Hacker News Reader

Comment by kstenerud | Hacker News Reader

I've got all my music ripped to FLAC, sitting on my file server, a little N40L microserver with 8tb mirrored storage (my music takes about 300gb total). On the server, I've got a script that uses ffmpeg to convert from flac to whatever format I want (currently aac and mp3, which my player devices support). The script also applies replaygain to keep the volume consistent. I run the script once and it processes the whole shebang.

This means that I control all of my music, stored in lossless format on a backed up server in my house, able to be deployed to any device I want whenever I want, in whatever audio format is popular at the time, in perpetuity.

View on HN

I know it was not your intention but your comment reminds me of the infamous HN comment on why dropbox is useless and would never be successful.[1] :))

"For a Linux user, you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem. From Windows or Mac, this FTP account could be accessed through built-in software."

[1] https://news.ycombinator.com/item?id=9224

> I know it was not your intention but your comment reminds me of the infamous HN comment on why dropbox is useless and would never be successful.[1] :))

Except the context of the argument is different. OP is doing this for convenience and privacy reasons along with the availability and no-cost argument. While it's a slight point to make it is an important one these days. With the Dropbox argument the old HN comment was making the point without the context of privacy, security or even availability of which we all have to consider today. We know service providers have removed or deleted purchases by users. Some providers remove content and make it paid for during adventageous times of the year (e.g. holidays) and, well, privacy doesn't need much expansion here. We also know vendors can lock you into hardware ecosystems through file formats. Finally if we all stop DIYing our media/files/etc the options will turn into expert level oniy. So the continuation of this being a thing keeps the open side of the conversation right side up. Lest we all end up boiled frogs.

The last sentence of that comment is:

> I know this is premature at this point, but without charging users for the service, is it reasonable to expect to make money off of this?

I googled for 'dropbox financials', and soon ran across a TC article about its IPO last year [1] which says:

> The company’s prospectus warns it has “a history of net losses”; anticipates increasing expenses and slowing revenue growth; and notes that it “may not be able to achieve or maintain profitability.”

I'm still not sure that snarky 10-year-old comment was far off.

[1]: https://qz.com/1214822/dropbox-is-filing-for-a-500-million-i...

Dropbox has been operating at a loss since its creation and is currently not making a profit, so the jury is still out on whether it is actually successful (or just a VC-subsidized service that got advertised a lot).

You are correct, but the metrics of success have always been different in tech. Regular businesses would be considered failures if they had 10 years in the red. But if you get a "unicorn" valuation and manage to IPO without even making a profit, well, that's the dream for a lot of HN'ers. If somebody out there makes you a multi-millionaire based either on faith in your company or in their ability to shift the risk to the next sucker, who're you to say no?

This gives me concerns about legacy. Henry Ford is still well known and quoted today not only because he built a well known company and iconic products a century ago, but because his company is still building cultural icons today (F-150 and Mustang).

I would much prefer to build a Ford than a Yahoo/Tumblr/Compaq, even if it meant my personal fortune was 10x or 100x lower.

>the metrics of success have always been different in tech.

It hasn't always been that way at all, but it certainly is now. I think this is one of the huge red flags about the modern incarnation of the tech industry.

Agreed, "always" is the wrong term. "Since the late '90s" would be more accurate.

This comes up all the time and it's silly.

If dropbox only had a high valuation on paper, but employees/founders/investors couldn't actually exchange that into real $ in their bank account then sure, you can say the jury is still out.

But that's not the case. They built a product, have tons of paying users, and have IPO'd with a market cap over $10B. It's not money that only exists on paper. Founders and employees can sell their shares for real money and buy real things.

Their shareholders aren't just a handful of VC's anymore - they are public. The jury has decided, it's successful

> The jury has decided, it's successful

Hypothetically, what if it never begins earning more money than it spends each year? At some point wouldn't it be unable to borrow any more (from banks or investors) and be forced to shut down? And if so, we still don't know if it can be profitable, right?

It’s successful to me as a user. Beats the heck out of that whole “set up an FTP server” recommendation. And how much is their loss due to growth focused investment?

My setup is far simplified from the parent's--I simply have all my music backed up onto an external HDD that is attached to my HTPC (roughly 2TB of MP3s and FLACs). The "Music" folder is shared over the network via the Windows share option, and all of my Linux machines are able to trivially mount the directory as well using a SMB/CIFS share.

There's so little effort required that I'd barely consider it a DIY solution, considering that most of the OS's can see it and access the drive without leaving the file explorer.

I also discovered it happily continues to work when my ISP had an outage, since all of the IPs were local and still accessible provided that the router was up.

Actually, I'm rather shocked that there isn't an all-in-one solution like what I've cobbled together. It's not like there's any super secret magic involved; you'd just need a nice user interface on top of it all.

There are even music streaming servers that you can deploy on a VPS (or locally) pretty easily. Pop in a little rsync script to move your converted lossy files to the streamer and you're set. It's just that nobody's built something that integrates everything yet, thus my Frankenstein setup.

You'll be surprised, but AFAIK Windows Media Player can do all of that.

