MIDI Clock Input #702

ywwg · 2015-09-02T00:10:41Z

This PR implements master sync using an external midi clock as the source. It does not add a midi output clock. It is still early and has some bugs, and the skin support is extremely basic and just for testing, but it works.

Includes unit tests.

Other cleanups.

Also add a couple simple sync tests.

Conflicts: build/depends.py

ywwg · 2015-09-02T00:11:17Z

src/controllers/midi/portmidicontroller.cpp

@@ -173,6 +173,7 @@ bool PortMidiController::poll() {
        if ((status & 0xF8) == 0xF8) {
            // Handle real-time MIDI messages at any time
            receive(status, 0, 0);
+            continue;


This is a big bug in current master.

Because there will also be MidiSendClock

ywwg · 2015-09-03T00:06:00Z

Ah, I forgot that I hadn't created a way of enabling or disabling midi clock source on a per-controller basis. I'll have to figure that out or disable this by default before merging.

daschuer · 2015-09-03T09:41:20Z

src/controllers/midi/midicontroller.cpp

    setDeviceCategory(tr("MIDI Controller"));
+    m_pClockBpm.reset(new ControlObjectSlave("[MidiSourceClock]", "bpm"));


Why you do not use the constructor in the init list?

daschuer · 2015-09-03T10:35:31Z

Do you have a clue how accurate this is to beat-match with a external device.
I am afraid this will not work without using the midi time stamp.
Do you think it will be hard to use it?

daschuer · 2015-09-03T10:39:40Z

src/controllers/midi/midisourceclock.cpp

+    const double bpm = static_cast<double>(tick_count - 1) / kPulsesPerQuarter
+                       / elapsed_mins;
+
+    return math_clamp(bpm, kMinMidiBpm, kMaxMidiBpm);


Clamping is always wrong here. Can we return just 0 and let the calling code decide?

Be-ing · 2018-04-18T03:31:02Z

This pull request has not been updated in years so I am closing it to reduce the clutter of open PRs. If you solve the merge conflicts and would like to have this merged in the future, please reopen it. If you want to continue working on this, I think we should consider how to make this more generic so the code could also be used for Ableton Link.

daschuer · 2018-04-18T20:56:45Z

This is still a desired feature.

Holzhaus · 2020-09-10T00:53:48Z

@ywwg Syncing from MIDI clock would be a nice feature. Any plans to continue working on this?

ywwg · 2020-09-16T12:11:28Z

blast from the past! I'll see how bad the conflicts are

kbdnr · 2020-09-16T12:44:57Z

Any reason to opt for midi over something like Ableton Link which will compensate for latency and solve output clocking? I see the original PR was made in 2015 which predates link.

In my experience, while the midi standard specifies tick/cycle, this is not as commonly standardized across devices (before even looking into how devices handle stop/start messages).

Holzhaus · 2020-09-16T13:04:41Z

Any reason to opt for midi over something like Ableton Link which will compensate for latency and solve output clocking?

Ableton Link support would be nice as an additional feature, but cannot replace MIDI clock because it's not nearly as widely supported by actual hardware. For example, MIDI clock output is necessary to sync the TR-S drum machine on the Roland DJ-505 to a playing deck in Mixxx.

…-clock-input

Holzhaus · 2021-02-07T00:51:09Z

Let's try to get this PR merged. I tried to fix the merge conflicts here: #702

Tests don't pass due to some debug assertions. @ywwg would be great if you could take a look.

Be-ing · 2021-02-07T00:55:32Z

Let's try to get this PR merged. I tried to fix the merge conflicts here: #702

I think you meant to link ywwg#11

Holzhaus · 2021-02-07T01:13:29Z

Let's try to get this PR merged. I tried to fix the merge conflicts here: #702

I think you meant to link ywwg#11

Oops ;-)

Resolve merge conflicts for MIDI clock output PR 702

Holzhaus · 2021-02-08T00:46:39Z

Currently this PR only considers a single input clock source. IMHO each attached controller might provide a separate input clock, hence each controller should act a separate clock provider, correct?

I'm not really familiar with the details of the sync API, but I have some PoC code for per-controller COs here: #2884

Be-ing · 2021-02-08T00:48:42Z

I'm quite concerned about how this PR expands the scope of MidiController which is already overburdened with implementing the legacy mapping system. I started considering how to decouple controller hardware IO from the legacy mapping system and this seems like it might make it more challenging...

daschuer · 2021-02-08T00:51:53Z

The idea of midi clock sync is to have a single clock source for all devices. Normally a sequencer is used or a dedicated device like this: https://www.musicstore.de/de_DE/EUR/E-RM-MIDIclock-/art-SYN0004625-000?campaign=GShopping/DE&ProgramUUID=rrLAqJarLjMAAAFl.gZyjI8h&gclid=EAIaIQobChMI7YzM6ojZ7gIVF-DtCh31ZQaREAQYAiABEgJ2FfD_BwE

daschuer · 2021-02-08T00:54:22Z

Of cause a controller can also do the job. Do we know controllers with this capability? In this case only one is the clock provider. the others are the clock receiver. Mixxx can also become the clock source, but this is probably a bad idea because all the jitter we have until the signal is on the midi line.

Holzhaus · 2021-02-08T01:19:19Z

Of cause a controller can also do the job. Do we know controllers with this capability? In this case only one is the clock provider. the others are the clock receiver. Mixxx can also become the clock source, but this is probably a bad idea because all the jitter we have until the signal is on the midi line.

