Add replaygain backend using ffmpeg's ebur128 filter.
The album gain is calculated as the mean of all BS.1770 gating block powers.
Besides differences in gating block offset, this should be equivalent to a
BS.1770 analysis of a proper concatenation of all tracks.
Just calculating the mean of all track gains (as implemented by the bs1770gain
backend) yields incorrect results as that would:
- completely ignore track lengths
- just using length in seconds won't work either (e.g. BS.1770 ignores
passages below a threshold)
- take the mean of track loudness, not power
When using the ffmpeg replaygain backend to create R128_*_GAIN tags, the
targetlevel will be set to -23 LUFS. GitHub PullRequest #3065 will make this
configurable.
It will also skip peak calculation, as there is no R128_*_PEAK tag.
It is checked if the libavfilter library supports replaygain calculation. Before
version 6.67.100 that did require the `--enable-libebur128` compile-time-option,
after that the ebur128 library is included in libavfilter itself. Thus we
require either a recent enough libavfilter version or the `--enable-libebur128`
option.
When setting up bpd tests, two servers are startet: first a control server, then
bpd. Both send their assigned ports down a queue. The recipient only needs bpd's
port and thus skips the first queue entry.
Use a `multiprocessing.Queue` instead of a `multiprocessing.Value` to avoid the
manual polling/timeout handling.
TODO: Strangely Listener seems to be constructed twice. Only the second one is
used. Fix that and then remove the code working around it.
The bpd test bind a socket in order to test the protocol implementation. When
running concurrently this often resulted in an attempt to bind an already
occupied port.
By using the port number `0` we instead let the OS choose a free port. We then
have to extract it from the socket (which is handled by `bluelet`) via
`mock.patch`ing.
Return a namedtuple CommandOutput(stdout, stderr) instead of just stdout from
util.command_ouput, allowing separate access to stdout and stderr.
This change is required by the ffmpeg replaygain backend (GitHub
PullRequest #3056) as ffmpeg's ebur128 filter outputs only to stderr.
*All* URLs were checked manually, but only once per domain!
I mostly concerned myself with URLs in documentation rather than source
code because the latter may or may not have impactful changes, while the
former should be straight forward.
Changes in addition to simply adding an s:
- changed pip and pypi references as their location has changed
- MPoD (iOS app) url redirects to Regelian, so I replaced those
- updated homebrew references
Notable observations:
- beets.io does have HTTPS set up properly (via gh-pages)
- beatport.py uses the old HTTP url for beatport
- as does lyrics.py for lyrics.wikia.com
- https://tomahawk-player.org/ expired long ago, but the http page
redirects to https regardless
- none of the sourceforge subdomains have https (in 2019!)
The real MPD ignores `noidle` when the client is not idle. It doesn't
even send a successful response, just ignores the command. Although
I don't understand why a client would fail to keep track of its own
state, it seems that this is necessary to get ncmpcpp working.
The playlistid command is supposed to list the whole playlist if no
argument is provided, but we were accidentally trying to look up an
impossible negative id in that case causing an error to always be
returned.
When using pytest's test collector, any class with a name starting with
Test is collected. If it notices that the class has an `__init__` member
then it skips it with a warning since it's probably a false positive.
This isn't a big deal, but we can avoid warnings like this:
test/test_ui_importer.py:33
beets/test/test_ui_importer.py:33: PytestCollectionWarning: cannot collect test class 'TestTerminalImportSession' because it has a __init__ constructor
class TestTerminalImportSession(TerminalImportSession):
simply by renaming TestX to XFixture.
By comparing `sys.path` as setup by nose vs. testall.py it seems that we
weren't adding the top-level beets directory to the path. The script was
also previously changing the working directory before running the tests.
This is a hacky but effective way to work around a problem when running
the tests in parallel where two different test executions want to use
the same port.
