Skip to main content

Media Methods

It can be confusing about what Ampache does and what you need to do as an API client.

This page will document a few things about how playback works and what a client can expect from Ampache.

There are 2 binary data methods used for basic playback function:

And there are 3 helper functions which allow you to manage playback history and state but do not return media.

Finally, you can also call the function now_playing to get details about what your user is currently stremaing. (json and xml)

Explaining stream​

Stream is used to play a song in the same way as the Webplayer.

The URL generated by stream will match the URL property in all song object repsonses

"url": "https:\/\/music.com.au\/play\/index.php?ssid=cfj3f237d563f479f5223k23189dbb34&type=song&oid=115&uid=4&transcode_to=mp3&player=api&name=Chi.Otic%20-%20Are%20we%20going%20Crazy.mp3",
<url><![CDATA[https://music.com.au/play/index.php?ssid=cfj3f237d563f479f5223k23189dbb34&type=song&oid=54&uid=4&player=api&name=Synthetic%20-%20Red-GreenSmoke.mp3]]></url>

When you play a song; stream will:

  • Record a play in the database which is recorded as a stream
  • If you stream a new song too quickly the last song will be marked as skip
  • Play count is incremented UP for a stream and back DOWN for a skip

It is very important to note that play history is compared to the client string. You can play a song on a browser, start a subsonic stream and then an api client stream right after and all 3 plays will be recorded as streams because they are different clients.

Explaining download​

A download is not considered a stream and does not go through the same stat checks.

When you download a song:

  • Record the action as a download
  • A download does not perform skip checks and ignores any recent play history
  • Play counts are NOT incremented.

Download is used for caching files without the server thinking you've played 40 songs in 3 minutes.

Use the helper functions when playing downloads so you can inform the server when you actually play the files you download.

The Ampache media playback process​

Stat checks happen every time you call a stream or download URL (including part stream content-range calls)

e.g. https://music.com.au/play/index.php?ssid=cfj3f237d563f479f5223k23189dbb34&type=song&oid=54&uid=4&player=api&name=Synthetic%20-%20Red-GreenSmoke.mp3

When you use the API stream/download calls you are redirected to the play_url for the object. (Ampache calls them play_url's because they redirect to the /play/ directory of the Ampache site)

The checks are there to stop duplicate stats being recorded or spam from repeated calls to the same URL.

If you are only using the stream action, Ampache will perform these checks for you.

If you cache your files locally with download or are using a different client you can use the helper functions to notify the server.

The Ampache config file also allows you configure what you consider to be a skip.

The skip_timer setting is what the server uses to decide whether the new play is going to override the last play.

; Skip Timer Threshold
; This allows custom times to decide when a track is skipped
; Allows an integer to denote seconds, or a float to denote percentage
; POSSIBLE VALUES:
; 20 = 20 seconds.
; 0.3 = 30%
; DEFAULT: 20
;skip_timer = 20

Example, when using the default settings:

  • You stream TRACK1
  • TRACK1 is recorded as a 'stream'
  • You start to stream TRACK2 10 seconds after TRACK1 started
  • The 'stream' for TRACK1 will change to a 'skip'
  • TRACK2 is now recorded as a 'stream'
  • TRACK2 is recorded as your current now_playing track.

Explaining scrobble​

The scrobble and record_play methods are used for recording plays from saved information/data.

They're generally not used during playback as the stream action manages that process.

Scrobble uses parameters similar to how the Last.fm API works. (Which is why it's called scrobble.)

The idea is that if you are using a different player where you aren't able to get the object id you can scrobble to your Ampache server.

Scrobble is useful for plugins in different music players as an alternative to Last.fm or Libre.fm clients.

Example streaming plugin for rhythmbox written a few years ago. rhythmbox-ampache-fm

Explaining record_play​

record_play does the same thing as scrobble but only requires the object id of the song instead of searching by string data.

This is useful with playing cached/downloaded songs from an API client that has access to the id.

If you are playing a downloaded file or a cached version of the file there is no stream or download action happening.

To tell the server what you're doing and to record a stream you would call record_play when you start a track.

Example:

  • I cache song with the id 1240 using a download action
  • I play the mp3 that was saved in my file cache
  • That mp3 play is NOT a stream so I use record_play to set the stream in the database

Explaining player​

There is a new method for Ampache 6.4.0+ called player. (json and xml)

Player is all about giving the server an active status of what the client is doing.

Player does a bit more than scrobble and record_play allowing me to do a few more things

  • Support for more media types. (podcast_episode and video objects instead of just songs)
  • Update now_playing data (remove on stop add it back on play)
  • Shift last item play time on resumption of playback. (e.g. i pause a song at 20 seconds for a day and restart the song. the item will shift to NOW - 20 seconds as i haven't played any other song.)

It's similar to scrobble/record_play but the play/stop parameter gives me a bit more information about what you're doing

Final thoughts​

Hopefully that helps you understand what happens when you are streaming or downloiading form Ampache servers.

The idea is that you shouldn't have to think about anything when you stream. The server will do all the checks and updates for you.

When you download, you need to tell the server that you're playing a cached file.

Use scrobble/record_play/player to tell the server what you're doing.

Once you tell the server that you're playing with stream/scrobble/record_play/player methods the server will update the play status/history depending on your your data and current playback status.