Hacker News Reader

Comment by mmastrac | Hacker News Reader

That is a blast from the past. MP3tag is older than Web 2.0, Twitter, Facebook and is still actively maintained.

https://web.archive.org/web/20010502171211/http://www.mp3tag...

My workflow back in the day was mainly thrift store CD to AudioGrabber. I still have a few CDs that only exist in high-bitrate MP3 format after losing the physical disk.

Lately I've been using MusicBrainz Picard to re-organize all of these ancient rips and then automedia to add parity. I'm still paranoid that Spotify will disappear one day and I'm afraid to lose my older music collection.

View on HN

My workflow is

1. Buying used and reasonably priced original music CDs

2. Ripping them with EAC[1] and an external LG BH16NS55 to FLAC format (takes 120 seconds per CD - this drive is FAST and ACCURATE)

3. Auto-import the ripped FLACs into my beets.io database via cronjob (which also unifies the metadata automatically in 99% of the cases)

4. Inplace-convert the FLACs to 192kbps mp3 via `beet convert`

5. Archiving the converted perfectly tagged FLACs to Bluray discs, as soon as the archive size hits 25GB

6. Point a self-hosted Navidrome instance and a Windows VM with iTunes to the beets folder

7. Use Substreamer App with Navidrome's smart playlists[2] and "favoriting" on my Android phone / iPhone as well as iTunes syncing my iPod Nano 7 via smart playlists

Works absolutely flawless and is less work than I expected. Since I automated everything possible, the only manual thing I need to do is the BUYING, the RIPPING and the Bluray ARCHIVING part.

1: https://pilabor.com/blog/2022/10/audio-cd-ripping-hardware/

2: https://github.com/navidrome/navidrome/issues/1417

> 5. Archiving the converted perfectly tagged FLACs to Bluray discs, as soon as the archive size hits 25GB

Do you have a rec for any long-life BR discs?

I use regular Verbatim 25GB disks. Since I keep the original CDs and usually use the Mp3s to listen, this is only a part of my 3-2-1 backup strategy. Nothing meant to be 200 years archival proof :-)

However, I never use anything but Verbatim, never had a bad experience with it.

Very cool archival setup! Funny you say that about Verbatim. I have a several old burned discs from the 2000s and 2010s where the dye has degraded and is no longer readable, but I have one specific blue Verbatim disc my childhood friend burned for me in the late 90s, and it still reads today!

> and it still reads today!

Exactly my experience. Verbatim is worth the additional cost in my opinion. I mainly do the archival as a hobby... not sure I ever gonna need it :-) However, I'm pretty scared of ransomware these days, so I tried to make my setup as ransomware proof as possible and zfs-auto-snapshot + self burned blurays with the most important data seemed like a good idea :-)

Verbatim is, was, has, and always will be, the gold standard.

Years ago I tried archiving to optical media and after a few short years, things were failing. Like all the things.

I am of the opinion that (multiple) hard disks might be the most recoverable. I might be wrong.

Na you don't. Of course I keep a HDD copy on my Backup-Server, I just mentioned the Bluray thing, because I try to keep my 24/7 System as clean as possible.

Why bother with MP3s in this day and age? I have a whole little flowchart like this too but one thing I'll mention is that I use cueripper with eac as a fallback. When I end up using EAC I run the result through cuetools to get the verification log and store it with the original rip in a sort of source directory. Then I split the single cue file to individual flac tracks (I convert them from wave if I had to use EAC) and tag them all/add images etc. the final destination is media monkey and an iPod running Rock box. I keep a spreadsheet with every CD and the rip results and whether metadata has been applied and whether it's been moved over to the iPod.

Great choice... I use Mp3 because its pretty much the works for everything format. Works on USB-Sticks for my car, the old Radio in the Kitchen, etc.

CueRipper looks nice, maybe I'll evaluate, but for now my workflow is totally fine.

The good part is, that if I one day choose to get rid of the MP3s, all I have to do is reimport my FLAC archive (of course I have a Hardisk-Version of the Bluray backup) and I'm done. With `beet convert` I can choose every other compressed format or quality and just need to wait a few hours to "recompile" my whole collection.

I use MP3 because it works on just about everything I own. Nothing is more annoying than when you have nicely prepared a USB stick with travel music and in the car nothing works and the computer is 500 km away.

This is why most people bother with MP3 still, we don't care that there is something more recent. Or rather, we do care but have no use for it yet. Not always is the new thing better than the old for your use case.

Many moons ago I decided to rip everything to AAC, until one day I brought cd full of mp4 files to my dad’s car…and realized that none of them could be played.

