midi controller question

[Galvonaut]

I am looking into ways of controlling lasers live in as accurate a manner possible and wondered if it were possible to daisy-chain controllers and if so, what software allows this and how could I set it up?

My idea is to have an apc40 and a midi keyboard - Main cues and control by apc40 and specific short cues controlled using a keyboard for more accurate triggering.

Cheers
Keith

[norty303]

Yes, midi is specifically designed for this sort of thing.
IF you specify which midi channel you want a device to output and listen on, you can have multiple controllers using the same note/cc values.

E.G. You set up the main cue grid to use the default channels (APC is a bit different in this respect as it uses a range of note values over a few midi channels iirc - read the APC midi map documentation to see this properly)
You should then still have a few midi channels remaining to assign to your short cues


To simplify the concept a bit (because i realise that probably doesn't make much sense)...

You have 2 cues. You can set one to be triggered by C1 midi channel 1, and the other to be triggered by C1 midi channel 2.

So you set one contrller to only output on midi1, the other to only do midi2.

You may need to force a cue to only listen to a specific midi channel, by default a lot of software is set to 'any', but once you've decided on your mapping, it is quite straightforward to do.

[Galvonaut]

Brilliant, thanks Norty - makes sense

Keith

[The_Doctor]

One thing to watch if daisy chaining is propagation delay. It's critical for audio, less so for visual events because vision can't resolve to below 5 ms but ears can. So put any audio gear first in the receiving chain, if any.

One thing that will reduce this risk, and increase reliability (because there is no chain to break this way) is to make a splitter box, or buy one. While MIDI signals converging to one line must be merged by a device with full understanding of MIDI protocol and timing, diverging one to many is as simple as connecting a bunch of 5-pin DIN sockets and inputting to one of them and feeding other inputs from the rest. This is safe too, in any system whose individual inputs are properly opto-isolated as the MIDI spec requires. (Done by the equipment maker, you won't have to add this).

In the case of a MIDI keyboard and a controller, likely both will have an input, and a merge capability built in otherwise that input makes no sense. In this case, plug one controller in series to the other, in whichever order works best for you, then plug the second one's output to a distributing net of parallel DIN sockets to feed everything else. That way you will even get the first controller's MIDI in for somethign interesting you haven't thought of yet.

But once you want more than three controllers it is DEFINITELY time to buy a good standalone MIDI merge box.

EDIT:
If you have a MIDI monitor, try this: Set it to receive a CC message from a continuous controller. (Note that CC is Control Change, NOT continuous controller, some are dedicated as switches even if their raw messages are capable of using all 128 values. Stupid distinction, but at times it gets important!)
Whip the controller through its range fast, cleanly, from one end to the other. A good MIDI controller should sample the value at rapid intervals, so a short burst will appear on the monitor with widely spaced values. A BAD controller will take every value it passes through even though it cannot render them meaningfully in the time it's meant to do it in, so you will see a delayed flood of closely spaced messages that takes its own sweet time to complete after you've whipped the control through its full range. This will cause nasty timing errors, and possibly damaged messages too, and even merging two controllers separately won't help if one of them does this. If you see this you really need a better keyboard or controller! There is no way round this, it is not a weakness of MIDI, it's just that many cheap tools use it stupidly, and this is the best test to sort the good from the crap.

I strongly suggest that anyone reading this does this test! That way we can all benefit from a shortlist of known good, and known bad, controllers.

[Galvonaut]

Thanks Doc!

Yes, timing is really important for this application as I have very little time in the track to play cues and they are spaced very close together, so any backing up won't be acceptable. Does the apc40 have a fast sample rate? What should I look for in the specs to check this?

Can you (or anybody else) recommend a good brand of midi keyboard? I don't want too many keys - maybe 20 or so.

Keith

[The_Doctor]

I wish I knew. I'm hoping someone will test the APC40 and post some results. One way to test is the MIDI-OX monitor by Jamie O'Connell and Jerry Jorgenrud (freeware). I'm using the v3 16 bit one, but there may be later versions. If you whip a controller through its range fast there will be about 7 messages or so instead of all 128. It shows raw data so the timestamps are hexadecimal, counting milliseconds since the port was opened.

I tested a Doepfer 'Pocket Control' (because I have one, it's convenient, and it shows that other Doepfer gear is likely to work well) and found 26 or 27 milliseconds apart. This is a device that multiplexes 16 controls so their keyboard electronics which scan fewer continuous controllers may be faster. Given that 24 ms is low enough to give a keyboard a feeling of near-instantaneous response for fast staccato performance, this is at least good enough, probably. (There are people who think they need 2ms latency or better, but given that ears can't resolve easily below 5 to 10 in most people, I don't think they do need 2 ms). Anything lower than 25 is usually great. (More examples: a Peavey PC1600X samples at 21ms, and a Yamaha SY99 as often as every 6 ms!).

A keyboard to avoid is anything from the Alto Professional range. I like them, but they have the poor quality sample-everything problem I described. It may even be worth fitting Doepfer's MKE electronics to a keyboard you like, as you're familiar with repair work, but it might be tougher to do than it may appear. I think Doepfer sell a few cheap small keyboards too. 2-octave keyboards are common now, but go for one of the better ones after testing it for time-based sampling. Oxygen-8 will NOT do. Been there...

EDIT: Remember that if you're automating much of your signals and can add delays in the receiving end, you can send setup signals early to keep the instant for manual commands free of traffic.

[Galvonaut]

Thank you Doc.

Looking into keyboards now.

I'm not sure I understand your edit.

Keith

[Galvonaut]

The APC40 doesn't have a midi connection - it's via usb. I've already run out of USB ports so need dacs to be via ethernet to free some space up.

I really can see why this hobby is so expensive!

Keith

[norty303]

I used an Evolution MK-225c for a while, and seemed to be ok, although i never did the tests Doc has detailed with the CC.
It's now sitting in the lockup for some time, so if you're interested in it, drop me a PM.

I've already run out of USB ports so need dacs to be via ethernet to free some space up.

Why not just use a powered usb hub?

All of my work machines have one, as 2 or 3 USB ports simply isn't enough, and I find a powered hub is better when using multiple DACs anyway.

[Galvonaut]

I used an Evolution MK-225c for a while, and seemed to be ok, although i never did the tests Doc has detailed with the CC.
It's now sitting in the lockup for some time, so if you're interested in it, drop me a PM.



Why not just use a powered usb hub?

All of my work machines have one, as 2 or 3 USB ports simply isn't enough, and I find a powered hub is better when using multiple DACs anyway.

Thanks Norty - Yes, powered hub is the way to go I think.

PM incoming