diff options
author | Robin Gareus <robin@gareus.org> | 2015-01-17 22:38:48 +0100 |
---|---|---|
committer | Robin Gareus <robin@gareus.org> | 2015-01-17 22:42:20 +0100 |
commit | 49047e4dc9a35cb8f586204fdbe8cca55fc1026f (patch) | |
tree | a07ee180b769659573ff130cf5ff3eb5333eca1f /gtk2_ardour/engine_dialog.cc | |
parent | 03523f6703dbc9e422adcb8c346ab5c79d90a5c2 (diff) |
consistent display of period-size.
Diffstat (limited to 'gtk2_ardour/engine_dialog.cc')
-rw-r--r-- | gtk2_ardour/engine_dialog.cc | 18 |
1 files changed, 14 insertions, 4 deletions
diff --git a/gtk2_ardour/engine_dialog.cc b/gtk2_ardour/engine_dialog.cc index 2dfcf0088c..394a52b193 100644 --- a/gtk2_ardour/engine_dialog.cc +++ b/gtk2_ardour/engine_dialog.cc @@ -1036,15 +1036,25 @@ EngineControl::show_buffer_duration () uint32_t samples = atoi (bs_text); /* will ignore trailing text */ uint32_t rate = get_rate(); - /* Translators: "msecs" is ALWAYS plural here, so we do not - need singular form as well. - */ /* Developers: note the hard-coding of a double buffered model in the (2 * samples) computation of latency. we always start the audiobackend in this configuration. */ + /* note to jack1 developers: ardour also always starts the engine + * in async mode (no jack2 --sync option) which adds an extra cycle + * of latency with jack2 (and *3 would be correct) + * The value can also be wrong if jackd is started externally.. + * + * At the time of writing the ALSA backend always uses double-buffering *2, + * The Dummy backend *1, and who knows what ASIO really does :) + * + * So just display the period size, that's also what + * ARDOUR_UI::update_sample_rate() does for the status bar. + * (the statusbar calls AudioEngine::instance()->usecs_per_cycle() + * but still, that's the buffer period, not [round-trip] latency) + */ char buf[32]; - snprintf (buf, sizeof (buf), _("(%.1f msecs)"), (2 * samples) / (rate/1000.0)); + snprintf (buf, sizeof (buf), _("(%.1f ms)"), (samples / (rate/1000.0f))); buffer_size_duration_label.set_text (buf); } |