Subsonic 6.0 released

[sindre_mehus]

I'll try to give some background for why I chose to change the software license.

Subsonic has never been a typical open-source project. For all practical purposes I have been the sole contributor and committer. I always intended it to be a one-man project. From time to time people have kindly offered to contribute, but I have respectfully declined (because I prefer writing code rather than reviewing it). As a consequence I am in my full
right to change the license as I see fit.

Subsonic is clearly a commercial undertaking and has been for many years. I established a company (Subsonic AS) in 2010 with which I try to make a living. The source code is the only asset of this company, and it is problematic when the exact same code (minus the need for a paid license key) is made available by others. I've also registered that some forks of Subsonic are (or at least plan to) sell their own licenses. It has also been confusing for many users that they have to pay for an open-source product.

Then there is the issue I mentioned earlier about third-party licenses I have obtained and that are part of the source code. In the future I may use more commerical third-party components (e.g., html audio and video players). This is more complex to deal with if the Subsonic source is open.

As a service to people/projects who would like to use Subsonic as a basis to build their own stuff I made a github repo with the source of the last stable release (5.3) on https://github.com/sindremehus/subsonic. This contains 11 years worth of software development.

Regards, Sindre

[gingerbeast121]

Thanks Sindre,

Before others come in (I'm sure whom have a differing opinion) I would like to say...

--love the look and feel of the new interface
--unlike other forks (madsonic for example), subsonic is lighter and more stable and of course is the development source anyway. When I first looked I raised an eyebrow to some forks charging a premium and using the same/a modified version of the subsonic source
--Not sure about others but I pay my yearly subscription to listen to music and (to a lesser extent) watch video. As long as the application continues to support this functionality (and work well), I will continue to pay to use it

Good look with the development, looking forward to some exciting new features.

[Exrace]

Please see my post here reporting a possible issue with the windows based installer.
viewtopic.php?f=2&t=16641

[surfrock66]

I would like to make a case for releasing 6.0 stable as GPL code, representing the final release of the software as open source.

As indicated above in this thread, 6.0 beta 1 was released as GPL code. This is not complete code, and is likely full of bugs and issues.

Many of us, myself included, upgraded to 6.0 stable upon the day of release, April 30. This upgrade completed before you announced that you were closing the source code.

So, to switch to a fork (especially one based off of 5.3) will require a downgrade of my database, potentially a settings rollback...all in all, it puts us in a very bad position.

Additionally, by providing the 6.0 beta 1 code on Feb 5 2016, you've left us with a starting point for 6.0, but it would require 3 months of your bugfixing that you already completed to get it workable. For people like me that paid you to work on this under the expectation that this was open source software, and had already upgraded my instance by the time you announced that you had changed the agreement, it feels like I was scammed.

So now we know what the agreement is. The final release of code at the time of announcement was 6.0 stable, and providing patches and commits from the time between beta 1 and stable would be an amicable way to treat the community that has supported you this far. Going forward, we know any patches and commits are NOT open source.



There's one more thing. I would like to see a public commitment to keeping at least the subsonic API open source. A lot of apps, like DSub, utilize the API, and should you change your mind on something, it would be unfortunate for apps to have to pivot in such a way that compatibility with open source implementations breaks. If you decide to change say the authentication engine, and apps like DSub either update and leave Libresonic or Madsonic or Ampache users behind or split themselves, it's a very bad deal. You have become the steward of an API that is so widely adopted that I think we deserve at least a commitment in openness of the API, separate from the source code of Subsonic itself.

[phillipmcmahon]

I would like to make a case for releasing 6.0 stable as GPL code, representing the final release of the software as open source.

As indicated above in this thread, 6.0 beta 1 was released as GPL code. This is not complete code, and is likely full of bugs and issues.

Many of us, myself included, upgraded to 6.0 stable upon the day of release, April 30. This upgrade completed before you announced that you were closing the source code.

So, to switch to a fork (especially one based off of 5.3) will require a downgrade of my database, potentially a settings rollback...all in all, it puts us in a very bad position.

Additionally, by providing the 6.0 beta 1 code on Feb 5 2016, you've left us with a starting point for 6.0, but it would require 3 months of your bugfixing that you already completed to get it workable. For people like me that paid you to work on this under the expectation that this was open source software, and had already upgraded my instance by the time you announced that you had changed the agreement, it feels like I was scammed.

So now we know what the agreement is. The final release of code at the time of announcement was 6.0 stable, and providing patches and commits from the time between beta 1 and stable would be an amicable way to treat the community that has supported you this far. Going forward, we know any patches and commits are NOT open source.



