload

* Rio uses mp3 tags to name the songs (instead of the filename which allows more characters and more precise matching)

* The command line tool supposedley will allow you re-upload the same directory, while grabbing only new files but that does not work.

* Command line tool does not report back the fid to be used until the actual sync is happening,.. making it impossible to script the fid easily back into my database for building sync lists (lists of files to upload, or remove on the Rio)

All of this makes me seriously wonder how anyone else keeps a large collection synced up with their car player. The reply posts I got the last time I posed this question were not very helpful overall,.. although I really appreciate everyone taking a stab at it.

Matter of fact the most promise post turned out not help at all. It was suggested that I simply use the following command:

upload -p -r /mp3/*

And that would skip upload on all the files already on the Rio. But the (undocumented) -p switch doesnt seem to do anything on my version of the command line tools (emptool version v1.03), as far as I can tell. But I should at least be grateful that it doesn't work as badly as the Windows tool which not only uploads the same files again, but creates duplicates of the same name on the Rio. At least emptool doesn't create duplicate files, it just wastes time uploading the same files again.

I've also saw that people were suggesting that I somehow use the CSV download feature to drive my syncronization scripts. Well unless somebody knows something I don't, the CSV database is useless to me. It seems that the CSV database is using the ID3 tag of the song, instead of the filename. Well, many songs have the same ID3 tag. This is because ID3 chops off the song titles (and artists sometimes too) because of character limitations. Songs like:

Five For Fighting - 04 - Superman (It's Not Easy, Bob Clearmountain Mix).mp3

.. end up getting chopped off and the end result is the regular mix and clearmountain mix get the same ID3 tag. Not exactly sure that example is 100% factual but you get the idea.

I also saw posts saying that I should use ls -R in emptool to produce a list to drive my sync scripts but this is just as useless as the CSV file, for the same reason. The artists and songs get chopped and are impossible to reverse match using a simple fixed length chop.

Hmm, i guess i could rename all of my songs to make sure that the first 30 characters of Artist and Title are unique, then a chop and reverse match might work,.. but what a pain in the ass that would be. Not to mention my collection would be full of mangled artist and titles simply to suit Rio. Then again I guess I could achieve the same thing with symlinking the entire collection using the mangled unique names. You can see how these "hacks" would get unweidly very quickly,.. script to generate the mangled symlinks, etc.

I even saw one post that said why dont you just use the Emplode tool and fish around in Windows explorer dragging and dropping the new files onto Emplode? Well I have 90GB worth of mp3s, so even though my directories are 100% organized and strict, fishing around through 16,000+ mp3s is no fun. And certainly out of the question. Uploading a few files becomes an hour long ordeal.

This isn't a bitch session, but I feel it's important for the Rio developers to know that out in the field, in real world, the management tools absoultely SUCK. I have had no problem coming up with sync scripts for keeping home, work and laptop mp3 jukeboxes in sync. But this Rio thing still poses major hurdles, due to the seemingly bizarre choices of the developers when they built the Rio database management system.

And although the last go-round yielded no helpful advice, I'm still listening so I'd love to hear from anyone else trying to manage a huge collection on the Rio. What are you guys doing out there? I'm getting ready to go on a rip/encode marathon, adding another 10,000 or so mp3's to my collection so I am really hoping someone out there can provide some insight on hy they are making it happen on their Rio.

Thanks, Aaron Newsome

[stop the technology madness]

correction, emptool DOES create duplicate entries with the same name,.. even when using the -p switch for the upload command.

[stop the technology madness]

Man, that's tough. There's a lot of requirements there.

Something to consider. One of your primary complaints is the 30-character ID3 limit, and the fact that they don't match the file name.

What about using MP3 Tag Studio to bulk re-write all your tags as ID3V2 tags to match your long file names? Then the file names would all match what was in the Tags, which is what ends up in the Empeg's database.

Anyone have any other tips?

___________
Tony Fabris

Anyone know the character limit of ID3V2 tags? If I remember correctly a good number of my files get chopped using ID3V2 tags as well, which is why all three of my jukebox systems use strict filename/directory structure to show the song/album/artist info and ignore the tag all together. This seems to work much better and actually does what the tag was supposed to do better than the tag does.