After that, it’s just mp3 (and flac)

I would use 320kbps mp3s. At this point the space savings from 192kbps isn't worth anything and it's one of those things you wish you'd thought through so you won't have to do these steps again.

Or V0 is acceptable as well. 192 is crazy...

I'm not an audiophile. For me 192kbps is good enough at the time, but I get your Point. Space should not be a problem these days and why not go as good as possible. Maybe I'll change that in the future.

Like I described above I keep the FLAC Archive on a Harddisk, so if any day I decide to change my library format (e.g. to AAC to use FLAC), it is like one rsync, one import and one convert.

Always use V0. 192 is not good enough for many things, 320 is a waste of space.

I'm gonna consider this for the future, thank you

I saw a study once, and it found 224kbps LAME mp3 worked for all types of music. So that is what I've always done.

What about the metadata? Do you listen to whole CDs at a time? I find that adding the tags for each song takes more time than all of the other steps you mentioned. Of course I use Mp3tag for that, but still, would be nice to auto populate somehow.

What metadata do you mean? Like the regular ID3 tags? Most tools so a decent job finding those on online DBs. I use musicbrainz Picard personally.

Track titles, contributing artists, year of issue, genre(s), that sort of data. Once filled in, say in EAC, it carries over mostly to other tools.

beets corrects all the metadata in a very automagic and decent way. If you never tried, it's open source software and totally free. It takes some time to configure and get used to, but it is worth it.

EAC (Exact Audio Copy) has a Metadata provider, which matches the audio CDs and pulls data from different sources.

That said, beet (beets.io) is much more convinient / accurate / easy to use and once it is configured properly. EAC has partly inserted the most important metadata stuff (artist, album, etc.) after ripping. So I use a cronjob, that runs `beet import` with "ignore if no match is found", that fully automates the process of embedding the metadata (also cover, lyrics and so on).

Works really well and I don't need to perform ONE mnaual step. Although Mp3Tag is a great tool (and I mean really great), doing all the tagging manually costs just too much time.

With a setup this bespoke, why not transcode to Opus or AAC or MP3 VBR?

I do something similar to OP but do the final transcoding to HE-AAC VBR 64kbps (ffmpeg params `-c:a aac_at -profile:a 4 -b:a 64k -aac_at_mode 2`). The tracks sound more than acceptable to me and this way I can easily store my entire collection (+80k tracks) into my iPhone. Modern codecs are wonderfully efficient.

Thanks for sharing your params... This is nice, maybe that's something I should consider. I need to test wether my car supports aac. My ipod does and kitchen radio is not that important.

You should See my audio books workflow with m4b-tool and audiobookshelf, which is probably even better ;)

Did you know ffmpeg has a non free Encoder (libfdk_AAC) thats sounds slightly better?

No problem, please do share any interesting part of your workflow if you have the time!

> Did you know ffmpeg has a non free Encoder (libfdk_AAC) thats sounds slightly better?

Ah yes. I'm not sure it sounds better than the audiotoolbox encoder though, at least it is reported that it doesn't here: https://trac.ffmpeg.org/wiki/Encode/AAC#aac_at although the hydrogenaudio source of this claim seems to be down, is still readable on the web archive at https://web.archive.org/web/20240407200855/https://wiki.hydr...

I'm desperately trying to publish a blog article about this topic for half a year now :-)

Maybe I'll never get it done, there is so much to cover... from

- what I know about iPods

- why modern Android Phones still can't compete in my opinion

- why the Apple Earpods can't change volume on Android devices and vice versa (https://tinymicros.com/wiki/Apple_iPod_Remote_Protocol)

- how you can harvest Earpods remote to create a durable good sounding headset using the right pinout

- that OneMore seems to have implemented the Apple Remote Protocol without anyone noticing

- why I wrote my own command line cross platform audio tagger (https://github.com/sandreas/tone)

- why I self-host my audio stuff

- and why nobody seems to care about an audio everything solution (music, audio books, podcasts, etc.)

So, let's hope I'll find the time :-)

That's a choice I made years ago. Maybe today I would change this, but I did not feel the pain to reencode my whole library yet. Maybe in the future. At least I know I can do it very easily by just reimporting the FLACs and change one config switch in beets :-)

Opus is more exotic and does not work on all my players though (e.g. my car or kitchen radio)

That is the beauty of a lossless archive :) makes sense

Thank you for this. I am trying to figure out what it is I want to do with my collection.

EAC doesn't seem like it would have anything set up for an autoloader like the Nimbie, but dbpoweramp does; on the other hand, the major music-sharing groups seem to prefer EAC.