There's one more thing. I would like to see a public commitment to keeping at least the subsonic API open source. A lot of apps, like DSub, utilize the API, and should you change your mind on something, it would be unfortunate for apps to have to pivot in such a way that compatibility with open source implementations breaks. If you decide to change say the authentication engine, and apps like DSub either update and leave Libresonic or Madsonic or Ampache users behind or split themselves, it's a very bad deal. You have become the steward of an API that is so widely adopted that I think we deserve at least a commitment in openness of the API, separate from the source code of Subsonic itself.

This seems a sensible and reasoned approach. Would represent also amicable parting terms.

Sent from my Nexus 6P using Tapatalk

[toolman]

The way a lot of people react to Sindre's announcement that Subsonic will no longer be open source strikes me as really bizar.
Sentences like : "For people like me that paid you to work" really are beyond any border of realism.
No, you never paid Sindre to work for you. You paid to be able to use his software.
And since it is his software, he is entitled to do whatever he wants to do with it.
Sindre has in extenso explained why he made that decision.
It's kind of sad that now Sindre has decided to put his interests first, he suddenly went from being Number One Nice person to being a scammer.
Way to go. First some name-calling and than demanding guarantees.
Sindre has no obligation, whatsoever, to anyone.
I, for one, can understand Sindre's point of view and will go on supporting him and Subsonic.

[mikes]

So, it looks like Subsonic 6.0beta1 source (r4715, 2016-02-04) and 6.0 final distribution files are on sourceforge ( https://sourceforge.net/projects/subson ... /subsonic/ ), and are all distributed under GPL v3 ( https://sourceforge.net/projects/subsonic/ ). So, at least there's a 6.0beta base for forking, and no issue with reverse engineering from the distribution files.

I'm all for Sindre making a living on it, and I paid for it long ago to support him. But, that wouldn't have happened if it hadn't been open source at the time. He certainly has a right to not release future versions under an open license, but users should be made aware of that before a new release goes into beta, not after.

[surfrock66]

The way a lot of people react to Sindre's announcement that Subsonic will no longer be open source strikes me as really bizar.
Sentences like : "For people like me that paid you to work" really are beyond any border of realism.
No, you never paid Sindre to work for you. You paid to be able to use his software.
And since it is his software, he is entitled to do whatever he wants to do with it.
Sindre has in extenso explained why he made that decision.
It's kind of sad that now Sindre has decided to put his interests first, he suddenly went from being Number One Nice person to being a scammer.
Way to go. First some name-calling and than demanding guarantees.
Sindre has no obligation, whatsoever, to anyone.
I, for one, can understand Sindre's point of view and will go on supporting him and Subsonic.

I can't speak for everyone, but the concept of software freedom is a critical feature for me. More has been written on this topic than I can summarize here, but the concept of a free and open license is absolutely a feature for many people. It has security implications, support implications, philosophy implications, and feature implications.

Even if what you said is true, and that we paid to use the software he made and weren't paying him...he removed a major feature without discussion or notification to his customers. And he did it very late in a major feature release cycle, which if nothing else is bad form since it led people to commit to a new version without knowing a critical feature was being removed.

I'm a lifetime subscriber, and often contribute to other free software products that I use. In fact, Subsonic's model where only Sindre contributed was likely a disservice to himself...the forks show there was desire to expand and modify features, and all he would have had to do is merge in changes. These are potential customers willing to do work for him to provide improvements and potential features, all without any obligation to be paid out. It's even likely some features could have brought in more users, leading to more donations and ultimately more money for Sindre. Everyone would have benefitted.

But he didn't do that, and he didn't accept patches, and now he's locked the software's source down because he wants to change the model. That's fine, I'd really like to continue to use the product but will not personally contribute to closed source software. I will switch to one of the forks, and I really do wish Sindre the best...I hope it works out for him, though I expect that a large bunch of potential customers that choose to run their own media streaming service (as opposed to a spotify/google play/itunes type solution) are the type of people that value free software and will rule Subsonic out. I believe this move will be financially negative in the long run, which is a shame.

[tocheeba]

If you don't like it, find another solution. It's his right to do what he wants with this software, and it's still the best DIY hosted music solution - so I'll continue to support it.