Media Monkey does all of the things you mentioned, and with a GUI. Rip, store, convert, transfer.

I'm pretty sure there are Synology applications that also do what Plex does and with a nice phone app.

Have you heard of Plex?

Plex doesn't convert files for storage on a portable device, nor does it replaygain. It also has a terrible music selector interface and doesn't firewall punch.

Streaming is only part of the issue. I'm talking about the whole lifecycle, from backed up lossless to lossy-on-device in whatever format, or to a streaming service that integrates with your player devices, all from a single source of truth that you control 100% with no silly DRM or ads or restrictions.

>Plex doesn't convert files for storage on a portable device

Yes it does, you can pick what quality you want when you sync[0]

> doesn't firewall punch

It does so via a reverse proxy if required[1]

[0]https://support.plex.tv/articles/201053678-sync-media-to-a-d...

[1]https://support.plex.tv/articles/216766168-accessing-a-serve...

I know it does encoding/decoding for video at least; I don't know how its music features compare.

> Plex doesn't convert files for storage on a portable device

I don't know about Plex, but Kodi does this.

As much as anyone that gives tech support to their parents I can assure you I get your very fine and subtle irony.

In systems development, the KISS principle (Keep It Simple, Stupid) is precious, it is your safeguard to mental and emotional health.

As the MIT / New Jersey schism taught us decades ago, there is no one concept of "simple". Simple in design? In implementation? For the user? For the architect? To set up? To troubleshoot?

Which is simpler: a pile of files on a hard disk, or a playlist in an online music service? It depends.

It's like asking what's the simplest mode of transportation: a 1985 Toyota, a 2018 Toyota, a Lyft, a bus, or a bicycle. You could make a good case for any of these, depending on your situation. I know people who can't operate a manual transmission, so the 1985 Toyota would be problematic for them, but on the other hand, I've never fixed a 2018 Toyota (or a bus) with just the tools in my backpack and a water bottle.

I use git annex for my library. This way I can easily replicate my 3 TB archive to multiple devices. Since my notebook is size constrained, I can clone my archive and just pull the files I actually need.

Having CD's that go back to their launch (plus records/cassettes from before), I had to rip a couple of thousand CDs to FLAC so I'm totally with you on this one. Unfortunately a lot of people don't understand what's thrown away when converting to a lossy format, and the consequential horrible degradation that comes with converting from one lossy format to another. It's not like storage or bandwidth is an issue now (soon even the 2:1 or 3:1 lossless with FLAC won't be worth it). Just save the raw data and move on.

Of course it's more 'convenient' to trust cloud service X but (as recently discussed here) 'convenient' services can turn out to be pretty venal - not to mention short-lived (Microsoft Zune,various google services, random shuttered startups etc).

I do something similar (ripping to FLAC, putting on a tiny fileserver, except mine is an Antec ISK-300 with a single 6TB disk and external backups), and then use Logitech Media Server to serve it out to other machines.

"Squeezer" is excellent as a remote control for this, and I just have a stream URL I point players at to all play the same thing, or they can use DLNA and slurp music from the archive.

Works nicely, but does require more setup than most users will want to do when they can just plug in a "smart" speaker and stream from Spotify or whatever.

My current setup is a Docker container[0] with a musikcube[1] daemon in an ARM board (ROCK960 Model C) and on my phone I use the musikdroid client. At first, it was an investment of time, but now requires almost no maintenance.

[0]: https://github.com/hectorm/docker-musikcube

[1]: https://github.com/clangen/musikcube

You've solved the storage/archival half of the problem. What do you use on the UX side for playback? Do you just stream everything, or use some sort of an app for choosing which songs to pin to your mobile devices? Personally I use syncthing and keep a complete copy of all my mp3s (only ~5GB in my case) on my phone. I'm curious what others use though.

Have you looked at some selfhosted media platform like Jellyfin[1] or mpd[2] ? Those seems to be good privacy-wise.

[1]: https://jellyfin.github.io/

[2]: https://www.musicpd.org/

Add ampache [3] to that list. What it lacks in UI, it makes up for in features.

I've been tempted to setup a jellyfin instance for more seamless in-browser video playback, though.

Edit: funkwhale is also quite interesting, especially for the federation and music discovery aspect [4]

[3]: http://ampache.org/

[4]: https://funkwhale.audio/

MPD is nice in that it has many clients. MPDroid (available on F-Droid) is quite decent.

That's what I'm using. A central machine (that has an output to main audio system), with it being controllable from anywhere using various MPD clients. I tried to go the streaming route with MPD to enable music to be played in multiple rooms, but it was very finicky and susceptible to network problems. Instead, another output of the main machine goes to an FM transmitter, and any other audio system can pick it up. The sound quality is not pristine, but it's acceptable. Music in any room, all properly synced and centrally controllable.

"I've got all my music ripped to FLAC"

And that's why I won't do it, because there's no way I can rip all my cds, let alone my vinyls... it takes way too much time.

I set up a machine in the living room, and ran a script that would rip a disc as soon as it was inserted, then eject it.