Sadly, there's nothing out there that will be relatively accurate and precise when it comes to extracting beats per minute, key, and the nebulous "energy" characteristics. I would love to have those.

I think I will also have to work to do things like pushing lyrics and artwork into the FLACs, seeing what metadata "makes it" in the re-encoding to something like a 320kbps MP3.

I use EAC, because I made this choice a while ago, but there are some other tools around, that can even better automate the process of ripping to flac.

Speaking of metadata: Beets does this pretty well. You should try it, even if it takes a few hours to get used to it. There are a few good youtube tutorials.

Part of the consideration is that I would eventually like to get into the music sharing scene (I have fought this for a very very long time) and there's considerations of which tools and which source of metadata complicating my decisions. Which is the "preferred" among which group?

Beets sounds like it should be part of the mix, but so does MusicBrainz Picard.

Have you tried "Mixed in Key" for detection of BPM and key? I used this a long time ago for EDM and for that it works great. I don't know if it works reasonably well for other genres.

How did you get EAC working exactly? I tried to use it on numerous occasions and each time it would randomly freeze or repeatedly show a dialog box (I forget what it was now, something about tips maybe) until I closed the software. Even while I did nothing the boxes kept appearing. This would cause it to overlay itself endlessly in a loop.

There is a pretty good tutorial linked in my article:

https://captainrookie.com/how-to-setup-exact-audio-copy-for-...

My apologies, I misspoke. It's the program itself that, without my intervention, will go haywire, regardless of whether I've followed a guide or started from scratch. I seem to be the only one with this particular problem, so I never knew how to fix it.

I'd recommend to setup a fresh Windows 10 VM, install all the available upgrades, create a snapshot and then follow the linked guide.

The Optical Drive can be passed through to the VM.

If you still get errors, it's probably a hardware issue.

I've been using EAC over 20 years and have never seen anything like what you're seeing. (that said, I might be on a pretty old version, I'm not sure) Try to reinstall it and make sure you do not select the GD3 plugin on installation. It's the only option I'm aware of that might come with a nagware type dialog box.

> Lately I've been using MusicBrainz Picard to re-organize all of these ancient rips and then automedia to add parity. I'm still paranoid that Spotify will disappear one day

Spotify doesn't even have to disappear, it just has to lose access to the songs you care about or raise their prices to the point where they are unfair, or degrade their service in other ways until it's not worth using.

Personally, I'm paranoid about using internet based services to fetch metadata for my local media collection. I imagine their servers log those requests, each filename, the metadata pulled, IPs, and timestamps, and that they keep that data around for at least some length of time. Not sure how many sell that data, but what I'm listening to isn't really anyone else's business and I don't want the RIAA or anyone else to use that information as evidence against me somehow or coming around asking for proof of purchase for every MP3 they got from those logs.

Spotify does this thing for me where, no matter what you do, after an hour or so of listening you find yourself hearing the same 50 or so songs, over and over and over until you go mad. I still pay for it, and use it a bit, but playing my own collection is so much nicer.

I'm a fan of YouTube music's algorithm. I don't even have playlists. I just like songs that I enjoy, and whenever I need music I click something in the homepage and enjoy old and new songs that YouTube thinks I will like.

That’s why I disable all of Spotify’s smart playlist and auto play features. When my playlist/album ends I want to select something new, not be spoon fed whatever everyone else is listening to!

Same for me, I could never understand the hype around their suggestions. Even random playlists like „coffee house music“ (which I exclusively turn on when friends are over) seem to be personalised to me - so much so that they pretty much immediately veer off into heavy rock or other stuff that I may have come across, but that definitely isn’t good as a chill backdrop.

And even the much-praised weekly playlist is hit or miss - sometimes it’s weeks with abhorring stuff, then for once I get a good one with two to three songs I actually like.

Don't get me started on the weekly playlist. Spotify refuses to accept the fact that I don't speak Ukrainian. Yet a big chunk of the Playlist consists out of Ukrainian music. Probably Spotify is not able to differentiate between Ukrainian and Russian, which I listen to a lot

Considering sanctions, Ukrainians are probably the largest Spotify consumers of Russian content right now, and obviously will listen to Ukrainian content too. So their engine probably goes "oh you like Russian stuff, so chances are you're Ukrainian, here's some Ukrainian stuff for you".

Good point, that's probably it

I found so many great new Songs (ProgMetal and similar) through Spotify, but the quality of recommended Songs really deckined gor me in the past two years.

If you like Prog Rock and Metal, you may like those two online radios. I've discovered some interesting bands thanks to them:

https://www.progulus.com

https://radioarg.com/tmb/

Well, it does have some nice features. It autocreates playlists of my girlfriends and my music. And playlists of different genres I listen to. And playlists by release year of genres that I like. And playlists of songs I used to listen to some years ago.