[kountaphit]

To piggyback on this, there's no longer the option to Select All / Select None

I already posted it into the thread for the beta2:
The options to remove selected items from the playing queue are missing. See screenshot:

[DaveWut]

The lack of support will push me to switch to an open source version of Subsonic. When the software was open source, I had no problem modifying part of the code that was buggy by myself, now I can't. See, this issue is still there since the time I've had created this post: viewtopic.php?f=5&t=8894. It's only one example among all the other that are ignored by the developer.

[BloodyIron]

I've been using Subsonic now for several years. It has been extremely reliable for me. But, closing the source, I'm not okay with that.

1) I am paying to support the developer move forward. I understand that, and that's why I'm paying in the first place. I have no issue with paying for access to the mobile features and others. Also, the price is extremely reasonable.

2) I also want to support Open Source projects, not closed source projects. Why? Because if Sindre stops for whatever reason, someone else can pick back up the work.

3) Should the time come where I want to, I cannot audit closed source software. This has become an issue for me, and if closing the source is permanent, I will move to another suite.

Sindre deserves my money, and the money of those who use Subsonic. He works hard, and while he is just one man, I think ensuring that he can get paid to do this is important. However, closing the source code is a deal-breaker for me. There are also beliefs that I have which support FOSS development over closed source. Namely, everyone benefits because it's out there. But closed source cannot be supported by people who don't have the code or ability to open it.

Please Sindre, find another way to make this work for you. Because closing the source doesn't work for me.

[BloodyIron]

I also wanted to make another v6 related post, but not related to the source code.

I've posted in a few areas how I cannot send video over chromecast any more. I'm getting pretty fed up that nobody has responded to my posts, even if to confirm it isn't working for them.

These forums are dead, barely used, even by Sindre. More and more this project feels just like Sindre's direction, and not that of the community, or what the paying people want.

For example, a feature like this, that was previously working, is just broken. No responses, no comments, nothing. And I am generally powerless without downgrading.

What about other basic feature requests? Such as cross-fading. There are many feature requests that have gone unheard or unaddressed for literally years.

I'm getting fed up with these things. You may call me entitled, but I'm a patient person, and my waiting time is almost over.

[acroyear]

Believe it or not, decent dj-quality cross-fading is *HARD*; even gapless playback is a harder problem than you might think, given the tools available.

HTML5's audio tag does not support gapless at all. I've ideas on how to hack an approximation of gapless by using two audio tags (where one is buffering up while the other is playing, and I try to start the second one as soon as the first one stops), but that only works if the browser decides to buffer them the way one might expect (this is not as reliable as you might think), and if you get the right events from the audio tag (also not as reliable as you might think). On webkit browsers on Android including the webview used when you package in Cordova/PhoneGap or Amazon's Fire, very little of that works right...though Chrome 51 is supposed to resolve a bit of that someday.

To cross-fade automatically, you're looking at determining when towards the end of one song to start the fade and how long into the other to end it and an algorithm (as in, a standard set of numbers, like 3 seconds out, 1 second in) that works fine for 'pop' songs that already fade will break utterly for other gapless playback situations like classical music or live albums. Good web radio stations and professional DJs that use mp3 kits work by having extra metadata associated with the file (either in ID3 or externally) that indicates when a fade-out should start and if a fade-in should be permitted (or not)...this works because they actually have a very small playlist (generally no more than a thousand songs), and by almost never playing from truly gapless sources like live albums. Other web stations either don't bother with cross-fading, or the cross-fading cuts across the material in a very distracting manner.

The HTML5 web audio library is not an easy to work with, and it isn't until you actually code something up that you figure out only 1 or 2 browsers may support some critical aspect of the actual standard (or they say they support it, but it totally breaks when you try it). There are wrapping libraries that may handle a lot of those, but the better ones cost money. This appears to be something Sindre has said he is interested in pursuing, if I read the announcement above correctly. Plus maybe he gets it to work in the web version, but then there's getting it to work in jukebox mode which routes through a totally different audio system since it has to work directly on the audio card through Java's audio system.

I personally think web audio is still 2 years away from really working the way developers would like in order to support applications like this with any cross-browser consistency and stability.

[alphawave7]

Insofar as Chromecast: I can play audio just fine on my ADT-1 (precursor to Nexus Player), but NOT on my Chromecast Audio's (living room/bedroom/home group). The audio appears to transfer to the device, but the throbber in the songlist is hung, whereas it spins when on the ADT-1 properly during playback. On video, my chromebook plays Subsonic vids great, but flowplayer appears to not have a Chromecast link/capability built in. Casting the tab 'works', but barely..choppy as hell. No guesses as to why ADT-1 works while CCAudio's do not.