Whenever I passed through, I would swap discs. If I had some free time, I sat in front of it and read while swapping discs.

It took a few months, but it was entirely doable.

I ripped my 1000+ cds in about 1 month. I specifically designed my scripts so that ripping was a separate task from "meta data" inputting. I'd bring a stack of 50ish discs to work and start ripping. The cd tray would open when it was done and I'd put in the next one and close the tray. That's the only interaction, and I could just deal with switching them all day. Then at night when I got home I'd do a meta-data pass, collecting the info from cddb/etc and finalizing everything.

I'm not going to say it's not work, but holy cow it's worth it in the end.

If you're running linux, just riping the CDs is fairly hands-off once it's configured:

abcde can be configured to save a flac with an embedded cue-sheet. You can make it totally non-interactive, and a simple "eject -t; abcde; eject" will load the tray, rip the CD and eject the tray.

Then just load a CD in every time the tray ejects. I ripped a bit under 300 CDs in a week with the only extra time spent on it being setting everything up.

No need to do it all at once, just when you’re playing the CD. But I guess your CD player is not attached to the computer?

Slightly of-topic but how do you discover new music to listen to?

Slowly and gradually. The music I love, I love. I don't like playing chase.

So sometimes they played a show/concert with an artist I already like. Sometimes it just comes up in the YouTube feed. Sometimes somebody points me to something and it takes.

I've been trying an experiment lately.

I took a few local venues that I like, and put the top song from each of their headliners for the next month into a spotify playlist, then just listened on shuffle while I was doing something else. Occasionally a song jumps out, and like that! I've discovered a new artist I like, and I can go to their upcoming show to boot!

Same way I have for the past 30 years. I listen from time to time to stations that play the kind of music I like. Sometimes I check out various places that have top 100 charts. If I find something I like, I go get it and add it to my collection. Rerun the script and resync my devices.

> I listen from time to time to stations that play the kind of music I like.

You're lucky! In my area, there are no such stations. All of the stations are essentially top 40 (in each of the major genres -- top 40 country, top 40 pop, etc.), talk, and sports.

There's nothing in any of those for me.

Internet radio may be able to solve your problem. Music podcasts, and internet radio sites/apps mean geographic location is no longer the limiting factor for what you can listen to.

There are big names like iHeartRadio and SlackerRadio but lots of smaller radio stations have internet streams and/or apps. Podcasts are nice because they can focus on niche genres or provide more information about artists/releases. Mixcloud is one of my favorite sites for listening to radio shows.

> Internet radio may be able to solve your problem

It's a problem that I've solved long ago, really. Internet streaming services bring their own set of problems, so I don't use them.

I do listen to podcasts (downloaded, not streamed), but not for music.

> There's nothing in any of those for me.

I can advice you to listen to fip, it's a french radio with every genre of music.

A couple of ways I do it:

* Use the Shazam app on my phone to identify music I hear in public places

* Go on Rateyourmusic.com, look up an album I like and find what "lists" RYM users have put that album on, or what message board posts have cited that album. It's a much more "human-centric" approach compared to websites where algorithms generate the recs.

You realize that those algorithms used by Spotify and whatnot are basically doing the same thing, right? “What songs did people who like your songs also like...”

They're not doing the same thing. I have no insight into how Spotify's algo works. When I go on RYM, I am reading human-written message board posts about WHY they're recommending this album. That additional context is far more valuable to me than a generic "Users also liked..." message that Spotify shows me.

That's true.

I attended a somewhat notable music festival this year and I noticed that the bands on the line-up had all the other bands from the same festival in the "similar" section, despite not being similar at all - (I'm assuming) just because other people interested in the festival were checking them out.

I'd like to know too. I also don't consider this off-topic considering we're contrasting streaming music with CDs, and one of the "value adds" of streaming is discoverability.

I think many of us curate a physical or digital collection and also subscribe to a streaming service.

I use Google Play Music and when I find something I like, I buy the CD.

RSS feed of music review blogs. Works well enough.

Bandcamp tags and notifications as well as following people with similar taste. Some reddit subs are also quite helpful

I was wondering if some media server [1] set up in local network and serving audio and media files in popular formats from an external USB drive would do the job for me. Don't have the time to experiment even with such a simplified setup.

[1] https://www.turnkeylinux.org/mediaserver

I just have all my music on Dropbox in max-quality MP3s. I use TuneBox on iOS to get a music player experience, and whichever MP3 players on my desktop OSes. Not lossless, but a simpler setup.

This feels like a business opportunity. Or, at least, an open source repo + a tutorial article that would attract donations.

How much money do you realistically see this generating? With how much effort for marketing and polishing? I would be impressed if OP could calculate even 1 dollar per hour and come out with a profit.

This is a niche of a niche of a niche.

There are already several projects that do this. Sub/Mad/Airsonic all allow hosting your own music and transcoding with ffmpeg.

There are numerous sites giving tutorials about how to set up such systems using OSS.

Obligatory question: could you share your ffmpeg script? Sounds like it perfect match my need for my classical music collection

Reply on news.ycombinator.com