Subsonic Forum

Any chance of getting Amazon Echo Support?

I've thought about this and it is not really an easy problem. Amazon has a lot of resources in getting vocal recognition to work against their own playlists, but the toolkit doesn't give developers quite all of that power. Maybe it can work by just playing playlists (which is how I first approached SubFire for the Fire TV and Stick, adding browsing and album views later), or work by using the tablet as a kind of remote - a pseudo-jukebox where App A (tablet or browser) browses server C, then tells App B (Echo) to load from the server C. The new Play Queue API would be great for this - you can use anything to save the play queue, then just have the Echo load it up and keep playing (and update the last song played as it goes).

Some ideas, all of which may be possible. I think I'm most thinking just doing the Play Queue is best for a start, followed by playlists, and not risk getting into full browsing 'til you're sure the voice recognition (and text reading) toolkits can deal with all of the names it has to read, some of which aren't necessarily English (or in the case of Ozric Tentacles, not any other language, either).

Please! The Echo is being touted as a voice commander to turn your lights on and for grandma to turn up the heat! Granted, it's not an audiophiliac-grade speaker, but still . . . .

. . . is there a way to maybe connect the subsonic library to the Amazon Music Library? Maybe store a playlist or two in the cloud. . . sort of use the Amazon Music Lib like a Subsonic Media Folder/Playlist in the cloud?!

I would happily make a one-time contribution for this feature.

It wouldn't need to be perfect... if voice recognition covered 90% of my library, I could live without the couple of albums released by "The Artist Formerly Known as an Unpronounceable Symbol".

Another app I use called "Our Groceries" on Android just added Alexa support. I am able to say "Alexa, tell Our groceries to add BROCCOLI to Vacation Food"

It understands the word broccoli (or anything else i throw at it) and reads the list of stores or other list names (e.g. Vacation food) I have put into the app

Why can subsonic not use something similar? I would think the index with the app would allow this.

Would love Echo support!

ok, for Echo / Alexa to work, a developer would need to do quite a bit.

https://developer.amazon.com/public/sol ... our-system

Note all that "oauth" stuff: Subsonic has none of that. It has no token-exchange authentication at all. As such, a developer would need to create an app (probably best to host it on Amazon AWS/S3) that would generate oauth tokens and use them to wrap the Subsonic server credentials. Those who use that service would need to trust that the author of that app doesn't do anything malicious with the logins they effectively have been granted, because there is no way that information can be used without being permanently stored by that app.

Granted, if a future release of Subsonic used oauth instead of its current (relatively insecure by comparison) mechanism, that would make parts of this much easier.

(There are reasons I try to make it very clear that neither SubFire nor SubFireRadio are sending any communications to my own servers, nor even do I run google analytics on it. Anybody who uses my apps? I have no idea where their servers are or how to log into them: none of that information ever gets to me.)

Only after the oauth wrapper and account linkage is done can the developer start writing alexa-skills for playing (and pause, skip, back), as wrappers around the Subsonic API that would then send the final music-playing URLs back to the device. In addition, of course, would be the need to figure out how the device can recognize when a song has finished in order to know when to start playing the next song in the queue.

There's also the issue of whether or not the Echo and Alexa systems would accept self-signed certs for HTTPS. The Fire Stick does not, in any way, and I can't override that (in spite of openly stating that, I still have 3 1-star reviews solely on that topic). The Fire TV and Tablets are very difficult to add a home-cert to. My guess is that Echo would not support it, so that's another thing that would disappoint many.

It isn't technically impossible, and I'm fascinated by the idea, but it is a LOT of work.

Well this seems promising!

https://developer.amazon.com/public/community/post/Tx1DSINBM8LUNHY/New-Alexa-Skills-Kit-ASK-Feature-Audio-Streaming-in-Alexa-Skills

It still doesn't solve the oauth problem, but it solves the problem of detecting when the song is complete. It should also support basic playback functionality.

A developer has created an Echo app for JRiver Media Center. See this video demonstration. This makes me think something similar should be possible for Subsonic.
https://www.youtube.com/watch?v=dqBOaoEssN0

I would love, love, love Echo/Subsonic integration.

Instead of starting off with something super complicated, where you could order up songs, albums, playlists, artists, etc. - what about starting off really simply?

There is a Chrome app called Perisonic that I've been playing with. All it does is play random tracks - that's it. If someone could code a similar Echo skill, the voice recognition aspect would just have to start the skill, and maybe recognize "Next" to advance a track.

Anyone willing to give this a shot? I would happily support a development effort along these lines.

+1