My Roland DJ-505 sends MIDI clock to the computer (for syncing the decks with the built-in TR-S drum computer), and can also receive MIDI clock from the computer to sync the TR-S to a playing deck.

Holzhaus · 2021-02-08T01:30:50Z

I think each MIDI device should have the same sync COs as a normal deck. This means that I can either ignore the controller's clock signal, or enable sync lock (i.e. Midi output clock follows sync clock), or even set it as explicit leader (sync clock follows midi input clock). If I have another drum machine plugged into my computer, it should also be possible to switch between the two clock sources (or ignore both) by toggling the sync button on my midi device.

daschuer · 2021-02-08T01:38:37Z

That sounds reasonable, but I think the use case for this PR is the central clock case, where Mixxx is synced to a central clock like any other instrument.

ywwg and others added 14 commits June 14, 2015 15:35

MidiClock: a class that listens to incoming midi clock messages

5d826c9

Enable -std=c++11 support.

e46cceb

MidiClock: Actually commit the source files

5518af1

Midi Clock master works

098c92e

MidiClock: remove atomic / threading stuff since it's not needed..

02cd418

Other cleanups.

MidiClock: Add indication that midi is running or not.

a9df9a9

MidiClock: Add a sync tweak button to adjust for midi offset

b1f93bd

Also add a couple simple sync tests.

Merge branch '1.12' into midi-master

26e4ecd

MidiClock: Fix midi not updating follower rate sliders

afdd57b

LateNight: Fix effect unit getting too big when fourth unit is populated

96936f4

move midi sync to below right decks

f9ff1ca

MidiClock: Always update bpm on every callback

8286af1

Merge branch 'master' into midi-master

421edb7

Conflicts: build/depends.py

MidiMaster: Cleanup, clamp bpm calculation to a range

4857f4a

ywwg reviewed Sep 2, 2015
View reviewed changes

ywwg added 2 commits September 1, 2015 20:38

MidiMaster: Rename MidiClock to MidiSourceClock

e1ebbb5

Because there will also be MidiSendClock

Merge branch 'master' into midi-master

eb405df

daschuer reviewed Sep 3, 2015
View reviewed changes

Be-ing closed this Apr 18, 2018

daschuer reopened this Apr 18, 2018

Be-ing added stalled and removed stalled labels Apr 18, 2018

Be-ing added this to the stalled milestone Apr 18, 2018

Be-ing changed the base branch from master to main October 23, 2020 23:54

ronso0 marked this pull request as draft November 17, 2020 16:01

Merge branch 'midi-master' of https://github.com/ywwg/mixxx into midi…

98f313d

…-clock-input

Holzhaus mentioned this pull request Feb 7, 2021

Resolve merge conflicts for MIDI clock output PR 702 ywwg/mixxx#11

Merged

Holzhaus changed the title ~~Midi master~~ MIDI Clock Input Feb 7, 2021

Merge pull request #11 from Holzhaus/midi-clock-input

e3468de

Resolve merge conflicts for MIDI clock output PR 702

github-actions bot added build code quality labels Feb 7, 2021

ywwg closed this Jul 28, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MIDI Clock Input #702

MIDI Clock Input #702

ywwg commented Sep 2, 2015

ywwg Sep 2, 2015

ywwg commented Sep 3, 2015

daschuer Sep 3, 2015

daschuer commented Sep 3, 2015

daschuer Sep 3, 2015

Be-ing commented Apr 18, 2018

daschuer commented Apr 18, 2018

Holzhaus commented Sep 10, 2020 •

edited

Loading

ywwg commented Sep 16, 2020

kbdnr commented Sep 16, 2020 •

edited

Loading

Holzhaus commented Sep 16, 2020 •

edited

Loading

Holzhaus commented Feb 7, 2021

Be-ing commented Feb 7, 2021

Holzhaus commented Feb 7, 2021

Holzhaus commented Feb 8, 2021

Be-ing commented Feb 8, 2021

daschuer commented Feb 8, 2021

daschuer commented Feb 8, 2021

Holzhaus commented Feb 8, 2021

Holzhaus commented Feb 8, 2021

daschuer commented Feb 8, 2021

		setDeviceCategory(tr("MIDI Controller"));
		m_pClockBpm.reset(new ControlObjectSlave("[MidiSourceClock]", "bpm"));

MIDI Clock Input #702

MIDI Clock Input #702

Conversation

ywwg commented Sep 2, 2015

ywwg Sep 2, 2015

Choose a reason for hiding this comment

ywwg commented Sep 3, 2015

daschuer Sep 3, 2015

Choose a reason for hiding this comment

daschuer commented Sep 3, 2015

daschuer Sep 3, 2015

Choose a reason for hiding this comment

Be-ing commented Apr 18, 2018

daschuer commented Apr 18, 2018

Holzhaus commented Sep 10, 2020 • edited Loading

ywwg commented Sep 16, 2020

kbdnr commented Sep 16, 2020 • edited Loading

Holzhaus commented Sep 16, 2020 • edited Loading

Holzhaus commented Feb 7, 2021

Be-ing commented Feb 7, 2021

Holzhaus commented Feb 7, 2021

Holzhaus commented Feb 8, 2021

Be-ing commented Feb 8, 2021

daschuer commented Feb 8, 2021

daschuer commented Feb 8, 2021

Holzhaus commented Feb 8, 2021

Holzhaus commented Feb 8, 2021

daschuer commented Feb 8, 2021

Holzhaus commented Sep 10, 2020 •

edited

Loading

kbdnr commented Sep 16, 2020 •

edited

Loading

Holzhaus commented Sep 16, 2020 •

edited

Loading