Totally separate from the "song radio" playlists what you probably mean

No, I mean the regular old playlists, like, road trip indie, the shower playlist, late night jazz, what have you. The stuff with cover images and descriptions that looks totally curated. Those are at least enriched with music you listen to, if not downright generated individually.

And it’s awful. Like I’m stuck in an echo chamber. I want to find new music!

I find YouTube Musics Discover Mix good for music discovery, I guess with the downside that you need to actively use YT Music a bit to feed the algorithm. Spotify has a tendency to suggest stuff that is overwhelmingly popular but YT Music isn't afraid to throw in something a lot more obscure that the overlaps with the artists or genres you listen to.

Spotify has introduced me to many new genres but the algo has gotten worse over time.

Plexamp. Does the same kind of thing but with your own collection.

Maybe it's the settings I have enabled but typically Plexamp will move from one album to the next. Maybe because I typically play an album at a time?

If you haven’t already looked in to it, beets might be a solution for you

https://beets.io/

I highly recommend Musicbrainz Picard: https://picard.musicbrainz.org/

It will match against the Musicbrainz database and will acoustically ID your files, so the tags can be completely wrong and it can ID the song from it's sound fingerprint. Just dump folders of albums into the client, it will group and sort things and ID them. It works great.

It's a miraculous project. I have something like 300+ albums from 170+ artists and it tooks me only a few days to cleanly retag everything, with about 99% of the albums just working.

Musicbrainz Picard: highly recommended

I'll 2nd Picard. Been using it for years (and contributing to musicbrains for years).

I wish there had been something like MusicBrainz but for movies, they now have BookBrainz[1] so that would have covered most aspects of digital content.

[1] https://bookbrainz.org/

I only wish it wasn't written in Python - writing good GUI apps is way too hard with it, also the lack of static typing makes development of anything beyond simple scripts a potential minefield.

> It will match against the Musicbrainz database and will acoustically ID your files, so the tags can be completely wrong and it can ID the song from it's sound fingerprint.

For songs that have covers, it will ID the song as any of a number of similar covers. I just tried to use it to tag something from the Grease 2007 revival soundtrack (which, as of this writing, doesn't exist in Musicbrainz), and it happily identified it as the same song from the 1994 revival, which is wrong. This makes me hesitant to use it to identify songs if I don't already know what the identification should be.

I remembered that taking care of metadata of 1000+ mp3 music and syncing them between music players and backing up with CD-RWs were time filler. They still are, but I enjoyed doing so. Digital garden in web 1.0 era I could say.

I was so proud of my meticulously tagged mp3 collection, and even took the time to add album art to everything. I always wanted mp3s tagged with the original album they came from, even if they were from a greatest hits CD or something. (Looking back, this wasn't quite the right mindset, as sometimes the versions on a greatest hits CD or similar will be slightly different than the "real" album version, but it was my collection!)

I maintain mine. It's the only way to get guaranteed gapless playback in the modern era.

> I'm still paranoid that Spotify will disappear one day and I'm afraid to lose my older music collection.

Services eventually disappear. Not many people recall about Vitaminic, a gem of the early 2K that along MySpace gave countless unknown artists and bands the opportunity to put their music online, then one day poof! and it was gone with all its content. I had saved a few quite interesting tracks on my hard drive and attempted to search for the artists in the hope they moved elsewhere, but no way: they were gone forever. So, screw online services: I'm not going to waste any energy in something whose kill switch is in someone else's hand.

If the artists didn't own their own copyrights, there's a good chance that their stuff is available somewhere else by now. There are many companies out there that do nothing but put your music onto spotify and youtube for exposure; anything owned by a label is probably there.

Fun fact is that I have at least one CD where there was an audio track misprint (one track repeated). I didn't actually notice until I ripped it and used Picard to analyze the tracks.

I also once found two tracks where Spotify's own data was wrong (wrong audio files associated with the album/track names).

After failing in my attempts to escalate it through Spotify support several times, I ended up having to reach out to Glenn MacDonald on Twitter to get it fixed.

I love MusicBrainz Picard but the UI doesn't seem to allow me to mass-manage my TB collection of music. Just crashes after trying to load it all into RAM for 20 minutes or something.

I'm sure there's a CLI or something that can accomplish what I'm trying to do though.

Try beets[1]. It's likely the CLI answer you're looking for.

[1]: https://beets.io

> automedia to add parity.

What do you mean by this?

I’m guessing they’re referring to this: https://github.com/mmastrac/automedia

Reply on news.ycombinator.com