So I've been trying to figure out how to use mp3tofid for a while now, as I can't seem to get emplode or jemplode to make it past 100GB of tunes on my SSD without all kinds of sigkill and segfault errors. It's infuriating. I'm pretty close to giving up.

I've got mp3tofid compiled and installed on my MacBook as well as a Debian Jessie VM, and both seem to work fine. They both seem to create the data files as well as the symlinks to the mp3s themselves. What I don't understand is how/where to put the music that has been symlinked?

Does anybody have sort of a "manual" for how to set up and actually use it? This may be my only hope.

The entire collection of FID files must get copied into "/drive0/fids" on the player:

What is a FID?

Getting them there.... Not sure what the correct procedure would be for you. If you happen to be able to hook up the drive to your PC directly, that would be the fastest I think.

Once that's done, then you need to get the player to rebuild the database:

how to rebuild databases

Right. But mp3tofid doesn't convert the .mp3 to a fid. It merely creates a symlink to that .mp3, which obviously doesn't do any good if it's not on /drive0 somewhere. That's my question. What do I do with the mp3s?

I've also noticed that the symlinks created refer to my local directory, where mp3tofid finds the mp3s, such as /Users/Trevor/Desktop/... which doesn't correspond to the empeg drive in any way. Is there some certain way that I should be setting up my desktop environment for mp3tofid conversions?

I have been hooking the drive directly up to the Jessie VM and have been able to sustain very high (22mb/s) write rates to the drive, so definitely the fastest.

I've just been trying it with a few folders and very small transfers for proof of concept, but unfortunately still nothing.

Hm. Not sure, I haven't ever done that.

But I do know that the MP3 files are just more files like the FID index files. The index file is the one ending in the number "1" and the actual data is the file ending in the number "0". For instance, on my player, the actual MP3 file of Tom Sawyer is "/drive0/fids/bf90" and the song's index/tag data is stored in "/drive0/fids/bf91".

The files that end in "1" are just ascii data containing the song or playlist information. For instance, on my player, /drive0/fids/bf91 looks like this:

duration=273946
bitrate=fs256
genre=Progressive Rock
file_id=01
type=tune
length=8767903
codec=mp3
artist=Rush
offset=1627
tracknr=1
ctime=1053562286
year=1981
source=Moving Pictures
rid=70d3fc79d7338bd1c897ebeae124d987
samplerate=44100
title=Tom Sawyer

(from what is a fid?)

I would have figured that the purpose of MP3toFID was to do that particular processing on everything for you. I guess not?

I never got around to using mp3tofid myself, however...

If I remember right the symlinks were made on the computer to avoid duplicating the collection. Once those links are made, you ignore the source mp3 files, and use rsync aimed at the symlinks to the empeg.

There may be specific flags needed on rsync to ensure it treats the symlinks as full files instead of trying to replicate the symlinks to the empeg.

Aha, there is more details on the rsync steps Pim took with his personal setup in the readme. His project was migrated to GitHub by other members of the community here a few years ago. Give this a lookover and see if it helps: https://github.com/empeg/mp3tofid/blob/master/README.html

Ah, hah! I suppose I didn't realize that rsync was a required part of the process. So rsync sees the symlink and then transfers the actual .mp3 targeted by the simlink. Does it also rename the file to xx0? I'll have to set up rsync and try again.

I just ran mp3tofid on a small folder again and I'm looking at the xx1 files. There is a 'loadfrom' entry which points to the local directory on my MacBook. As this is the index file and NOT the actual .mp3data file, does the empeg simply ignore this? What is the purpose of this entry, if this is the file read by the empeg? Here's an example:

artist=M83
bitrate=vs196
codec=mp3
ctime=1444155365
duration=303177
file_id=8/15
genre=Alternative & Punk
length=7422614
loadfrom=./M83/Before The Dawn Heals Us/08 Teen Angst.mp3
offset=2229
rid=df49f2f94bdb8c6bee09e294072b7c46
samplerate=44100
source=Before The Dawn Heals Us
title=Teen Angst
tracknr=8/15
trailer=626
type=tune
year=2005

I've also noticed that it's used the shorthand directory which I used in the mp3tofid command, i.e. ./M83 vs /Users/Trevor/Desktop/M83. I suppose I should get in the habit of using full directories when running it? Would this affect rsync? This question seems to fall in line with the purpose of the 'loadfrom' entry. I checked, and fids created by emplode or jemplode do not have this entry.

I have read your FAQ multiple times, Tony. That's usually where I start with any issue that I have. But I'm just lost here with this. I realize that it's 15 years later, but I'm still trying to use my empeg every day and I'm finding it more and more difficult.

Okay, so I think I'm getting there. Definitely need to use full file directories for sure, otherwise the symlinks are broken. There's a lesson learned. I don't frequently work with symlinks.

Having said that, a 'cp -L' command successfully copies the index and data files (the actual targeted .mp3 instead of the symlinks themselves, hence the -L flag) of an mp3tofid-created /drive0 properly, as well as renaming said .mp3 so everything seems to be quite well there.

Just need to get rsync up and running properly and then I will have managed to bypass emplode entirely. Good stuff. Thanks for all of the help.

Keeping an eye on this. When you get it figured out, I'd like to give it a shot as well. Dealing with Emplode is frustrating these days and most of the time I'd rather just pull the drive and plug it in via USB for the massively faster speeds.

If I don't already have a FAQ entry on a step-by-step instruction guide for MP3toFid, I'm happy to take someone's submitted instructions and post them.

You don't need the "cp -L" step; rsync will reify the symlinks for you (also with the "-L" switch).

Correct. This was for a manual test with the empeg's drive physically connected to the Linux machine on which I ran mp3tofid. I didn't even use rsync as I haven't gotten that to work yet. I seem to have figured everything else out though so as soon as I get rsync to go, I can put together an actual step-by-step guide for getting it to work from scratch. Should help guys like me out.

You might want to read Using rsync to Synchronise empegs; it's not directly applicable, but might be useful...