We had to work through an issue with Spotify this week. There seems to have been a change in Spotify’s protocols that wasn’t publicized, which seems to have happened before. We’ve run into issues with Pandora in the past.
For both, we use unofficial open libraries to access the services. We really wish they’d support official open libraries (Spotify used to) that would make it easy for the millions of open developers in the world to access their services. We make sure to remind them of the Communities they’re missing out on when we can.
We’ve reached out many times to try gaining official support from these and others, and won’t stop working for official support. We see people using Spotify and Pandora in skill metrics, so clearly people are using the skills. But we also see suggestions for Google Play Music and Logitech Squeezebox support. There’s the Internet Radio Skill, and a skill for your Emby Server.
But, what does the whole Community really want for music on Mycroft? Obviously, one solution won’t work for everyone, but having a good ranking of what the Community wants will help us focus on the most impactful and desireable platforms.
obversaly there isnt one solution for ll. Spotify is the streaming service I prefear, but others would like others. And some services arrent posible in every contry. So there have to be different services to choose from.
Beeing a smart speaker music is a key funktionality.
Yes, one solution? I don’t think so. My wife uses Spotify, and I use Pandora.
The issues I have with Spotify is it requires some sort of an app to play.
I use the Mycroft flavors picroft and different versions of linux also. I realize that it should download the raspian version of Spotify, but without a gui, I cannot get it to play through my speakers. It automatically connects to the next Spotify app it can find…my androidTV.
At least with Pandora, it plays through my speakers.
Also, I have quite a large personal music collection that I would like to be able to access and play also. I think there are a couple of skills that might allow that, but I haven’t looked too close yet.
Is the a dab style radio player mycroft skill,
Example - hay mycroft play radio one,
Or play music from a Nas home server,
Not shore how accessing a play list would work.
Our Spotify Skill uses the spotipy python package, which may be what raspotify also uses. If you just want to use Spotify as a Skill you don’t need to install anything but the Skill, it should handle the rest for you.
I would say rather low. As far as I understand the Apple Music API is intended to be used by iOS/macOS apps and their token based authentication system. Theoretically you can generate the required application keys yourself, but I think this won’t be feasible for the Joe Average users as you have to register for the Apple Developer program - and you will be charged 99 bucks for one year…
The real question for us is if there is enough demand in the community for Apple Music? If it’s a key service then we’ll of course look into how we can make it happen. What we’re trying to work out is how many people actually use it, compared to the alternatives?
Current test results are that the skill works for Amazon-EU domain accounts but does not work for Amazon-US accounts (for reasons that are beyond the skills scope). Nevertheless any additional tester - EU or US - is highly welcome.
I’d voted for Plex and MPD as I personally don’t like to use third party services and I host all my online needs, yet I could use the free spotify account from time to time to discover some new albums.
I would LOVE having the music (or any sound that’s not the mycroft voice) reduce to 25% volume (or a predetermined level) when the wake word is detected and it’s listening for commands. I have the mark I and it’s almost impossible for the mark I to hear me when it’s playing Spotify or pandora.
We call this ducking, and we’ve started to include it in the code. The challenge on the Mark 1 is the limitations of the hardware. Just hearing that first wake word is quite hard as it’s unable to differentiate what it’s playing from what it’s hearing from other sources.
The mic array on the Mark 2 will improve this dramatically as it includes native support for voice activity detection including direction of Arrival, beamforming, noise Suppression, de-reverberation and acoustic echo cancellation. We’ve previously made attempts to perform similar things at a purely software level for the Mark 1, but to make a long story short, this is extremely difficult.