Some of my filenames can get very, very long. 250+ characters, like:

Bring On The Pain (Monday Night Madness REMIX, featuring Billy Blanks and Morgan Stanley).mp3

So I'll look into this Tag Studio thing,.. but probably will not end up using it simply because it is a Windows based and those tools really don't lend themselves to automation very well.

Since my current tagging script is Linux/Perl based,.. I guess I will probably have to track down the ID3v2 mp3 tagging module and give that a shot. Last I checked though, it had warnings all over it that said it might corrupt the mp3 file as it tried to apply the tag. I don't need to explain why corrupting 16,000+ mp3 files would leave me feeling a bit disapointed.

[stop the technology madness]

The character limit of ID3V2 tags is huge compared to V1. I blow the buffer on most of my mp3s in V1, but not a single one in V2. Give it a shot.

Why are you uploading the same file several times? What I mean is, why don't you have a separate directory for uploaded mp3s and one for mp3s that haven't been uploaded to the RIO. That is a simple way of keeping it sync'ed with your collection, unless I totally misunderstand what you are trying to do.

Anyone know the character limit of ID3V2 tags?

What's more important is what the limit is for the database entries in Emplode. I believe each field is more limited than even a V2 tag. Remember that the empeg does not use the tags or the file names at all. It only uses its internal database. So all the important limitations are the ones imposed by the database.

For instance, maybe you might have a V2 tag with a 300-character field, but Emplode will truncate that to 255 characters upon input into its database. I think 255 is the limit, not certain.

Remember, though that this is per field, not for the entire tag. Your example of a long song title was only 90 characters long. The artist and album titles have separate limitations.

I wonder, are some of your problems due to the fact that you're trying to cram all the information into the single Title field instead of splitting up the artist/album/year data into their proper fields?

So I'll look into this Tag Studio thing,.. but probably will not end up using it simply because it is a Windows based and those tools really don't lend themselves to automation very well.

MP3 Tag Studio does a decent job of automating mass rewrites, assuming you know what you're doing with it. And you really only have to do it once. Run it on your big collection and be done. Then just make sure that any new files get added with the proper tag data.

The only scary part is the sheer size of your collection. One mistake in Tag Studio and you might ruin the tags on your entire collection. So I recommend that you practice with it a bit on a small subset copy of your collection, and check its results carefully, until you understand how it works.

The Linux/Perl option sounds good too, if you're comfortable with that.


___________
Tony Fabris

My Linux machine has vented its holy smoke and I am currently stuck with Windows, so I am talking from memory:

Uploads: I kept a list of uploaded files (a slightly massaged output of recursive ls emptool command) and created emptool command scripts (by a Perl program) that 'automanually' created each necessary playlist and uploaded songs individually. It took quite some time to evolve this from album-by-album to more automatic version. (And I am affraid I lost them now... )

Tags: grip I used for ripping and encoding used a V1-only tagger, though I believe the current version is V2 capable. I think I used a demo tagging utility that comes with id3lib to do ID3V2 tagging (from a Perl script) and it worked OK. Alternately (IIRC), one can use a series of emptool set directives (from the upload script) that identify songs by the beginning of the title (once in the correct playlist, of course), but then you have problems with different versions of same song etc. Much better to prepare ID3V2 tags correctly.

MP3TagStudio Tony is swearing on is very good and lends itself to a certain degree of automation (not the kind Linux people are used to, though: no scripting support, just set of rather clever tags-from-dir/filename and vice versa modes). If you have VMWare or a Windows machine with Samba mounts, it can be very handy fot V2 tagging if your dirname/filename structure is systematic and informative enough.

Finally, it should be understood that emptool has never been advertised as nothing but a stop-gap measure. Did you give Jemplode (Java semiclone of emplode that uses emptool low-level code - the only thing Mac users have) a try?

Good luck!

Dragi "Bonzi" Raos
Zagreb, Croatia
Q#5196, MkII#80000376, 18GB green

I did find a perl module to do the tagging so I will write the script to tag this weekend.

I do have vmware and samba mounts to my repository of files (on a linux server).

I downloaded the MP3TagStudio and gave it a try. The Interface was pretty slick and it apeared to be able to do what I needed, without needing to write any scripts at all. However when I tried it simply didnt work. It reported success in tagging the files but it did not tag them. My file structure is like this

/mp3/d/Depeche Mode/Ultra/Depeche Mode - 01 - Barrel Of A Gun.mp3

The tag studio seemed to follow the path just fine with its format specifier, but alas, no tag was written to the file.

Second problem is (and a minor one at that), is that about 50 of my files have a question mark in them, which is not allowed in the Windows filesystems, or over a samba mount so these files get badly mangled names. Like the Jill Scott CD for example,.. the Title of her CD is "Who Is Jill Scott?", so every one of those tracks get mangled.

As far a Jemplode,.. I actually saw that BEFORE I bought my empeg but after I finally got my empeg (Rio actually), I could not find a version of jemplode that was being actively maintained and the version I did find was buggy and crashed alot.

--Aaron

[stop the technology madness]

You might want to try Tag&Rename, which you can find here.

I've had better luck with it than with mp3tagstudio, and it's everything I need for tagging.

Riocar 10gig (for now) S/N : 10101747 amber

In that case, your own Perl tagger is probably the way to go.

FWIW, I didn't have any problems with MP3TagStudio. The problems with MSW-illegal characters in filenames should be somewhere in Samba, should'n it. Isn't it possible to configure Samba to simply replace those chars with something benign (I know, not ideal, but better than having filenames totally mangeld). Of course, this is irrelevant if your Perl tagging module works OK.

I didn't know about the Jemplode status. Pitty. Anybody knows more?

Dragi "Bonzi" Raos
Zagreb, Croatia
Q#5196, MkII#80000376, 18GB green

There is no real character limit as I remember for ID3v2; the empeg has a 254 character per field limit though (ie, title can't be >254, artist can't be >254, etc).

Using ID3v2 there should be no truncation at all.

Hugo

Are you sure the tags weren't written? If the mp3 files previously had v1 tags and you didn't delete them, then they are probably still there and will often show up first. The best way to save tag confusion is often to get rid of all the v1 tags first, then you will be able to see if the v2 tags are actually there easier.

(list 6284, Mk1 S/N 00299 4GB blue [for sale]. Mk2 S/N 080000094 26GB blue)

Yes I am sure there was no tag after I ran mp3tag studio,.. or better yet I'll clarify, there was a v2 tag present and all the fields were blank. This is what I did.

* Played a file with mpg123 and the -v flag (shows tags). Tag shows up fine.

* Use mp3tagstudio to remove the tag (success)

* Play the file again with mpg123 -v ,.. tag is completley gone as expected.

* Use mp3tagstudio to tag the file according to my file pattern,.. success reported.

* Play the file again looking for the tag installed by mp3tagstudio, again with mpg123 -v. Tag is printed but all fields are blank.

So to clarify, it does install the tag but it just puts all the fields blank.

--Aaron Newsome


[stop the technology madness]

Hi.

* Use mp3tagstudio to tag the file according to my file pattern,.. success reported.

* Play the file again looking for the tag installed by mp3tagstudio, again with mpg123 -v. Tag is printed but all fields are blank.

So to clarify, it does install the tag but it just puts all the fields blank.

Sounds as if the pattern you specified didn't actually match the pathname. At least that is what happened to me when using mp3tagstudio some while ago. I now use mpTagger which stops if a pattern does not match and asks for a new pattern. However mptagger only works for a single directory on each pass, so it is certainly not the tool you look for with you huge collection.

cu,
sven

proud MkII owner (12GB blue/green/smoked, was #080000113 is #090001010)

I usually check tags using Browse tags option of MP3TagStudio. If there is nothing there, it always turned out it was my mistake when specifying patterns

Dragi "Bonzi" Raos
Zagreb, Croatia
Q#5196, MkII#80000376, 18GB green

IIRC, the only reason the player uses the ID3 tag is if it is present in the MP3. If it's not in the file, it takes the filename and uses that. This means you could bulk strip the 3v1 tag for your entire collection; I use MP3 Tag Studio. As Tony recommends elsewhere, it's a good tool and worth paying for. The author is receptive to suggestions and your ideas may be taken on board for a future version. At the moment it is not script controlled and requires manual work, however.

One of the few remaining Mk1 owners... #00015